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 |
|---|---|---|---|---|---|
We previously showed that newborns congenitally infected with Trypanosoma cruzi ( M+B+ ) display a strong type 1 parasite-specific T cell immune response , whereas uninfected newborns from T . cruzi-infected mothers ( M+B− ) are prone to produce higher levels of proinflammatory cytokines than control neonates ( M−B− ) . The purpose of the present study was to determine if such fetal/neonatal immunological environments could alter the response to standard vaccines administered in early life . Infants ( 6–7 months old ) living in Bolivia , an area highly endemic for T . cruzi infection , and having received Bacillus Calmette Guerin ( BCG ) , hepatitis B virus ( HBV ) , diphtheria and tetanus vaccines , were enrolled into the M+B+ , M+B− , M−B− groups mentioned above . The production of IFN-γ and IL-13 , as markers of Th1 and Th2 responses respectively , by peripherical blood mononuclear cells stimulated with tuberculin purified protein derivative of Mycobacterium tuberculosis ( PPD ) or the vaccinal antigens HBs , diphtheria toxoid ( DT ) or tetanus toxoid ( TT ) , as well as circulating levels of IgG antibodies against HBsAg , DT and TT were analyzed in infants . Cellular responses to the superantigen SEB were also monitored in M+B+ , M+B− , M−B−infants and newborns . M+B+ infants developed a stronger IFN-γ response to hepatitis B , diphtheria and tetanus vaccines than did M+B− and M−B− groups . They also displayed an enhanced antibody production to HBsAg . This was associated with a type 1-biased immune environment at birth , since cells of M+B+ newborns produced higher IFN-γ levels in response to SEB . M+B− infants produced more IFN-γ in response to PPD than the other groups . IL-13 production remained low and similar in all the three groups , whatever the subject's ages or vaccine status . These results show that: i ) both maternal infection with T . cruzi and congenital Chagas disease do not interfere with responses to BCG , hepatitis B , diphtheria and tetanus vaccines in the neonatal period , and ii ) the overcoming of immunological immaturity by T . cruzi infection in early life is not limited to the development of parasite-specific immune responses , but also tends to favour type 1 immune responses to vaccinal antigens .
Infectious diseases are a leading world-wide cause of morbidity and mortality in childhood , against which vaccination remains the best prevention measure [1] . However , protection induced by vaccines is of limited effectiveness in early life owing to the relative immaturity of the neonatal immune system . Moreover , the fetal/neonatal immune system is initially polarized toward a Th2 immune environment which appears essential for the survival of the fetus [2] , [3] . Indeed , both dendritic cells and T cells present quantitative and qualitative defects in the neonatal period , limiting the development of CD4+ Th1 cell responses essential for the control of intra-cellular pathogens [2] , [3] , as well as the production of antibody responses [4] . Nonetheless , neonates are in some cases able to develop mature T cell responses . This has been demonstrated in congenital infections with Trypanosoma cruzi [5] and cytomegalovirus ( CMV ) [6] , in infection with Bordetella pertussis in early life [7] , and after early vaccinations with Bacillus Calmette-Guerin ( BCG ) [8] or the whole cell pertussis vaccine [9] , [10] . Additionally , BCG vaccination at birth has been shown to increase both cellular and humoral responses to other vaccines such as hepatitis B and poliomyelitis vaccines [10] . Active maternal infections may also modulate neonatal immune responses to vaccines , as demonstrated in newborns of mothers chronically infected with helminths , who developed a Th2-biased response to BCG vaccination , by contrast with those born to non-infected mothers [11] , [12] . The modulation of immune responses to vaccines in infants from mothers infected with intracellular parasites , and having experienced such congenital infection has heretofore not been investigated . Chagas disease , or American trypanosomiasis , caused by the protozoan parasite Trypanosoma cruzi , is a major cause of morbidity and mortality in Latin America where 8–10 million people are currently estimated to be infected [13] . It has also become an important health issue in the United States and Europe due to large-scale migration of Latin Americans over the last few decades [14] . The parasite is primarily transmitted by insect vectors ( in endemic areas of Latin America ) , blood transfusion and congenitally ( in all areas ) [15] . Even the frequency of transmission by vectors and blood transfusion decreases as a result of vectorial control programmes and improvement of blood bank screening , mother-to-child transmission of T . cruzi presently cannot be prevented and has thus become an important route of transmission [16] . Recent estimations indicate that at least 15 , 000 newborns are likely to be congenitally infected with T . cruzi each year in Latin America [17] and 2 , 000 in North America [18] . In Europe , such transmission also becomes a problem in migrants originating from endemic countries [19]–[21] . In Bolivia , a highly endemic area for Chagas disease , we have reported that 17% of pregnant women are chronically infected with T . cruzi and that congenital transmission occurs in 5 to 6% of the cases [22] . We have showed that congenitally infected newborns develop a parasite-specific T cell immune response comparable to that of adults [5] as well as phenotypic and functional modifications of their NK cells [23] . On the other hand , newborns of T . cruzi-infected mothers are prone to produce higher levels of pro-inflammatory cytokines in comparison to those born to non-infected mothers [24] . The present study aimed to determine if such modifications of the immune environment in infected or uninfected newborns of T . cruzi-infected mothers could modulate immune responses to vaccines administered at birth or in early life . Cellular and/or humoral responses to BCG , hepatitis B virus ( HBV ) , diphtheria and tetanus vaccines were compared in infants living in Bolivia .
Two different patient groups from Cochabamba ( Bolivia ) were enrolled in this study: one of neonates ( born at the German Urquidi Maternity ) and one of infants ( 6 to 7 months old; followed at the Manuel Ascenci Villaroel paediatric Hospital and the Viedma Universitary Hospital , UMSS ) . The scientific/ethic committee of the Universidad Mayor de San Simon ( UMSS ) approved the study , and informed written consent was obtained from the mothers before blood collection . Maternal T . cruzi infection was assessed by standard serological techniques . T . cruzi infection in neonates and infants was diagnosed by detection of parasites in umbilical cord or peripheral blood by microscopic examination of heparinized microhematocrit tubes , or in some cases by hemoculture , as previously described [16] , [22] . Congenital infection was assessed at or close to birth for 28 neonates and 10 infants . For 3 other infants , T . cruzi infection was diagnosed later on when they presented at 6–7 months old at the consultation . Congenital infection was inferred because these infants lived in an area free of vectorial transmission , had not travelled in endemic areas , and had not received blood transfusion . Three groups of neonates and infants were established: congenitally-infected ( M+B+ ) ; non-infected born to T . cruzi-infected mothers ( M+B− ) ; and uninfected controls born to uninfected mothers ( M−B− ) . A total of 68 newborns and 60 infants were included in the study ( M+B+: 28 newborns and 13 infants; M+B−: 19 newborns and 36 infants; M−B−: 21 newborns and 11 infants ) . Mothers of neonates and infants included in the study were all asymptotic , even when infected with T . cruzi . They originate from the same rural region surrounding Cochabamba and live in similar socioeconomic environment . The sex ratio in all groups of newborns or infants was similar ( data not shown ) . Clinical data of the newborn groups have been previously reported [22] and Table 1 shows mean ages and weights at the time of blood collection , as well as historical clinical data at birth , of newborn and infant groups . At the time of blood collection , the infant age in the 3 groups was comparable and M+B− and M−B− infants were all asymptomatic . At birth , such infants had similar APGAR scores and maturity parameters . However , M+B+ infants presented at birth slightly lower mean gestational age , weight and APGAR scores; three of them showed one or two symptoms known to be associated with congenital T . cruzi infection ( i . e . , splenomegaly , anasarca and petechia ) . When examined at 6 to 7 months , M+B+ infants were asymptomatic , but had still slightly lower mean weight as compared to the other infants . As soon as the diagnosis of congenital T . cruzi infection was established , treatment of infected neonates/infants ( 7–10 mg/kg/day of benznidazole for 30 days , [22] ) was proposed to the mother . From the 13 M+B+ infants under study , 9 were treated within the two first months after birth , and 4 remained untreated at 6–7 months of age when blood was collected for the present immunological study . One remained untreated since her mother refused the treatment , and 3 were treated from 6–7 months old onwards since the diagnosis of infection was not established previously ( see above ) . Cure was confirmed in treated cases by the subsequent negative results of direct parasitological examination ( microhematocrit tubes ) and T . cruzi serology . All M−B− and M+B− infants presented with negative serology for T . cruzi at 6 to 7 months of age , indicating that they had not acquired the infection after birth . In accordance with the recommended Bolivian vaccination programme , all newborns received the BCG vaccine ( BCG , Serum Institute of India Ltd ) at birth by intradermic injection . Infants were then vaccinated against diphtheria , tetanus , whooping cough , hepatitis B and type B Haemophilus influenzae ( Hib ) by receiving intramuscular injections of Tritanrix mixed with Hiberix ( GlaxoSmithkline Biologicals , Rixensart , Belgium ) at the ages of 2 , 4 and 6 months . At the same time , they also received the oral vaccine against poliomyelitis ( attenuated vaccine , trivalent OPV , Sanofi Pasteur , Paris , France ) . Infants included in this study have all received the BCG at birth . At the time of blood sampling , 53 . 5% of the infants had completed the vaccinal schedule ( 3 doses of Tritanrix + Hiberix + OPV ) , the others having received 2 doses . The proportion of infants having received the 3 doses was similar in each group . Cord blood ( CB ) and infant peripheral blood ( PB ) were collected in sterile endotoxin-free heparinized tubes . Plasma was stored at −20°C until use for determinations of antibody levels . Mononuclear cells ( CBMC and PBMC ) were immediately isolated by Nycoprep density gradient centrifugation ( Nycomed Pharma AS , Oslo , Norway ) . CBMC and PBMC ( 2×106 cells/mL - duplicates ) were then cultured in RPMI containing 10% fetal calf serum ( Biowhittaker , Lonza , Verviers , Belgium ) , either alone to measure spontaneous production of cytokines , or in the presence of one of the following antigens: tuberculin purified protein derivative of M . tuberculosis ( PPD; 5 µg/mL; batch RT49 , Statens Seruminstitut , Denmark ) , diphtheria toxoid ( DT; 10 µg/mL ) , tetanus toxoid ( TT; 10 µg/mL ) , or the surface S antigen of HBV ( HbsAg; 10 µg/mL ) . DT , TT and HbsAg were kindly provided by GlaxoSmithkline Biologicals ( Rixensart , Belgium ) . A positive control for T lymphocyte stimulation was included by incubating cells with the staphylococcal enterotoxin B ( SEB; 10 ng/mL; Sigma-Aldrich ) activating non specifically T cells ( superantigen ) . After 6 days of culture at 37°C in 5% CO2 humidified air , cell cultures were centrifuged and supernatants conserved at −70°C for further cytokine analysis . IFN-γ and IL-13 cytokine concentrations were used as markers of Th1/Th2 responses , respectively . Cytokine levels were measured in culture supernatants by ELISA using commercial kits ( IFN-γ: antibody pairs and standard from Biosource , Invitrogen , Merelbeke , Belgium; IL-13: ultrasensitive kit from Bender Medsystems , VWR International , Leuven , Belgium ) . Standards were non-glycosylated recombinant cytokines ( Biosource and Bender Medsystems ) . Detection limits were 2 pg/mL for IFN-γ , and 1 pg/mL for IL-13 . Spontaneous release of cytokines by unstimulated cells was generally undetectable or very low . In the latter case , cytokine concentration was subtracted from levels obtained after stimulation with SEB or vaccine antigens . IgG antibodies against HbsAg , DT and TT were measured in plasma samples by ELISA using commercial kits ( HbsAg: Diasorin , Brussels , Belgium , DT and TT: Genzyme Virotech GmbH , Rüsselsheim , Germany ) . Results in international units ( IU ) were derived from WHO standard sera provided with the kits . Antibody concentrations above 10 IU/mL for hepatitis B , and 0 . 1 IU/mL for diphtheria and tetanus were considered to be protective [25]–[27] . IgG sub-classes against HbsAg were measured by ELISA as follow : Nunc Maxisorp plates ( VWR ) were coated with HbsAg ( GSK Biologicals ) at a concentration of 1 . 5 µg/mL in PBS overnight at 4°C . After blocking with 1% bovine serum albumin ( BSA ) in PBS for 2h at 37°C , plasma samples were diluted in PBS containing 0 . 3% BSA and 0 . 05% Tween 20 and incubated for 2h at 37°C . Sample dilutions were 1/50 for IgG1 and 1/10 for other IgG sub-classes . Mouse monoclonal antibodies specific for the Fc fragment of each human IgG subclass ( clones HP6001 , HP6014 , I7260 and HP6050 for anti-IgG1 , G2 , G3 and G4 antibodies respectively , Sigma-Aldrich , Bornem , Belgium ) were diluted 1/500 , and then incubated for 1h at 37°C , followed by rat monoclonal antibodies specific for mouse IgG coupled to horseradish peroxidase ( clone LO-MK-1 , Imex , UCL , Brussels , Belgium ) at 1 µg/mL for 1h at 37°C . After each step , plates were washed with PBS containing 0 . 05% Tween 20 . Finally , substrate and chromogen were added ( hydrogen peroxide and 3 , 3′ , 5 , 5′ tétraméthylbenzidine , BD Biosciences , Erembodegem-Aalst , Belgium ) and absorbance at 450 nm was measured after 30 min colour development and stopping the reaction with 2N sulphuric acid . Positive and negative internal controls were added in each ELISA plate in order to correct the absorbances for plate variability . We found a very good correlation between the levels of total HbsAg-specific IgG measured with this ELISA and those obtained using the commercial Diasorin kit ( r = 0 . 842 , p<10−4 , n = 25 ) . Absorbances lower than 0 . 01 were considered a negative result . Results were expressed either as arithmetic means±SEM or geometric means . Statistical comparisons of concentrations of cytokine or antibody levels between groups of newborns or infants were performed either with the Mann Whitney test or by two-way analysis of variance ( ANOVA ) , followed by Tukey test for multiple comparisons . Comparisons of proportions were carried out using the Fisher exact test . Statistical analyses were conducted using GraphPad Prism software ( GraphPad Prism 4 , San Diego , CA ) or the Statistical Software Package SPSS version17 . 0 .
Fig . 1 shows that , upon SEB stimulation , cells from M+B+ newborns produced significantly more IFN-γ in comparison with the M−B− neonate group ( geometric means 4372 and 551 pg/mL respectively ) . Interestingly , such congenitally infected neonates displayed at birth comparable levels to those observed in controls at 6–7 months ( geometric means in M−B− infants : 4726 pg/mL ) . Meanwhile , IL-13 levels remained similar in all neonate groups . Comparable results were seen in the M+B− and M−B− groups . This indicates that the in utero immune environment associated with congenital Chagas disease induces an early shift in the neonatal immune response toward a type 1 response . IFN-γ and IL-13 were not detected at birth when CBMC were incubated with DT , TT and HbsAg vaccinal antigens . However , as shown in Table 2 , these cytokines were detected in 6 to 7 month old infants . Two way Anova followed by a test for multiple comparisons ( Tukey ) indicated that the IFN-γ responses in the different infant groups was not influenced by the used antigen ( p>0 . 05 ) , allowing us to compare the responses after globalizing the results of each antigen . This showed IFN-γ responses to vaccinal antigens to be significantly higher in congenitally-infected M+B+ infants as compared to each other groups ( M+B+ vs . M−B− : p = 0 . 007 ; M+B+ vs . M+B− : p = 0 . 0003 ) whereas they were similar between M−B− and M+B− infants ( p>0 . 05 ) . Thus , M+B+ infants produced meanly 10–12 times more IFN-γ to HbS and DT , and 3–4 times more of this cytokine in response to TT . IL-13 release to these 3 vaccinal antigens did not differ between groups . This highlights a trend towards a type 1 response for vaccines against hepatitis , diphtheria and tetanus administered in congenitally infected neonates . Since the benznidazole treatment received by M+B+ infants before vaccination might have some confounding effect on stimulating immune responses [28] , data were also analyzed separating treated and untreated infants ( at the time of blood collection ) . Similar results were observed in both M+B+ subgroups ( data not shown ) . As indicated in Fig . 2 , mean levels of IgG antibodies against the vaccinal antigens HbsAg , DT , TT at birth were similar in the 3 groups of neonates , and high enough to be considered protective ( 67 to 100% above the protection thresholds of 10 IU/mL for hepatitis B , and 0 . 1 IU/mL for diphtheria and tetanus ) . This indicates that mothers had been immunized and that the transfer of such maternal antibodies during pregnancy was not affected by T . cruzi infection . In the 3 infant groups , HbsAg-specific antibody levels were significantly higher than in newborns . This sustains HbsAg-specific antibodies to be produced by the infants in response to vaccination , although maternally transmitted antibodies may still be present at low levels their sera [29] . Inversely , DT- and TT-specific antibody levels were lower in infants than at birth , suggesting that maternally transmitted antibodies may have restricted their production [30] . Nevertheless , they are close to antibody levels described by others in vaccinated infants [31] . Interestingly , M+B+ infants produced noticeably more antibody in response to the vaccine against hepatitis than those of the other groups . Though not statistically significant , a similar trend was seen for the M+B+ DT- and TT IgG antibodies . For the reasons mentioned above , M+B+ HbsAg-specific antibody levels were also analyzed by separating treated and untreated infants , and , again , results were similar in both M+B+ subgroups ( data not shown ) . Since the polarization of the immune response is known to modulate the isotypes of antibodies produced [32]–[34] , we next determined the IgG subclass profile of the HbsAg-specific antibodies in vaccinated M+B+ infants . As shown in Fig . 3 and as expected from previous reports [35] , [36] , IgG1 was the main subclass of IgG antibody found in infants vaccinated against HBV regardless of the group . Comparisons between groups show that M+B+ infants harboured 2 to 4-fold higher levels of IgG1 , IgG2 and IgG3 than infants of the other groups , while IgG4 levels remained low in all groups . Such variations likely account for the greater production of total HbsAg-specific IgG described above in these M+B+ infants , likely reflecting their stronger type 1 immune environment . We also studied the Th1/Th2 cytokine responses to the BCG vaccine by stimulating mononuclear cells with mycobacterial PPD . CBMC ( newborns ) did not produce either IFN-γ or IL-13 after PPD stimulation , regardless of group ( a total of 32 newborns were tested ) . By contrast , PBMC from most infants having received the BCG vaccine at birth produced cytokines in response to PPD ( Fig . 4 ) . The proportions of IFN-γ responders was ≥70% and similar in the 3 infant groups . However , M+B− but not M+B+ infants , produced on average 5-fold more IFN-γ in response to PPD in comparison to control M−B− infants . IL-13 levels produced by M+B− infants in response to PPD were low , and the proportion of IL-13 responders was slightly but significantly inferior to that of both other groups . This suggests that , even in the absence of congenital T . cruzi transmission , the maternal-fetal immune environment associated with maternal infection contributes to enhance the Th1 response to the BCG vaccine administered at birth .
We have investigated the effect of congenital Chagas disease and/or maternal T . cruzi infection on immune responses to standard vaccines regularly administered in neonates and infants in Bolivia . Our results show that: i ) congenitally T . cruzi-infected newborns ( M+B+ ) display an early ability to produce the type 1 cytokine IFN-γ; ii ) their specific responses ( M+B+ ) to the hepatitis B vaccine , and to a lesser extend to diphtheria and tetanus vaccines , administered later after birth are characterized by higher levels of IFN-γ and/or protective IgG antibodies; and iii ) non-infected infants born to T . cruzi–infected mothers ( M+B− ) secrete higher amounts of IFN-γ in response to BCG vaccination administered at birth . The responses to PPD that we observed in BCG-vaccinated infants were Th1-oriented in the 3 studied groups . This agrees with previous observations showing that BCG vaccine given at birth induces a strong Th1 response that modulates the neonatal Th2 immune environment , with the CD4+ subpopulation of T cells being the main source of IFN-γ [8] , [37] . However , M+B− infants mounted a still stronger Th1 response to BCG . This likely relates to the prenatal activation of the immune system induced by maternal T . cruzi infection described by us and others , leading to a non-specific activation of various cell types as monocytes , lymphocytes and probably other cells [24] , [38] . Such activation might result from the transplacental transfer of soluble molecules released by parasites present in the infected mother , such as the Tc52 molecule known to activate directly dendritic cells or pro-inflammatory glycophosphatidyl-inositol anchors [39]–[41] . Our study shows that chronic maternal infection with an intracellular pathogen can influence the neonatal immune environment sufficiently to enhance the Th1 regular infant response to BCG . This is also in line with other work showing , conversely , that neonates born to mothers chronically infected by helminths , known to promote Th2 immune responses , display diminished responses to PPD [11] , [12] , [42] . By contrast , M+B+ infants failed to exhibit an increased response to BCG vaccine . This is surprising , since they presented at birth , when BCG was administered , a marked type 1 immune bias ( indicated by their increased IFN-γ response to SEB ) , and harboured high amounts of circulating live parasites [43] , susceptible to release the pro-inflammatory molecules mentioned above . Different hypotheses could explain such divergence between M+B− and M+B+ infant responses to BCG . First , the treatment received by M+B+ infants might have contributed to modify their response; the similar PPD responses observed in treated and untreated infants may not rule out a treatment effect since the number of cases in each subgroup was low . The M+B+ IFN-γ−producing T cells might also have been temporary exhausted as indicated by their high level of apoptosis resulting from their intense response to parasite multiplication [5] , subsequently limiting the responses to other antigens , as also shown in other infections [44] , [45] . Another possibility relates to the higher frequency of prematurity and low birth weights in the M+B+ newborn group [22] , known to reinforce immune immaturity [46] , [47] . Furthermore , congenital T . cruzi infection might have down-regulated the cord blood T cell response , as observed in malaria [48] . Finally , it is also possible that the IFN-γ response to BCG increases earlier in M+B+ infants than in other groups , and is no more detected at 6–7 months , in agreement with the observation that there is no more difference in the response to SEB between groups at 6–7 months . Nevertheless , the immune responses of M+B+ infants to hepatitis B vaccine , and to a lesser extend against diphtheria and tetanus vaccines , given after birth , were stronger and more type 1-oriented than in both of the other groups of uninfected infants . This group produced significantly higher specific IgG levels in response to HBV vaccine . A similar trend was observed for antibody responses to diphtheria and tetanus vaccines . Enhanced antibody responses in M+B+ infants were associated with increased IFN-γ production in response to vaccine antigens . The observation that , amongst the HbsAg-specific IgG , the IgG1 , IgG2 and IgG3 subclasses , known to be associated with type 1-responses to intracellular pathogens [32]–[34] were increased still emphasizes the Th1-biased profile of the response . The mechanism of such type 1-shift of the immune response to DT , TT and HBV vaccines associated with congenital Chagas disease , and why it is not observed in M+B− infants , remains to be elucidated . Some non-exclusive factors might be considered . First , the fact that the immunological modulation observed at birth in M+B− newborns is transient [38] could imply that it could have a short term impact on immune responses to vaccines given from 2 months of age onwards . Secondly , the benznidazole treatment received by M+B+ infants might release parasitic pro-inflammatory molecules ( see above ) from lysed parasites favouring IFN-γ production [28] . However , high parasite levels were no longer present at the time of diphtheria , tetanus and HBV vaccine administration ( 2 , 4 and 6 months ) . No differences in the IgG and IFN-γ responses to DT , TT and HBV antigens were observed between treated and untreated infants , suggesting that treatment has a minimal effect on the Th1 shift observed in M+B+ infants . Third , the type of vaccine might also play a role on the different responses between M+B+ and M+B− infants . Indeed , BCG is a cellular , live attenuated vaccine , meaning that non virulent mycobacterias multiply in the host receiving the vaccine , which induces mainly a T cell response ( CD4+ Th1 and CD8+ ) , whereas the vaccines against hepatitis B , diphtheria and tetanus are acellular vaccines eliciting mainly a B cell production of IgG antibodies with the help of T cells . Our results suggest that at birth , the immune environment in uninfected newborns from T . cruzi-infected mothers ( M+B− ) favours preferentially a cellular response , while that of congenitally-infected infants ( M+B+ ) is more prone to modulate antibody responses . If the increased competence to secrete IFN-γ in both M+B− and M+B+ infants might confer better or earlier protection against infection may be discussed . Indeed , our study was limited to blood samples collected from infants at 6–7 months of age . Earlier sample collection would be useful for kinetic studies to appreciate a possible earlier protective effect . A follow-up might also permit to compare the functional implication of the protective effect of the different vaccinations between infant groups . Though the response to PPD does not correlate with protection against tuberculosis [49] , we may assume that it reflects a globally better response to BCG , a vaccine known to limit the severity of tuberculosis if acquired during early infancy , although it does not prevent the acquisition of infection [50] . For the other vaccines , the fact that M+B+ infants produce higher amounts of protective antibodies suggests that protection is probably reached earlier in life in infants suffering from congenital Chagas disease . Although this study involved a limited number of cases , our results show that both maternal infection with T . cruzi and congenital Chagas disease have no major impediment on vaccinal responses to BCG , hepatitis B , diphtheria and tetanus vaccines in the neonatal period . Nevertheless , both conditions induced a trend towards a type 1 immune polarization known to be essential in fighting intracellular pathogens [3] . In addition , our results also show that the overcoming of immunological immaturity by T cruzi infection in early life is not limited to the development of parasite-specific immune response [5] , but has also effects on responses to parasite-unrelated vaccinal antigens . This suggests adjuvant properties of some T . cruzi molecules , offering new prospects in neonatal vaccinology . | Vaccines are of crucial importance to prevent morbidity and mortality due to infectious diseases in childhood . A modulation of the fetal/neonatal immune system ( considered immature ) toward Th1 or Th2 dominance could modify responses to vaccines administered in early life . T . cruzi is the agent of Chagas' disease , in Latin America currently infecting about 2 million women at fertile ages who are susceptible to transmitting the parasite to their fetus . In previous studies we showed that T . cruzi-infected mothers can induce a pro-inflammatory environment in their uninfected neonates ( M+B− ) , whereas congenitally infected newborns ( M+B+ ) are able to develop a pro-Th1 parasite-specific T cell response . In the present study , we analysed the cellular and/or antibody responses to Bacillus Calmette Guerin ( BCG ) , hepatitis B birus ( HBV ) , diphtheria and tetanus vaccines in 6- to 7-month-old infants living in Bolivia . M+B− infants produced more IFN-γ in response to BCG , whereas M+B+ infants developed a stronger IFN-γ response to hepatitis B , diphtheria and tetanus vaccines and enhanced antibody production to HBs antigen . These results show that both maternal infection with T . cruzi and congenital Chagas disease do not interfere with responses to BCG , hepatitis B , diphtheria and tetanus vaccines in the neonatal period and that T . cruzi infection in early life tends to favour type 1 immune responses to vaccinal antigens . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"microbiology/immunity",
"to",
"infections",
"immunology/immunomodulation",
"infectious",
"diseases/neglected",
"tropical",
"diseases",
"immunology/immune",
"response",
"public",
"health",
"and",
"epidemiology/preventive",
"medicine",
"public",
"health",
"and",
"epidemiology/infe... | 2009 | Maternal Infection with Trypanosoma cruzi and Congenital Chagas Disease Induce a Trend to a Type 1 Polarization of Infant Immune Responses to Vaccines |
Cell cycle is a complex and highly supervised process that must proceed with regulatory precision to achieve successful cellular division . Despite the wide application , microarray time course experiments have several limitations in identifying cell cycle genes . We thus propose a computational model to predict human cell cycle genes based on transcription factor ( TF ) binding and regulatory motif information in their promoters . We utilize ENCODE ChIP-seq data and motif information as predictors to discriminate cell cycle against non-cell cycle genes . Our results show that both the trans- TF features and the cis- motif features are predictive of cell cycle genes , and a combination of the two types of features can further improve prediction accuracy . We apply our model to a complete list of GENCODE promoters to predict novel cell cycle driving promoters for both protein-coding genes and non-coding RNAs such as lincRNAs . We find that a similar percentage of lincRNAs are cell cycle regulated as protein-coding genes , suggesting the importance of non-coding RNAs in cell cycle division . The model we propose here provides not only a practical tool for identifying novel cell cycle genes with high accuracy , but also new insights on cell cycle regulation by TFs and cis-regulatory elements .
As one of the most important cellular processes , the cell division cycle is under precise regulation in all organisms . Mis-regulation of the cell cycle can lead to catastrophic cellular events , e . g . premature apoptosis or abnormal proliferation of cells , which are the causes of some human diseases such as cancer [1] , [2] . Cell cycle regulation has been studied intensively , with focuses mainly on two aspects . First , cell cycle regulated genes have been identified systematically using microarrays to detect periodic expression of genes in cell cycle time course data [3] , [4] . Second , the genes , particularly , the transcription factors ( TFs ) that modulate cell cycle have been investigated , e . g . identifying their genomic occupation using chromatin immunoprecipitation followed by microarray hybridization ( ChIP-chip ) or massively parallel sequencing ( ChIP-seq ) [5] , [6] . These studies have provided many insights into cell cycle regulation during normal biological processes and in cancers . Genome-wide gene expression during the cell cycle has been investigated using DNA microarrays in a wide range of organisms , including bacteria [7] , yeast [3] , [8]–[10] , mouse [11] , human [4] , [12] , [13] and Arabidopsis [14] . Microarray cell cycle time course data has been very successful at identifying a wide range of cell cycle-regulated genes . Despite its success , the microarray-based method has a few limitations . First , it is not effective for determining if a gene expressed at low levels is periodic due to low signal/noise ratios . Second , the synchronization procedure itself may change the expression pattern of some genes during the cell cycle , leading to false positive or false negative results . Third , limited by probe design , it is often difficult to distinguish expression patterns of different transcripts from the same gene . For example , for a gene with alternative promoter usage , it is possible that one isoform is cell cycle regulated while others are not . Consequently , the two isoforms may not be distinguished by a microarray based method if they share most of the exons . Moreover , previous microarray-based studies have focused on identification of cell cycle regulated protein-coding genes , while the non-coding RNAs have been largely overlooked . These issues can be overcome by measuring cell cycle gene expression using RNA-seq experiments , which unfortunately has not been performed . The cell cycle is under precise gene regulation at different levels of expression [15] , [16] . Particularly , at the transcriptional level it has been shown that a series of TFs act at different phases of the cell cycle and coordinate the sequential transcription of cell cycle genes [17]–[19] . The periodic expression pattern of cell cycle genes is encoded in cis in their promoters and can be manifested in trans by the TFs that bind to them . Namely , we would expect cell cycle genes to be bound by cell cycle regulating TFs . In this work , we raise and verify the hypothesis that cell cycle genes can be predicted by their genomic features ( the motif occurrence in their promoters ) and TF binding features ( binding affinity of TFs ) . Recently , the ChIP-seq genomic binding data for a large number of human TFs have been published . In particular , the ENCODE ( the Encyclopedia of DNA Elements ) project has published binding profiles for more than 120 human TFs in different cell lines and more ChIP-seq data are being produced [20] . Motivated by these data , we aim to construct a model that integrates microarray cell cycle expression data with ChIP-seq TF binding data to predict new cell cycle genes and to understand the function of TFs in cell cycle regulation . In this article , we present a computational method that predicts human cell cycle genes based on genomic and TF-binding features of genes . The model uses a supervised machine learning approach to integrate microarray cell cycle data , ChIP-seq TF binding data and motif information from sequence analysis ( Figure 1 ) . We first apply the model to all human RefSeq genes , which are well annotated with high accuracy . We validate the effectiveness of the model for cell cycle gene prediction by cross-validation , and we explore the relative importance of different predictors in the model . We then apply the model to the GENCODE TSS annotations , which provide a more comprehensive list of human promoters for both protein-coding genes and non-coding RNAs [21] . This systematic analysis enables us to explore the human genome to predict a full list of cell cycle-driving promoters . Our approach is effective in identifying cell cycle genes with low expression levels and is not sensitive to synchronization treatment . Since it is applied at the TSS level , it can distinguish the different isoforms of a gene regulated by alternative promoters . Furthermore , it can also be used to predict cell cycle regulated non-coding RNAs , which we believe will substantially promote our understanding of cell cycle regulation .
With the rationale that periodic expression of cell cycle genes is driven by a subset of transcription factors ( TFs ) , we first examined whether cell cycle genes and non-cell cycle genes show different binding strength by TFs . We collected the ChIP-seq data from the ENCODE ( The Encyclopedia of DNA Elements ) project [20] , which provided high-resolution binding events of more than 120 human TFs in multiple cell lines . The binding strength of a TF to the promoters of genes was calculated by a probabilistic model called TIP ( Target Identification from Profiles ) we proposed previously [22] . This model provides a significantly more accurate measure of TF binding affinity to particular genes than the peak-based method used in many studies [23] . We prepared a dataset of cell cycle genes and non-cell cycle genes in HeLa cells that have been experimentally verified with high confidence based on the meta-analysis described in Cyclebase [24] . In the 424 ENCODE ChIP-seq TF binding profiles , 46 were performed in HeLa cells , for which we calculated the regulatory scores using TIP and the average binding signals in a 2 kb DNA region centering at the TSS of all genes ( see Methods for details ) ( Figure 1 ) . Corresponding values of each measurement method were compared between cell cycle genes and non-cell cycle genes using Student's t-test . The regulatory scores for cell cycle genes are significantly contrastive ( generally higher than ) from those of non-cell cycle genes . As shown in Figure 2A , comparative analysis of the regulatory score distributions of both CMYC and E2F1 show that cell cycle genes tend to have substantially higher regulatory scores than non-cell cycle genes ( P = 2e-55 and P = 1e-50 , respectively ) . This is indicative of the significant regulatory roles CMYC and E2F1 have on the expression of cell cycle genes , thus suggesting that they are important features to be used in a cell cycle prediction model [25] , [26] . In comparison , the average TF binding signals can also discriminate cell cycle versus non-cell cycle genes , but with much lower significance levels . For example , when average signals of CMYC and E2F1 binding were calculated , we observed less significant difference in values between cell cycle and non-cell cycle genes . The P-values of average signal comparisons are 2e-8 for CMYC and 9e-7 for E2F1 ( Figure 2B ) , indicating that average signals are less effective classifiers for predicting cell cycle genes than regulatory scores . Other than CMYC and E2F1 , many other TFs also reflect significant differences in binding strengths between cell cycle and non-cell cycle genes , especially when TIP is utilized ( Figure 2C and Suppl . Table S1 ) . This suggests that the discriminatory efficacy of regulatory scoring is maintained throughout a high percentage of TFs and is not confined to a particular subset of cell cycle regulatory TFs . Thus , we will use regulatory scoring of TF to genes to predict cell cycle genes . It should be noted that cell cycle genes tend to have higher expression levels than non-cell cycle genes; some of the TF binding difference may reflect the expression level difference rather than their involvement in cell cycle ( see Discussion for details ) . We also note that the cell cycle regulatory function of a TF may not be reflected at the transcriptional level . Among the 46 TFs we investigated , only 6 showed significant periodical expression pattern in cell cycle: E2F1 , BRG1 , CJUN , RAD21 , GABPB and CTCF . The known cell cycle regulators , E2F4 and E2F6 , are not significant at the transcriptional level ( P>0 . 01 ) . The model that relates TF binding with cell cycle expression pattern , however , can be used to elucidate the function of TFs in cell cycle regulation by calculating their relative importance . Under the presumption that certain genomic features can discriminate cell cycle genes from non-cell cycle genes , we constructed Random Forest classification models to predict cell cycle genes using ENCODE ChIP-seq-derived TF-binding data and TRANSFAC-derived motif matching data as predictors . More specifically , we calculated the regulatory scores for all human RefSeq genes as described above , resulting in 424 TF binding profiles , each corresponding to a ChIP-seq dataset from the ENCODE project . These binding profiles represent binding strength of TFs to RefSeq genes in a number of different cell lines such as K562 , HESC , HeLa , etc . In addition , we also examine the existence of all TRANSFAC TF binding motifs in the promoters of RefSeq genes ( from TSS to upstream 1 kb ) , resulting in a total of 546 motif matching score profiles ( see Methods for details ) . To train the model , we used the cell cycle and non-cell cycle genes identified by microarray experiments in HeLa cells [4] . Consistently , from the 424 TF binding profiles we only included the 46 profiles from HeLa cells in our model . We examined three models for classifying cell cycle versus non-cell cycle genes using Random Forest method . In a TF model , the trans TF-binding features were used as predictors; in a Motif model , the cis motif features are used as predictors; and a TF+motif model uses a combination of all the features . The performance of these models was evaluated by 10-fold cross-validation ( see Methods for details ) . Our results suggest that both TF binding features and motif features are informative for cell cycle gene prediction , with a prediction accuracy AUC = 0 . 768 achieved by the TF model and AUC = 0 . 642 achieved by the motif model ( Figure 3A ) . This also suggests that the ChIP-seq derived TF binding features are considerably more predictive than motif features from in silico sequence analysis . Strikingly , a combination of both sets of features results in a prediction accuracy that surpasses that of both TF and Motif models , leading to an AUC = 0 . 861 by the TF+motif model . This indicates that the trans- information captured by ChIP-seq data and the cis- information provided by the motif analysis complement each other during cell cycle prediction . In a Random Forest model , the contribution of an individual feature to the overall predictive power of the model can be estimated by its relative importance , measured as the Mean Decrease in its Gini Coefficient ( MDG ) ( see Methods for details ) . Hence , we calculated the relative importance for all TF binding ( Figure 3B ) and motif features ( Figure 3C ) in the TF+motif model . Overall , TF features exhibit higher relative importance than motif features , with the best TF feature achieved by SYDH_E2F4 ( SYDH is the Lab ID ) ( Figure 3B ) and the best motif feature achieved by V$GEN_INI2_B ( Figure 3C ) . These data confirm that TF-binding regulatory scores are much better predictors than motif matching scores . The high relative importance of E2F4 is consistent with the critical roles it plays in cell cycle regulation [26] . To investigate whether the predictive accuracy of the model is predominantly determined by a few features or by many , we removed features one by one from the model and examine the change in prediction accuracy . In each step we removed the most predictive feature based on their relative importance , then recalculated the accuracy of the new model and re-estimated the relative importance of all remaining features . Our results show that many TF binding features are predictive of cell cycle genes . As shown in Figure 3D , removing the most predictive feature one by one only slowly reduce the AUC score of the model . Such a situation changes until most of the TF binding features have been removed , which leads to a sudden drop in prediction accuracy . At this point , most of the predictors remained in the model are motif features . In fact , we can achieve fairly accurate predictions by selecting a small set of predictors . For instance , when the top 10 TF binding features and the top 10 motif features with highest relative importance in the full TF+Motif model are selected as predictors , we achieve a AUC = 0 . 850 , only slightly lower than the full model ( AUC = 0 . 862 ) . Apart from the Random forest model , we also implemented other machine learning methods , including support vector machine ( SVM ) and penalized logistic regression ( PLS ) . Results from all these methods confirm the conclusions from the Random Forest model , e . g . higher predictive accuracy of TF binding features than motif features . Overall , Random Forest gives rise to the best predictive accuracy and thus in this paper we focus on this method in our analysis . Since experimentally verified cell cycle and non-cell cycle genes ( required to train the model ) were determined based on microarray experiments with HeLa cells , we restricted our analysis to HeLa cells in that we only include HeLa TF binding profiles as features in our models . In fact , in the 424 profiles from ENCODE ChIP-seq data , there are 68 from GM12878 , 94 from K562 , 37 from HESC and 55 from HEPG2 cell lines , respectively ( Suppl . Table S2 ) . We thus examined the cell line specificity of our cell cycle gene prediction model . If cell cycle regulation is cell line specific , we would expect to achieve the best prediction accuracy using HeLa TF binding profiles; and otherwise a similar accuracy throughout different cell lines . Our results exhibit highest prediction accuracy when the TF binding features from the HeLa cell line are used for predicting HeLa cell cycle genes , which is the case in both the TF+motif model ( Figure 4A ) and the TF only model ( Figure 4B ) . The TF sets with ChIP-seq profiles in distinct cell lines contains different TFs . We thereby compared the prediction accuracy of models using the 32 common TFs in HeLa and K562 as predictors . ChIP-seq data from HeLa cells achieve AUC = 0 . 756 in the TF only model and AUC = 0 . 860 in TF+Motif model , whereas ChIP-seq data from K562 cells achieves AUC = 0 . 722 and AUC = 0 . 831 , respectively . These results suggest that at least a subset of cell cycle genes is cell line specific . Furthermore , we investigate the binding strength of TFs to their target gene promoters in different cell lines . As shown in Figure 4C , the regulatory scores from E2F4 binding to the known cell cycle genes ( those used in our model as positive set ) are used as a metric to compare differential cell cycle regulation in K562 and HeLa cell lines . Although the scores calculated in HeLa and K562 cells are highly correlated , there is a small set of genes that show differential binding by E2F4 , most of which show higher regulatory scores in the HeLa cell line . In addition , when the target genes of E2F4 identified by TIP method in K562 and HeLa are compared , we find that many targets are unique to a single cell line ( Figure 4D ) . This indicates cell line specific binding of TFs to genes and as such , it is not surprising to observe cell line specificity of our cell cycle gene prediction model . Due to the periodicity of cell cycle genes , genomic features may vary in predictive power across cell cycle phases . Therefore , we examined whether genomic features are phase-specific by applying the Random Forest classifier model to categorize genes expressed in each phase . To generate training data , known cell cycle genes were partitioned into G1/S , G2 , G2/M , M , and S categories based on the annotation in Cyclebase [24] . Model accuracy was assessed via 10-fold cross-validation to yield an ROC curve for each cell cycle phase ( Figure 5A ) . Phase-specific cell cycle gene classification via Random Forest proved to be robust as shown by relatively high AUC scores for each phase ( Figure 5A ) . AUC scores of 0 . 858 , 0 . 793 , 0 . 864 , 0 . 859 , and 0 . 858 were obtained for G1/S , G2 , G2/M , M/G1 , and S cell cycle phases , respectively . The normalized relative importance of each genomic feature was calculated to deduce its predictive differentiability in each cell cycle phase ( see Methods for details ) . In all phases , TF features show significantly higher relative importance than motif features . Out of all TF features measured through all cell cycle phases , E2F4 is predominantly the most important predictor in G2/M , G2 , S , G1/S phases . However , in the M/G1 phase the prediction accuracy is driven by multiple TF features; interestingly E2F4 still has high relative importance but is not the most predictive feature any more ( Figure 5B ) . In line with these results , we observed that E2F4 targets were enriched in cell cycles genes with peak expression around G2/M and G1/S ( Suppl . Figure S2 ) . We note that the ChIP-seq data were performed in unsynchronized cells and reflect TF binding status in a mixed population of cells . We would expect an improvement of phase specific cell cycle gene prediction if phase specific TF binding features were available and utilized as predictors . Having shown the effectiveness of our model in predicting cell cycle genes using cross-validation , we applied it to identify new RefSeq genes that are potentially cell cycle regulated . The model was trained and then utilized to predict the cell cycle regulation of a total of 17 , 023 unclassified RefSeq genes ( gene dataset used in model training were excluded ) . Each gene was assigned a probability indicating the likelihood of a gene to be cell cycle regulated . By setting the threshold to 0 . 7 , we predicted 726 new cell cycle genes with a precision of 92% ( positive predictive value , PPV = 0 . 92 ) . Many of them are subunits of a protein complex that is known to be cell cycle regulated . For instance , Whitfield et al . measured the expression patterns of 12 centromere-associated proteins [27] in HeLa cells , among which 6 were identified as periodically expressed in the cell cycle ( CENPA , CENPF , CENPM , CENPL , CENPO , CENPQ and CENPT ) [4] . Our analysis predicts 2 additional subunits , CENPK and CENPN , to be cell cycle regulated , suggesting that the model is complementary to microarray based analysis . To further evaluate the reliability of these predicted cell cycle genes , we carried out Gene Ontology ( GO ) enrichment analysis on them . The results strongly support cell cycle related functions of these new predicted genes ( Suppl . Table S3 ) . As shown , the top enriched GO categories are all cell cycle related , such as chromosome ( GO:0005694 ) , cell cycle ( GO:0007049 ) , cell cycle process ( GO:0022402 ) , cell cycle phase ( GO:0022403 ) , M phase ( GO:0000279 ) , etc . Another method of evaluating prediction reliability is to compare them with RNAi knockdown experimental datasets . We downloaded two genome-wide RNAi knockdown datasets published by Mukherji et al . [28] and Kittler et al . [29] , in which cell cycle regulators are identified by knocking down individual genes and examining cell division defects that may result . We find that the novel cell cycle genes we predict tend to exhibit increased likelihood of cell division defect upon RNAi-induced loss-of-function perturbation . In fact , the new cell cycle genes are highly enriched in the cell cycle regulators identified by the two knock-down experiments . A total of 686 and 901 cell cycle regulating genes were identified by Mukherji et al . and Kittler et al . , respectively , among which 47 were identified by both experiments ( P = 4e-4 ) . Out of the 726 novel cell cycle genes we predicted , 50 and 55 were reported to be cell cycle regulating genes by Mukherji et al . ( P = 2e-4 , Fisher's exact test ) and by Kittler et al . ( P = 5e-3 , Fisher's exact test ) ( Figure S1 ) . Moreover , we examined the interaction partners of known cell cycle genes , the predicted cell cycle genes , and the predicted non-cell cycle genes . We expect that cell cycle genes are more likely to interact with one another and will therefore have more cell cycle partners in the protein-protein interaction ( PPI ) network . As shown , the known cell cycle genes interact with more partners than other genes ( Figure 6A ) , presumably due to the fact that they are more intensively studied in their interactions , e . g . by yeast two hybrid experiments . Moreover , the known cell cycle genes tend to have more cell cycle partners in terms of both number ( Figure 6B ) and percentages ( Figure 6C ) . We note that after excluding these known cell cycle genes , the remaining genes used for prediction have substantially fewer partners , cell cycle partners and lower percentage of cell cycle partners . However , compared with predicted non-cell cycle genes ( Probability <0 . 3 in our model ) , the predicted cell cycle genes ( Probability >0 . 7 ) interact with significantly more and higher percentage of cell cycle partners ( Figure 6B and 6C ) , implying their functions in cell cycle control . Having shown the effectiveness of our model for predicting cell cycle genes , we then applied it to the GENCODE annotation data , which provides a complete list of human transcripts including protein-coding genes , several categories of non-coding RNAs and so on . For all these transcripts , the precise positions of their TSSs were determined and the expression level associated with each TSS was quantified by CAGE ( Cap Analysis of Gene Expression ) experiments [30] , [31] . We calculated the regulatory scores of these TSSs based on the ENCODE ChIP-seq data and their motif-matching scores for all motifs as we have done for RefSeq promoters ( see Methods for details ) . Finally , the TF+motif Random Forest model trained using the above-mentioned RefSeq cell cycle and non-cell cycle genes was applied to the GENCODE dataset . Thus , by using the regulatory scores and motif-matching scores as features , the model predicts whether a TSS is cell cycle regulated and assigns a probability score to each TSS . We predicted the probability of cell cycle regulation for all GENCODE annotated human TSSs using our model . These TSSs are associated with different genomic feature categories including protein-coding genes , microRNAs , lincRNAs , snRNAs , snoRNAs and pseudogenes . As negative controls , we also included 10 , 013 randomly selected genomic locations ( i . e . artificial TSSs ) from the genome and predict their probability to be cell cycle regulated using our model . Certainly , the number of positive predictions is determined by the threshold setting and the precision ( also called PPV , positive predictive value , the percentage of true positives in all predicted positive predictions ) at each threshold can be estimated by cross-validation in our training data ( Figure 7A ) . To have a confident set of predictions , we set up a stringent threshold ( Probability score >0 . 7 ) in the following analysis , corresponding to a PPV = 0 . 92 . At this threshold , we identify 3 , 322 protein-coding , 83 lincRNA , 6 miRNA , 8 snoRNA , 4 snRNA , 16 pseudogene , and 9 artificial TSSs that are predicted to be cell cycle regulated ( Suppl . Table S4 ) . The percentage of cell cycle regulated TSSs for each genomic feature category is shown in Figure 7B . As shown , the percentage of positive artificial TSSs is very low ( <0 . 1% ) , indicating a high precision of our predictions . Similarly , the percentage of positive pseudogene TSSs is also very low ( 1% ) , since most of them are untranscribed “junk DNA” . But compared to the randomly selected artificial TSSs , it is possible that some pseudogene TSSs are actually active and expressed in cell cycle a dependent manner . Strikingly , lincRNA and protein-coding genes show similar percentage of cell cycle regulated TSSs ( ∼3% ) ( Figure 7B ) , indicating that lincRNAs might also be important in cell cycle regulation . Cell cycle regulated miRNA , snoRNA , and snRNA are identified in relatively low percentages , possibly due to low quality of annotation in their TSSs . For instance , annotation of miRNAs usually begin at the +1 start site of the corresponding pre-miRNA ( ∼110 bp ) as opposed to the genuine TSS of pri-miRNA . GO enrichment analysis was performed on the predicted cell cycle regulated TSSs associated with GO terms , most of which are for protein-coding genes . The results suggest that these positive predictions are highly enriched in gene categories involved in or related to cell cycle functions ( Suppl . Table S5 ) . Almost all of the top enriched GO terms are cell cycle related , e . g . cell cycle ( GO:0007049 ) , chromosome ( GO:0005694 ) , mitosis ( GO:0007067 ) , etc . Many genes possess multiple transcript isoforms with alternative TSSs and our model can predict the probability of each TSS to be cell cycle regulated . In fact , our results indicate that different isoforms of the same gene may be either cell cycle regulated or not cell cycle regulated , namely have distinct functions with respect to cell cycle regulation . For example , the gene DBF4 ( with Ensembl ID ENSG00000006634 ) is annotated to have 8 different TSSs by GENCODE , which forms two TSS clusters . The first cluster contain 6 TSSs , which are all predicted to be cell cycle regulated with a probability score >0 . 7; whereas the second cluster ( 11 kb away from the first cluster ) contains 2 TSSs with probability score of 0 . 296 and 0 . 190 respectively . The DBF4 protein is known to be essential for initiation of DNA replication [32] and the transcription of its promoter is activated through cell-cycle box ( MCB ) transcription elements [33] . Assuming the TSS annotation is correct , our analysis imply that only the first cluster of transcript isoforms are regulated in a cell cycle dependent manner; and that the two isoforms in the second cluster may not be periodically expressed during the cell cycle , either not being involved in cell cycle regulation or impacting cell cycle in a different way from the first cluster of isoforms .
Compared to the average binding signals of TFs in promoters , the regulatory scores we define are more informative for predicting cell cycle genes ( Figure 1 ) . Regulatory scores can be regarded as weighted average binding signals of TFs around the TSS of genes . For each TF , a specific weight is assigned to each nucleotide position in the 10 kb DNA region centering at TSS based on the characteristic binding profile of the TF . Thus regulatory scores can more accurately capture the regulatory potential of a TF to genes than average signals . When utilized as predictors for classifying cell cycle versus non-cell cycle genes , they generally reveal greater differentiability between the two gene classes , suggesting they are more powerful classifiers . In fact , the Random Forest model that utilizes average signals for the same set of TFs as predictors achieves a classification accuracy AUC = 0 . 683 , which is similar to the accuracy of the motif only model ( AUC = 0 . 642 ) , and is significantly lower than the TF only model that is based on regulatory scores ( AUC = 0 . 768 ) . Thus , it seems that by combining with machine-learning methods , the regulatory scores calculated from ChIP-seq data might also be promising in other applications , for example , predicting tissue specificity or conditionally expressed genes . Our analysis indicates that cell cycle regulation in different cell lines may not be exactly the same but shows certain cell line specificity: cell cycle genes identified in Hela cells can be best predicted by ChIP-seq TF binding profiles from the same cell line . Although the binding strengths of E2F4 to HeLa cell cycle gene promoters in both HeLa and K562 cell lines are comparable , there exists an observable small subset of genes exhibiting highly differential E2F4 binding affinities; with the majority of them showing more vigorous binding in HeLa cells . ( Figure 4C ) . In fact , a large percentage of E2F4 target genes identified by ChIP-seq experiment are HeLa or K562 specific ( Figure 4C ) . Moreover , we compared the cell cycle genes identified via cDNA microarray experiment in HeLa cells by Whitfield et al . ( 588 genes ) [4] and in fibroblast cells by Iyer et al . ( 480 genes ) [12] , and discover that only 155 are cell cycle genes in both cell lines . Contrastingly , TF binding data from other cell lines also prove predictive to HeLa cell cycle genes with reasonably high accuracy , indicating somehow a similar language of cell cycle regulation between cell lines . From these observations , it seems that to some extent , cell cycle regulation is cell line specific yet there may exist a core set of genes that are cell cycle regulated across all cell lines . The relative importance of predictors in our model suggests that E2F4 is essential in cell cycle gene regulation . In addition , the TF binding profiles for E2F1 and E2F6 also show significantly differential binding strengths between cell cycle and non-cell cycle genes , and exhibit high relative importance in our model . These results are in accordance with existing literature , which assert that the E2F family of transcription factors plays an inextricable role in driving and regulating cell cycle [34] . E2Fs are regulated by the pRB-family and pRB-related proteins ( e . g . p130 and p107 ) , that are inactivated upon CDK-mediated phosphorylation [35] . Additionally , E2F exhibit dual properties in that E2F1–E2F3 act as activators and E2F4–E2F8 as repressors [36] . In particular , E2F4 is shown by ChIP-chip analysis to have a plethora of gene targets involved in every phase of the cell cycle [26] , [34] , [37] . Principal targets of E2F4 include genes involved in cell cycle regulation , DNA replication , DNA repair , chromatin remodeling , and cell cycle checkpoints [26] . Evidently , these cellular processes are all associated with cell cycle genes thereby forming an integrated network of gene regulation [26] . Because E2F4 is a negative regulator , it must be constantly repressed by pRb and only expressed intermittently to allow the cell cycle to progress . This allows cellular processes controlled by E2F4 to occur in a phase-specific fashion ( i . e . DNA repair during S/G2 , DNA replication during S , and chromosome remodeling during G2/M ) [26] . Additionally , a comparative genomics study carried out by Linhart et al . proposes that there is substantial decrease in E2F4 binding during M/G1 phase of the cell cycle [38] . This is in accordance with our results which show a decrease in normalized relative importance of E2F4 during the M/G1 phase ( Figure 5B ) . Overall , these results suggest that E2F4 is repressed upon termination of mitosis and subsequently de-repressed upon initiation of G1 in daughter cells . The fact that E2F4 binding is an effective discriminatory cell cycle-associated TF binding feature demonstrates that our prediction model is indeed capable of utilizing key inherent cellular predictors to classify a wide variety of genes . Previous studies have shown that TF binding signals are predictive to the expression level of genes , accounting for >60% variation of gene expression [39]–[42] . Here we show that TF binding data can be used to predict cell cycle genes . The predictive power of regulatory scores which capture the trans- information of genes , can be further improved by the cis- information of these genes , or the motif matching scores in their promoters . Our model which uses TF+motif predictors achieves a classification accuracy of AUC = 0 . 861 , suggesting the regulatory code for cell cycle genes is largely harbored in their promoter regions . The TF binding data and the motif information complement each other , because ( 1 ) none of the two data are complete ( e . g . the ChIP-seq data of many critical cell cycle regulatory TFs are not available ) and ( 2 ) the trans- TF binding data from ChIP-seq captures regulatory information not only at the transcriptional level but also at the epigenomic level , since TF binding is significantly affected by epigenomic modifications ( e . g . histone modifications and DNA methylation ) . The periodical expression pattern of cell cycle genes is also regulated at the post-transcriptional level , e . g . by miRNAs , and we believe that predictive accuracy of our model can be further improved when such information are included . One caveat of the model is that ChIP-seq experiment captures TF binding in a population of unsynchronized cells , which limits our model from more precisely elucidating the cell line specific and phase-specific regulation of TFs . As we have described , microarray-based methods are less effective in identifying cell cycle genes expressed at low levels . For this reason , cell cycle genes detected from microarray experiment tend to have higher expression levels compared to those of non-cell cycle genes . In fact , when we statistically compare the expression levels of cell cycle genes versus non-cell cycle genes in HeLa cells , we observe significant expression disparity ( P = 3e-42 , Wilcoxon rank sum test ) . Furthermore , it has been demonstrated previously that TF binding is predictive of gene expression levels [39]–[42] , which makes expression level a confounding factor to account for when classifying cell cycle versus non-cell cycle genes: the model we propose here may be restricted to prediction of high versus low gene expression rather than cell cycle genes . This also explains why most of the TFs show very differential regulatory scores between cell cycle and non-cell cycle genes ( Figure 2 ) , but only 6 of them show periodical expression pattern during cell cycle in HeLa cells . To address this confounding issue , we prepare a set of non-cell cycle genes that have similar expression levels with cell cycle genes in HeLa . When the regulatory scores are compared between these genes and cell cycle genes , fewer TF features show significant difference but the key cell cycle regulators ( e . g . E2F1 and E2F4 ) still maintain significant difference . This suggests that these key regulators do in fact , bind differentially with cell cycle versus non-cell cycle genes even after decoupling them from the influence of expression levels . If average signals of TF features are compared , none of the TFs show differential binding at the 0 . 001 significance level , again demonstrating the advantage of regulatory scores . More importantly , when these expression-matched non-cell cycle genes are used as the negative training set , we can still accurately predict cell cycle genes with AUC = 0 . 706 using the TF only model and AUC = 0 . 814 using the TF+motif model . Thus , we can conclude that the model is effective for cell cycle gene prediction when the influence of expression level is eliminated and a very conservative training set is used . Most previous cell cycle research is focused on protein-coding genes , while in-depth systematic investigation of cell cycle non-coding RNAs have not been conducted . A recent paper examined the promoters of 56 cell-cycle genes using tiling array and revealed extensive non-coding transcription near these genes [43] . This explorative study highlights the potential importance of regulation by non-coding RNA during cell cycle division . We applied our model to more than 130 , 000 human TSSs annotated by GENCODE project and systematically predicted the probability of these TSSs to act as cell cycle driving promoters . The GENCODE TSS list contains TSSs for not only protein-coding genes but also for several classes of non-coding RNAs such as miRNAs , lincRNAs , snoRNAs , and snRNAs . Our predictions suggest that there is at least equal percentage of lincRNAs that are cell cycle regulated as there are protein-coding genes . Further experimental investigation of these non-coding RNAs should provide further insight into the non-coding world of cell cycle regulation . The enormous amount of genomic data from the ENCODE project provide valuable resources for biological studies . However , how to more efficiently make use of such data to facilitate hypothesis driven studies is still an open question . Here we show an example that combines large-scale ChIP-seq data from ENCODE with motif data from genome sequence analysis and cell cycle microarray data from small-scale laboratory studies . The framework introduced in this paper may also be applied to address other biological questions such as identifying tissue specific expression of genes , gene classes , and environment-induced gene expression and so on .
In this work , we used a supervised model to predict human cell cycle genes . To train the model , we obtained the known cell cycle genes and non-cell cycle genes from the data produced by Whitfield et al [4] , which measured gene expression during the cell division cycle in HeLa cells using microarray experiments . The data contain four different cell cycle time course series , each providing a list of cell cycle genes . To have a confident cell cycle gene list , we referred to the meta-analysis performed by Cyclebase [24] , which combined the results of all these four time courses . The cell cycle genes ( positive training set ) were selected as those with a significant combined P-value for periodicity ( P<0 . 011 ) , while the non-cell cycle genes ( negative training set ) were selected as those that were not significant in any of the four time courses ( P>0 . 1 ) . In total , we obtained 853 cell cycle Refseq genes and 1051 non-cell cycle Refseq genes . The phase-specificity of cell cycle genes were determined based on their peak expression time provide by Cyclebase . In Cyclebase , each cell cycle genes is assigned a value of 0–100 indicating their peak expression time with G1 ( 0–47 ) , S ( 47–70 ) , G2 ( 70–90 ) and M ( 90–100 ) . Accordingly , we selected 138 M/G1 ( 95–100 or 0–20 ) , 257 G2/M ( 80–95 ) , 253 G2 ( 70–90 ) , 175 S ( 47–70 ) and 185 G1/S ( 20–60 ) specific Refseq genes for model training . We calculated the binding affinity of transcription factors to the promoter of a gene based on their corresponding ChIP-seq data . The ChIP-seq data provides the binding signal of a TF at each nucleotide of the genome . We utilized the method called TIP ( Target Identification from Profile ) to quantify the regulatory relationships between TFs and target genes [22] . Given the ChIP-seq data of a TF , TIP builds a characteristic , averaged profile of binding around the TSS of all genes and then uses this to weight the sites associated with a given gene , providing a ‘regulatory’ score of this for each gene . From the ENCODE project [20] , we downloaded a total of 424 ChIP-seq data , representing the binding data for about 120 different TFs in more than 10 cell lines such as HelaS3 , HESC , K562 , etc . For each of them , we calculated the regulatory scores for all RefSeq genes , giving rise to a matrix of 34 , 299 ( RefSeq genes ) rows and 424 columns ( ChIP-seq datasets ) . The average binding signals of a TF with a gene is calculated by averaging the ChIP-seq signal of all nucleotide position in the promoter DNA region ( a 2 kb DNA region centering at the TSS ) of the gene . We downloaded 565 vertebrate motifs from the TRANSFAC database [44] , which represent the potential binding sites of DNA binding proteins , mostly transcription factors . We also downloaded the promoter sequences ( from TSS to upstream 1000 bp of a gene ) of 34 , 229 human RefSeq genes from the UCSC Genome Browser [45] . For each promoter sequence , we used the MATCH program [46] to examine the presence of these TF binding motifs . The pre-calculated cut-offs provided by MATCH were used to minimize the false positive rate . The MATCH program provides all the potential binding sites and their matching-scores of all of the RefSeq gene promoters . Based on these outputs , we constructed a binding score matrix [B_i , j] of size N×M , in which each row representing a RefSeq gene ( N = 34 , 229 ) and each column corresponding to a motif ( M = 565 ) . Each element B_ij was calculated by aggregating the matching-scores of all the binding sites of the motif j in the promoter of the gene i . The score was set to 0 if there is no binding site in the promoter of a gene . The Random Forest ensemble classifier was used to as a machine-learning model to predict genes as cell cycle or non-cell cycle . A prepared dataset containing known cell cycle genes ( annotation derived from RefSeq and Cyclebase ) and their associated TF features derived from ENCODE and TRANSFAC databases was used to train the model . This dataset contained 863 known cell cycle genes and 1051 known non-cell cycle genes . To generate the final training dataset , 81 TFs were chosen as pre-selected features resulting in a total of 69 , 903 cell cycle TF-gene pairs and 85 , 131 non-cell cycle gene pairs . Each cell cycle gene was assigned a positive binary value ( 1 ) and each non-cell cycle gene was assigned a negative binary value ( 0 ) . Model accuracy was assessed using 10-fold cross-validation in the following procedure . First , each fold was carried out by randomly dividing the training dataset into 10 partitions , irrespective of binary assignment . Second , 9 partitions were used to train the Random Forest model and the remaining partition was used as a test set to determine model performance . This is repeated 9 more times to yield the averaged sensitivity ( [#True Positives]/[#True Positives+#False Negatives] ) and specificity ( [#True Negatives]/[#True Negatives+#False Positives] ) of the model . This allows construction of a Receiver Operating Characteristic ( ROC ) curve , which is a direct representation of the relationship between sensitivity and specificity . The area under the ROC curve ( AUC ) is calculated via Riemann summation of 100 trapezoidal partitions . Calculation of the AUC is the main evaluator of Random Forest model accuracy in this study; AUC = 1 corresponds to 100% model accuracy and AUC = 0 . 5 corresponds to random classification by the model , thus completely non-discriminatory . The relative importance ( RI ) of a predictor in a Random Forest model can be measured by the metric “%IncMSE” ( increase of mean squared error ) [47] . Given a trained model , “%IncMSE” measures the increase of prediction error in the test data when the values for each individual predictor are permuted . The permutation of an important predictor will be expected to lead to a considerable prediction error increase and therefore a large “%IncMSE” . The relative importance of each genomic feature is model specific . The relative importance of features for predicting phase-specific cell cycle genes is calculated respectively from the corresponding phase-specific model . The R package “randomForest” was utilized to implement these models . Transcription start sites ( TSS ) information was derived from the GENCODE gene annotation project [30] and the high confidence TSS sets from GENCODE version 7 was used . This set includes a total of 137 , 874 TSSs for different gene categories , including protein coding genes ( 100 , 417 ) , miRNAs ( 1 , 755 ) , lincRNAs ( 2 , 751 ) , pseudogenes ( 13 , 164 ) , etc . In the dataset , many genes are associated with multiple TSSs which corresponds to these genes having alternative promoters . As we have done for the RefSeq genes , we calculated the TF regulatory scores and the motif matching scores for all these TSSs using the above-described methods . We trained Random Forest using the RefSeq gene training data described in the preceding section , and then apply the model to predict the probability of these TSSs function as cell cycle driving promoters . Meanwhile , we generated ∼10 , 000 random TSSs that are evenly distributed in the genome and fed them into the model as controls . Given a cut-off , the false discovery rate ( FDR ) of our model can be estimated by calculating the ratio of F_rand to F_real , where F_rand and F_real are the fractions of predicted cell cycle driving random TSSs and real TSSs ( i . e . TSSs with probability above the cut-off ) . We investigated the distribution of transcription factor target genes in the cell cycle . First , we sorted the cell cycle genes in HeLa cells according to their peak expression times . Then we examined the enrichment of the target genes of a given transcription factor in each sliding window of the cell cycle . We used a window size of 30 degrees with 10 degrees overlapping between neighboring windows . We used the Fisher's exact test to determine the significance of enrichment of target genes for a transcription factor in each cell cycle window ( Suppl . Figure S2 ) . Systematic gene knockdown data for cell division genes screening are available from Mukherji et al . [28] and Kittler et al . [29] . In the two studies , the majority of human protein-coding genes were knocked down in U2OS and HeLa cells , respectively , to identify cell cycle regulating genes . We examined and calculated the significance the enrichment of our predicted cell cycle genes in gene sets identified by Mukherji et al . and Kittler et al . using Fisher's Exact test ( Suppl . Figure S1 ) . To examine the enrichment of genes of different gene ontology ( GO ) categories in our predicted cell cycle gene set , we performed GO enrichment analysis by using the web-based tool from DAVID database ( the Database for Annotation , Visualization and Integrated Discovery ) , which calculated significance of enrichment based on Fisher's exact test [48] . In the analysis , we removed the cell cycle and non-cell cycle genes in the training set to avoid their impact and limit bias The human protein-protein interaction data is downloaded from the Human Protein Reference Database ( HPRD , Release 8 ) [49] . Human RefSeq gene annotations are obtained from the UCSC Genome Browser database [45] . | Cell cycle is a complex and highly supervised process that must proceed with regulatory precision to achieve successful cellular division . Microarray time course experiments have been successfully used to identify cell cycle regulated genes but with several limitations , e . g . less effective in identifying genes with low expression . We propose a computational approach to predict cell cycle genes based on TF binding data and motif information in their promoters . Specifically , we take advantage of ChIP-seq TF binding data generated by the ENCODE project and the TF binding motif information available from public databases . These data were processed and utilized as predictor for predicting cell cycle genes using the Random Forest method . Our results show that both the trans- TF features and the cis- motif features are predictive to cell cycle genes , and a combination of the two types features can further improve prediction accuracy . We apply our model to a complete list of GENCODE promoters to predict novel cell cycle driving promoters for both protein-coding genes and non-coding RNAs such as lincRNAs . We find that a similar percentage of lincRNAs are cell cycle regulated as protein-coding genes , suggesting the importance of non-coding RNAs in cell cycle division . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"genomics",
"biology",
"computational",
"biology"
] | 2013 | Transcription Factor Binding Profiles Reveal Cyclic Expression of Human Protein-coding Genes and Non-coding RNAs |
As Bangladesh , India and Nepal progress towards visceral leishmaniasis ( VL ) elimination , it is important to understand the role of asymptomatic Leishmania infection ( ALI ) , VL treatment relapse and post kala-azar dermal leishmaniasis ( PKDL ) in transmission . We reviewed evidence systematically on ALI , relapse and PKDL . We searched multiple databases to include studies on burden , risk factors , biomarkers , natural history , and infectiveness of ALI , PKDL and relapse . After screening 292 papers , 98 were included covering the years 1942 through 2016 . ALI , PKDL and relapse studies lacked a reference standard and appropriate biomarker . The prevalence of ALI was 4–17-fold that of VL . The risk of ALI was higher in VL case contacts . Most infections remained asymptomatic or resolved spontaneously . The proportion of ALI that progressed to VL disease within a year was 1 . 5–23% , and was higher amongst those with high antibody titres . The natural history of PKDL showed variability; 3 . 8–28 . 6% had no past history of VL treatment . The infectiveness of PKDL was 32–53% . The risk of VL relapse was higher with HIV co-infection . Modelling studies predicted a range of scenarios . One model predicted VL elimination was unlikely in the long term with early diagnosis . Another model estimated that ALI contributed to 82% of the overall transmission , VL to 10% and PKDL to 8% . Another model predicted that VL cases were the main driver for transmission . Different models predicted VL elimination if the sandfly density was reduced by 67% by killing the sandfly or by 79% by reducing their breeding sites , or with 4–6y of optimal IRS or 10y of sub-optimal IRS and only in low endemic setting . There is a need for xenodiagnostic and longitudinal studies to understand the potential of ALI and PKDL as reservoirs of infection .
The concomitance of anthroponotic transmission of visceral leishmaniasis ( VL ) , a single species of sandfly as the only known vector for transmission , the largely localized geographic endemicity of the disease , the availability of field-based diagnostic tests and highly effective drugs for treating VL , together , favour the elimination of the disease as a public health problem in the Indian subcontinent through effective surveillance , early detection and treatment , and integrated vector control strategies [1] . Furthermore , historical evidence of near-eradication of VL in the 1970s following insecticide spraying for malaria control in the 1950s and 1960s in India supports the rationale for VL elimination in the Indian subcontinent [2] . In 2005 , the Governments of Bangladesh , India and Nepal signed a memorandum of understanding to eliminate VL and set the target to reduce its annual incidence to less than 1 per 10 , 000 population ( at the upazila level in Bangladesh , block level in India and district level in Nepal ) by 2015 [3] . This political commitment was recently reinforced at a meeting of the Ministers of Health in September 2014 with the aim to make the Southeast Asia region including Bhutan and Thailand free of VL by 2017 or earlier [4] . Substantial progress has been made towards the elimination target in Bangladesh with only two out of 98 endemic upazilas reporting a incidence rate greater than 1 per 10 , 000 in 2015 ( Table A in S1 Text ) . An external assessment of the national VL control program in Nepal conducted in 2015 indicated that all the 12 previously endemic districts have achieved the elimination target ( Table B in S1 Text ) . On the other hand , despite a declining trend in the number of reported VL cases , 90 out of 456 blocks continue to report an annual incidence of more than 1 per 10 , 000 in India ( Table C in S1 Text ) . Despite substantial progress , a major challenge evident from recent outbreak investigations and surveillance data has been the increasing emergence of new ecological niches of indigenous transmission in previously non-endemic regions of Bangladesh and Nepal [5–9] . Research on VL and post Kala-azar dermal leishmaniasis ( PKDL ) has focused largely on the clinical and epidemiology aspects of the disease . A large body of research has evaluated diagnostics [10–13] , potential biomarkers for treatment response of VL and PKDL [14] , treatment options [15] , and vector control [16 , 17] . The parasite , vector species and alternative animal reservoirs for VL infection in Africa differ from that in the Indian subcontinent and research findings cannot be simply applied from one to the other [18] . Many questions remain about the natural history , the progression of asymptomatic Leishmania infection ( ALI ) to symptomatic VL disease , development of PKDL , the pathogenesis , the immune response to infection and disease [19] . Moreover , data on transmission dynamics , infectiveness and vector bionomics , role and duration of acquired immunity after infection are scarce , which limits the use of mathematical modelling to predict and inform treatment and vector control strategies for VL elimination [20] . Furthermore , the complex interactions of co-infection with HIV alters the transmission dynamics and increases the vulnerability of both infections to treatment failure and relapse and has the potential to thwart elimination efforts [21–25] . The emergence of parasite resistance to antimonials that led to a sharp increase of up to 65% treatment failure in a case series [26–28] seen in Bihar , India between 1980 and 1997 , also suggested the potential for development of resistance to miltefosine [29–31] and liposomal amphotericin B [32 , 33] . This could further alter the transmission dynamics and is a major concern for elimination efforts [34 , 35] . As countries progress towards the elimination target using current strategies of early detection and treatment of clinical disease and vector control , it is necessary to understand the consequences of under-reporting on planning and evaluating VL elimination strategies , the contribution of ALI to sustain transmission and emergence of new hotspots for infection [36] . It is equally important to understand the contribution of PKDL to transmission and its potential role as a reservoir of infection , to inform how long elimination efforts need to be continued and how they should be targeted to prevent recrudescence of new VL epidemics in the future [1] . The objective of this systematic review was to synthesize existing literature on transmission dynamics and relapse rates of VL caused by L donovani in the Indian subcontinent . In particular , the review focused on the role of ALI and PKDL as potential reservoirs for transmission so as to inform current strategies for achieving and maintaining VL elimination in the Indian subcontinent .
Data was extracted from the full text articles directly into a structured table under variables such as diagnostic used , seroprevalence , negative sero-conversion , asymptomatic to symptomatic ratio , risk factors / markers for ALI , progression to symptomatic VL disease , relapse , risk factors for relapse , infectiveness , etc . Simple proportions and risks ( hazard ratio , risk ratio , odds ratio ) as applicable and their range across different studies were used to describe the outcome variables of interest . We also compared the different mathematical models used for quantifying transmission dynamics with respect to their structures , data sources , assumptions , limitations , and predictions . We did not attempt a meta-analysis as the number of studies that focused on transmission dynamics , infectiveness were limited in the context of the Indian subcontinent . The risk of bias was ascertained using the Newcastle-Ottawa bias assessment scale for observational studies [38] and the Cochrane risk of bias assessment tool for trials [39] .
A total of 31 articles including two reviews were identified . ALI lacked a reference standard and appropriate biomarker . It was variably ascertained by a positive serology test ( rk39 ICT , rk39 ELISA or DAT ) , PCR , qPCR or LST in an otherwise healthy individual from an endemic area [41 , 42] . Only one study established the specificity of the assay on non-endemic controls [43] . Individuals with past history of VL were often but not always excluded from serological surveys . Table 1 gives the prevalence and incidence of ALI by country and by the different tests and thresholds used to ascertain infection . The majority of studies from which prevalence was estimated were not population based . Seropositivity among endemic healthy controls in diagnostic evaluation studies were also included for estimating prevalence of ALI . The seropositivity as measured by antibody response to rk39 antigen was 7 . 4% [44] . The seropositivity as measured by antibody response to the saliva antigen of the sandfly vector ranged between 43 . 5–63 . 2%; this is a proxy for human exposure to sandfly but not necessarily infection [45] . The prevalence of ALI was 34 . 8% and 3 . 8% for a parasitaemia of >0 and >1 parasite genome/mL on qPCR respectively [46] . The prevalent ALI cases outnumbered that of prevalent symptomatic VL cases by a factor of 4 . 0 in Bangladesh , 13 in Nepal and ranging from 7 . 6 to 17 in India [47 , 48] . The ratio of incident asymptomatic infection to incident clinical disease increased with decreasing incidence rates of VL ( Table J in S1 Text ) . However , more standardized and validated tests are needed to establish more accurately the prevalence of ALIs [46 , 49] . The risk of ALI was significantly higher ( OR ranging from 1 . 25–5 . 5 ) in individuals in close contact ( household member ) with a known VL case [50 , 51] or with the presence of other seropositive or recently sero-converted individuals in the household ( OR 1 . 37–2 . 22 ) [52] indicative of spatial and temporal clustering of infections [49 , 53] . Livestock ownership was associated with a lower risk for infection ( OR 0 . 4–1 . 0 ) [54] in Nepal but a significantly higher risk ( OR 1 . 16–2 . 1 ) in India [55 , 56] . In contrast , a higher cattle density in the community had a protective effect against infection in Bangladesh ( OR 0 . 97 ) and Nepal ( OR 0 . 63 ) [47 , 51] . Risk factors associated with VL such as poverty , low socioeconomic status , malnutrition , poor housing conditions , damp floors , mud walls , sleeping on the floor , sleeping outside and proximity to water bodies , also significantly increased the risk for ALI [51 , 52 , 55 , 57] . Most infections remained asymptomatic [42] . Spontaneous resolution ( sero-reverting from positive to negative status ) was seen in 33–86 . 7% of ALI within a year [41 , 48 , 62 , 66] . The proportion recovering spontaneously was lower for ALI with higher antibody titres [47] . On the other hand , the proportion of ALI that progressed to symptomatic disease within one year ranged between 1 . 5–23% [44 , 48–50 , 58 , 62 , 68] . It was higher ( 18 . 8–69 . 1% ) amongst seropositive contacts of VL compared to seropositive non-contacts ( 5–38% ) [58 , 66 , 77] . Seropositive contacts were 1 . 64–4 . 82-fold more likely to progress to clinical disease compared to seropositive non-contacts of a known VL case [47 , 49 , 62] . Anti-leishmania antibody titres were strong predictors of progression of ALI to symptomatic VL disease . Healthy but seropositive individuals with moderate antibody titres as measured by DAT were up to 5-times , and those with high titres were 8–40-times more likely to progress to symptomatic disease than those who were seronegative [44 , 46] . The risk for progression to symptomatic disease was significantly greater ( up to 9 and 111-fold for moderate and high titres respectively ) despite the small numbers amongst sero-converters [44 , 48] . Raised levels of immune cytokines and chemokines such as interferon-γ , nitric oxides , C-reactive protein and lowered levels of TNF-α , interleukin-2 and interleukin-4 were other potential markers for progression of ALI to clinical disease [47 , 81] . Parasitaemia levels were 500-fold lower in ALI than in active VL disease; individuals with a parasitaemia >5 parasite genome / mL were at higher risk of developing clinical VL [46 , 76] . Xenodiagnostic studies or artificial feeding experiments were limited in scale and number . In one study , 5 . 3% of a total of 258 laboratory-bred Ph argentipes became infected when fed on active VL patients [82] . There were no studies from the Indian subcontinent that quantified the infectiveness potential of infected asymptomatic individuals . Mathematical transmission models estimated the infectiveness of an early ( PCR+ , DAT- ) and late asymptomatic infected ( PCR+ , DAT+ ) individual to be 2 . 5 and 3 . 3% assuming that a VL patient would infect 100% of sand flies feeding on them [83 , 84] . The model further assumed that the relative infectiveness of an early asymptomatic infected ( PCR+ but seronegative ) individual was half of that of a late asymptomatic infected individual . PKDL is hypothesized to be the reservoir for the Leishmania parasite and was incriminated in the resurgence of VL in West Bengal , India in the 1990s following discontinuation of insecticide spraying [85] . A total of 35 studies on PKDL including 6 reviews were identified . Except for one longitudinal study each in Nepal and India , most studies were cross-sectional surveys or based on surveillance data . There was no standard case definition for PKDL diagnosis , though operational case definitions are available from the WHO since 2010 [86] . All rk39 positive cases , with or without a past history of VL , with skin lesions suggestive of PKDL were considered probable PKDL by most studies . Confirmed PKDL required the demonstration of parasite in the skin lesion . Table 2 summarizes the findings of burden , natural history , risk factors and infectiveness of PKDL . The prevalence of confirmed PKDL ranged between 4 . 4–4 . 8 per 10 , 000 population in Bangladesh and India [87–89] . The incidence of PKDL has been estimated to be 0 . 1 per 10 , 000 ( Table K in S1 Text ) . The development and natural history of PKDL showed wide variability . The proportion of PKDL without a preceding history of VL was between 3 . 8–28 . 6% [90–94] . The proportion of treated VL cases who developed PKDL within a year averaged 1–3% [95 , 96] . The mean period from VL treatment to development of PKDL was 1–4 years . The duration to development of PKDL did not differ by the drug used for VL treatment . The duration was slightly longer for nodular PKDL ( 34mo ) compared to macular or papular PKDL ( 22 . 8–23 . 8mo ) . Active surveillance of a population of 24 , 814 individuals in Bangladesh between 2002 and 2010 identified 98 untreated PKDL patients , 48 ( about 49% ) of whom resolved spontaneously with a median time of 19 months [91] . The younger age group was more likely ( 17 . 1% ) to develop PKDL compared to older VL cases ( 12 . 5% ) . They also developed PKDL earlier ( 27mo compared to 50mo ) [87] . Family history of VL and clustering was a significant risk factor for development of PKDL [88 , 94] . Inadequate treatment of VL with antimonials was associated with a 11 . 6-fold higher risk of developing PKDL [96] . We identified two xenodiagnostic studies done on PKDL patients from the Indian subcontinent . The proportion of sandfly getting infected after feeding on PKDL patients in an experimental setting ranged from 32–53% with the highest rate of infection of the sandfly seen on the 4th day post-feed [101 , 102] . For lack of data , transmission dynamics modelling studies assumed the infectiveness of PKDL to be either 50 or 100% in order to estimate other parameters of interest such as infectiveness of the asymptomatic stage of infection [84] . One of the model variants was structured in a way to assume that only VL and PKDL ( but not ALI ) contributed to transmission . With this and other assumptions of sandfly density , biting rate , life expectancy etc . , the infectiveness of PKDL relative to VL was estimated by this model variant to range between 2 . 32–2 . 72 [83] . A total of 20 studies relevant to relapse including eight drug trials and six cohort studies were identified from the Indian subcontinent . There was no standard case definition for relapse . Most studies required demonstration of the parasite to confirm a relapse following either a clinical cure or parasite cure at the end of treatment for VL or PKDL . We reviewed seven modelling studies that used data from the Indian subcontinent [34 , 80 , 83 , 84 , 124–127] . Transmission was modelled to quantify the levels and consequences of under-reporting , to quantify and predict the effect of different treatment or vector control strategies on VL incidence and / or prevalence and to ascertain the potential of ALI and PKDL as reservoirs of infection . All models were deterministic albeit with slightly varying compartments for the different transmission and clinical stages . The major differences in the model structure , data sources used to fit the model , assumptions , fixed and estimated parameters , scenarios simulated , main findings and limitations are summarized in Table L in S1 Text . Under-reporting can be a problem for planning and evaluation of elimination strategies . The first attempt at estimating under-reporting ratios was based on mathematical modelling that predicted a 90% under-reporting rate in 5 and 8 of the 21 endemic districts in Bihar , India in 2003 and 2005 respectively [124] . As a result , 3–5 districts were misclassified as high or low risk . Furthermore , the model predicted that the population density , health infrastructure , literacy level of the district had no effect on the extent of under-reporting which was sensitive to changes in VL endemicity levels . Community-based surveys reported an actual under-reporting of VL cases by a factor of 8 . 13 in 2003 [128] and 4 . 17 in 2005 in Bihar , India [129] . More than 85% of VL patients sought treatment from the public sector and was consistent with a downward trend in under-reporting , which was largely attributed to the free availability of VL drugs in government facilities under the elimination program [35] . Any attempt at interpreting the current reported disease trends should take into account this drastic change in underreporting ratio . Assuming that clinical cases were responsible for the bulk of transmission , country-specific empirical data on health seeking behaviour was used to parameterize a transmission dynamics model to predict the effect of very early diagnosis ( when non-specific symptoms such as fever appeared before the classical signs and symptoms of VL ) and to characterize the profile of a potential diagnostic product [126] . Patients in Nepal , typically first presented with VL symptoms to the health system and had a shorter duration of onset of symptoms to diagnosis and treatment , whereas in India , patients sought care earlier at the stage of non-specific symptoms , and experienced delayed diagnosis and remained untreated for a longer duration [130] . The study shows the importance of earlier diagnosis and prompt therapy in VL elimination but also the risk that reduced transmission will expose the population to future epidemics , with waning herd immunity if vigilance is not maintained and diagnostic delays increase–a factor which might further delays the detection of an epidemic . The models shows that the importance of novel diagnostics that can detect the infection in asymptomatic carries before they develop full VL , where high specificity is at a prime , even if sensitivity is relatively low . The reason is that the challenge of early testing with the intention of treating is to avoid false positives , especially with decreasing prevalence . Transmission dynamics was modelled to predict the effect of treatment of VL and PKDL patients on VL elimination , simulating different scenarios of detection delays , varying duration of treatment regimens , varying rate of treatment failure and relapse [84] . The model was fitted to the KALANET data to predict a best-case treatment scenario ( early case detection , shorter duration , and more efficacious treatment ) to reduce VL prevalence but no effect on the prevalence of ALI . Such a scenario reduced PKDL incidence from 1 to 0 . 6 per 10 , 000 but had a minimal effect on VL incidence ( the benchmark for the elimination program ) . The model predicted that transmission was predominantly driven by asymptomatic infected individuals and early case detection and treatment had no substantive effect on overall transmission . A variation of this model which tested the assumption of PKDL ( as opposed to ALI infection ) as the reservoir of infection predicted that ALI contributed to 82% of the overall transmission compared to 10% by symptomatic VL and 8% by PKDL patients [83] . Transmission dynamics was similarly modelled to predict the effect of different vector control strategies on VL elimination simulating different scenarios of optimal and sub-optimal IRS under varying endemicity levels and different assumptions of infection reservoirs . The transmission of the parasite between the host and sandfly was dependent on the infectiveness of the host or of the sandfly with a single bite , the mean biting rate and the sandfly density . The biting rate was assumed to be 0 . 25 / day ( inverse of the feeding interval assumed to be 4 days [131] ) and the latency period in the sandfly was assumed to be 5 days [132] . The elimination target could be achieved if the sandfly density was reduced from 5 . 27 to 1 per human , the life expectancy of the sandfly halved from 14 to 7 days , or the interval between blood feeds for the female sandfly doubled from 4 to 8 days or by a combination of any of these [84] . Entomology surveys estimated a prevalence of infected sandfly to range from 4 . 9–12% [64 , 71 , 133 , 134] . Table 4 gives details of sandfly abundance , distribution and feeding behaviour and risk factors affecting sandfly density . The effect of vector control on reducing PKDL prevalence would be delayed due to the latency between recovering from VL and developing PKDL . The same model estimated the basic reproduction number ( R0 ) as 4 . 71 ( 95% CI: 4 . 1–5 . 4 ) . The effective reproduction number ( Re ) was reduced non-linearly by IRS and LLINs and linearly by environmental management for vector control [125] . The model predicted that VL would be eliminated if the sandfly density was reduced by 67% ( 95% CI: 60–72% ) by killing the sandfly with IRS or LLINs or if the sandfly density was reduced by 79% ( 95% CI: 75–82% ) by reducing their breeding sites with environmental management for vector control or by a combination of these . Compared with these model predictions , the actual reduction in sandfly density of 24 . 9–43 . 7% with LLIN [135 , 136] and 42% with environmental management for vector control [136] seen in intervention trials ( Table 4 ) , would not be sufficient to reduce transmission to achieve the VL elimination target . However , the sandfly reduction of 72 . 4% seen in the intervention trial [136] with IRS would be adequate to reduce the transmission level ( Re<1 ) to achieve elimination . A more recent model designed to test different scenarios of optimal and sub-optimal application of IRS ( sandfly density reduced by 63% and 31 . 5% respectively ) in varying endemicity settings predicted that optimal use of IRS reduced the VL incidence by 25% and 50% at 1y and 2y respectively at all endemicity levels [83] . VL incidence continued to decline as the burden of ALI became less . However , the decline in VL incidence was slower if PKDL ( not ALI ) was assumed to be the main reservoir of infection . The model predicted VL elimination with 4 – 6y of optimal IRS or 10y of sub-optimal IRS and only in low endemic ( VL incidence < 5 / 10 , 000 ) setting whereas VL was not eliminated even with 12y of optimal IRS if PKDL were assumed to be the main reservoir of infection . A longer period to development of PKDL , a longer PKDL duration increased the transmission pressure to slow down the decline in VL incidence . Model predictions of VL elimination by IRS depended on the assumptions about the main reservoir for infection ( ALI or PKDL ) and were sensitive to other model assumptions such as the proportion of ALI progressing to symptomatic disease and the proportion of VL developing PKDL . However these predictions were robust to assumptions of infectiveness of early asymptomatic relative to that of late asymptomatic stage .
The burden of ALI is considerable . Longitudinal studies are necessary to identify biomarkers for infectiveness and for progression of ALI to symptomatic VL disease . More research is needed on the immune response to VL and PKDL to identify biomarkers for development of PKDL . Xenodiagnostic studies are necessary to quantify the infectiveness of ALI and PKDL to sandfly relative to symptomatic VL , and their contribution to overall transmission . Even though domestic animals are seen to be infected , there is no evidence of their role in anthroponotic transmission in the Indian subcontinent . Relapse rates need to be monitored for their potential to contribute to transmission and for the emergence of drug-resistant parasites in the context of HIV co-infection . Availability of better data from large well-designed longitudinal studies for modelling would contribute to a better understanding of the impact of treatment and vector control strategies and potential threats to VL elimination in the Indian subcontinent . | The role of asymptomatic Leishmania infection ( ALI ) , PKDL and VL relapse in transmission is unclear as VL elimination is achieved in the Indian subcontinent . ALI , PKDL and relapse studies lacked a reference standard and appropriate biomarker . ALI was 4–17-fold more prevalent than VL . The risk of ALI was higher in VL case contacts . Most infections remained asymptomatic or resolved spontaneously . The natural history of PKDL showed variability . Twenty nine percent had no past history of VL treatment . The risk of VL relapse was higher with HIV co-infection . Modelling studies predicted different effects . Early diagnosis was unlikely to eliminate VL in the long term . ALI was predicted to contribute to 82% of the overall transmission , VL to 10% and PKDL to 8% . Another model predicted that VL cases were the main driver for transmission . VL elimination was predicted if the sandfly density was reduced by 67% by killing the sandfly or by 79% by reducing their breeding sites , or with 4–6y of optimal IRS or 10y of sub-optimal IRS and only in low endemic setting . There is a need for more studies to fully understand the potential of ALI and PKDL as reservoirs of infection . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"medicine",
"and",
"health",
"sciences",
"tropical",
"diseases",
"vector-borne",
"diseases",
"india",
"sand",
"flies",
"geographical",
"locations",
"parasitic",
"diseases",
"pharmaceutics",
"neglected",
"tropical",
"diseases",
"infectious",
"disease",
"control",
"insect",... | 2016 | Transmission Dynamics of Visceral Leishmaniasis in the Indian Subcontinent – A Systematic Literature Review |
Single nucleotide polymorphisms ( SNPs ) are among the most common types of genetic variation in complex genetic disorders . A growing number of studies link the functional role of SNPs with the networks and pathways mediated by the disease-associated genes . For example , many non-synonymous missense SNPs ( nsSNPs ) have been found near or inside the protein-protein interaction ( PPI ) interfaces . Determining whether such nsSNP will disrupt or preserve a PPI is a challenging task to address , both experimentally and computationally . Here , we present this task as three related classification problems , and develop a new computational method , called the SNP-IN tool ( non-synonymous SNP INteraction effect predictor ) . Our method predicts the effects of nsSNPs on PPIs , given the interaction's structure . It leverages supervised and semi-supervised feature-based classifiers , including our new Random Forest self-learning protocol . The classifiers are trained based on a dataset of comprehensive mutagenesis studies for 151 PPI complexes , with experimentally determined binding affinities of the mutant and wild-type interactions . Three classification problems were considered: ( 1 ) a 2-class problem ( strengthening/weakening PPI mutations ) , ( 2 ) another 2-class problem ( mutations that disrupt/preserve a PPI ) , and ( 3 ) a 3-class classification ( detrimental/neutral/beneficial mutation effects ) . In total , 11 different supervised and semi-supervised classifiers were trained and assessed resulting in a promising performance , with the weighted f-measure ranging from 0 . 87 for Problem 1 to 0 . 70 for the most challenging Problem 3 . By integrating prediction results of the 2-class classifiers into the 3-class classifier , we further improved its performance for Problem 3 . To demonstrate the utility of SNP-IN tool , it was applied to study the nsSNP-induced rewiring of two disease-centered networks . The accurate and balanced performance of SNP-IN tool makes it readily available to study the rewiring of large-scale protein-protein interaction networks , and can be useful for functional annotation of disease-associated SNPs . SNIP-IN tool is freely accessible as a web-server at http://korkinlab . org/snpintool/ .
Being one of the most prevalent types of genetic variation in humans , single nucleotide polymorphisms ( SNPs ) occur in both coding and non-coding regions of the genome and have been associated with a number of Mendelian diseases and complex genetic disorders [1] , [2] . With the rapid advancement of DNA sequencing and genotyping technology , millions of SNPs have been determined [3] , [4] . An average gene is estimated to have several non-synonymous missense SNPs ( nsSNPs ) , each substituting an amino acid residue [5] . Nevertheless , our knowledge of SNPs that cause a disease is very limited . Understanding whether or not a mutation or a group of mutations induce changes of a molecular function is often the first step towards finding the missing link between the genetic variation and the disease . Recent studies of disease networks have linked many nsSNPs with protein-protein interactions [6] , [7] . Understanding how these mutations can rewire the interaction network mediated by proteins associated with the disease is critical in studying complex genetic disorders , such as cancer , autism , and diabetes [8]–[10] . Unfortunately , the interaction landscape determined by the genetic variants of the disease-associated genes is far from being fully reconstructed . Thus , computational methods can play an important role in modeling nsSNP-induced rewiring of a disease network . The growing interest in understanding the relationship between a genetic variation and its functional effect on a protein has lead to a number of recent in-silico methods . A group of methods introduced the idea of computational mutagenesis to study the structure-function relationship [11] , predict the changes in enzyme activity [12] , [13] , detect disease potential of a SNP [14] , and characterize other functional effects [15] . Most recently , a number of computational alanine scanning methods were developed to study protein-protein interactions ( PPIs ) and protein-peptide interactions [16] . These methods aimed at finding residues in the interaction interface that would disrupt the interaction when mutated to alanine; they did it by estimating the relative free energy change ( ΔΔG ) between the wild-type and mutant PPI complexes . Another group of methods focused on predicting the effects of general nsSNPs on protein function and distinguishing them from functionally neutral mutations [17]–[30] . Finally , several works studied the effects of disease-associated nsSNPs on protein-protein interactions by investigating the changes in binding energy using force field and electrostatic calculations [31] , [32] and understanding the structural effects caused by nsSNPs that lead to the disruption of PPI [6] , [33] . However , in spite of the tremendous progress , developing an accurate approach that predicts the effect of an nsSNP on the protein function , including protein-protein interaction , remains an open problem . The goal of this paper is to introduce a novel computational approach for the characterization of effects on PPIs caused by nsSNPs ( nsSNP-induced effects ) . The idea of our approach is to consider prediction of such effects as a classification problem . Specifically , we defined three related classification problems that differ in the available input information and the types of nsSNP-induced effects to be identified and characterized . Leveraging the machine learning methodology , we formulated each of the three problems as the supervised and semi-supervised learning tasks . The comparative assessment of the independently built classifiers using a variety of the supervised and semi-supervised methods has demonstrated feasibility of the machine learning approach in addressing each of the above problems .
Comprehensive analysis of the mutation effects on PPIs on a large scale by experiments is a difficult task . As a result , while several datasets have been used by the computational methods [34]–[36] , no golden standard currently exists . Here , we use one of the largest such datasets , SKEMPI [35] , which includes mutations on structurally-defined heterodimeric complexes that were experimentally characterized and extracted and manually curated from the literature . For each mutation , the database provides the changes in thermodynamic parameters and kinetic rate constants between the wild-type and mutant PPIs . From the initially collected set of 3 , 047 mutations occurring in 158 heterodimeric complexes , we keep 2 , 795 mutations after removing the redundancy , where the redundant mutations are defined as the same mutations obtained from different references . Finally , since in this work we focus on the effects caused by a single nsSNP , we filter out from the sets those entries that include multiple mutations , resulting in the final dataset of 2 , 079 single SNPs and 151 corresponding protein complexes ( This training dataset is available for download at SNP-IN tool website: http://korkinlab . org/snpintool ) . Next , each mutation is characterized as one of three interaction-associated types: beneficial , neutral , or detrimental . The types are assigned based on the difference , , between the binding free energies of the mutant and wild-type complexes . Specifically , we calculate , where and are the mutant and wild-type binding free energies , correspondingly . Each energy value is calculated as , where is the gas constant , is temperature , and is the known binding affinity . For our dataset , is obtained from the SKEMPI dataset at http://life . bsc . es/pid/mutation_database/datatable . html ( column 7 for the mutant and column 8 for the wild-type ) . This value can also be calculated by , where , can also be found at the same link above . The beneficial , neutral , or detrimental types of mutations are then determined by applying two previously established thresholds to [35] , [37] , [38]: Intuitively , a neutral mutation will not change the interaction's properties , whereas the beneficial mutation will significantly increase the binding affinity , and the detrimental mutation is expected to disrupt the associated PPI . Using these three mutation types , the labeled dataset for each supervised and semi-supervised classifier is formed ( see subsection Training and evaluation of supervised and semi-supervised classifiers in Methods ) . We note that these mutation types are introduced to characterize the effect on a protein-protein interaction rather than the biological function associated with the interaction . For instance , an nsSNP that has a beneficial effect on protein-protein interaction may have a detrimental functional effect by transforming a transient complex to a permanent one . Finally , the dataset of unlabeled mutations is generated for the semi-supervised learning classifiers . Specifically , for each of the 2 , 079 mutations , all other 18 possible mutations , excluding the original mutant and wild-type residues , are introduced at the same location in the corresponding complex as the original nsSNP . For these mutations , no values are available , thus they cannot be assigned a specific interaction-associated type . The final set includes 17 , 692 mutations ( mutations for which some of the software packages failed to generate the features are excluded ) . Each nsSNP in the labeled and unlabeled sets is represented as a 33-dimensional feature vector . To calculate the set of features , we first model the structure of the mutant PPI complex using FoldX [39] , [40] and using the structure of the wild-type complex as a modeling template . Next , for each nsSNP a set of features is calculated for the modeled mutant complex as well as the wild-type native structure , and the difference of these features is included into the final feature vector . Several software packages are used to generate the features ( Table 1 ) [39] , [41]–[45] . The first group consists of 22 energy terms calculated in FoldX: Total energy , Backbone Hbond , Sidechain Hbond , Van der Waals , Electrostatics , Solvation Polar , Solvation Hydrophobic , Van der Waals clashes , entropy sidechain , entropy mainchain , sloop_entropy , mloop_entropy , cis_bond , torsional clash , backbone clash , helix dipole , water bridge , disulfide , electrostatic kon , partial covalent bonds , Energy Ionisation , Entropy Complex [39] . The second group of three features includes energy terms ( OPUS-PSP terms 1–3 ) calculated in OPUS-PSP [44] . Accessible surface area of the mutant amino acid residue is computed by NACCASS [41] , as a descriptor to measure the changes on solvent accessibility during this mutation . The next feature , Interaction energy , is defined as the sum of interaction energies of the protein chain carrying the mutation against all other chains in the complex . Interaction energy for each pair of chains is also calculated in FoldX . The remaining features include three energy terms ( Goap terms 1–3 ) from software Goap [45] , Geometric score from Geometric tool [42] , energy term from Dfire2 [46] , and Decomplex energy score [43] . Two supervised and two semi-supervised approaches are implemented and compared . The supervised learning methods include Support Vector Machines ( SVM ) and Random Forrest ( RF ) classifiers , which have been consistently among the top performing methods for a number of bioinformatics tasks [47]–[49] . Random Forests have been shown to outperform other feature-based supervised learning approaches in bioinformatics and other domains [50]–[53] , although in some cases they perform worse than SVM methods [48] , [54] . The SVM approach , in addition to being among most widely used supervised learning methods in bioinformatics , lies in the core of the top performing semi-supervised learning algorithm [55] . For SVM , we assessed three popular kernels: ( i ) linear , ( ii ) polynomial kernel , , where d is degree of the polynomial , and ( iii ) radial basis function ( RBF ) , . The polynomial kernel is then selected with d = 3 as the most accurate one , as it has the highest f-measure value . SVM models are implemented using the libSVM package [56] and the RF classifier is implemented in Weka software [57] . Semi-supervised learning has been only recently introduced to the field of bioinformatics [49] , [58]–[61] . The basic idea is to rely not only on the labeled training data , but also to incorporate an additional , unlabeled , dataset ( often of a significantly larger size ) as a part of training to improve learning accuracy . We first apply semi-supervised learning by low density separation ( LDS ) [55] , which is considered one of the most accurate semi-supervised methods [62] . The LDS approach relies on clustering to guide the unlabeled dataset by combining ( i ) graph-based distances that emphasize low density regions between clusters and ( ii ) optimization of the Transductive SVM objective function [63] which places the decision boundary in low density regions using gradient descent . Specifically , a nearest-neighbor graph G = ( V , E ) is first derived for both labeled and unlabeled feature vectors . Then a modified connectivity kernel is computed , defined as follows:where p is a path of length |p| from the set Pi , j of all paths connecting two feature vectors xi and xj , and is a parameterized ρ-path distance defined between the set of all labeled vectors on one hand and set of all vectors on the other hand . The computed kernel is then used to train an SVM in the supervised part of the algorithm [55] . Based on assessment of the supervised methods ( see Leave-one-out cross validations subsection ) , the RF classifier shows superior performance over the SVM classifiers . Thus , we would like to further improve the accuracy of this approach , by developing a simple RF-based semi-supervised learning protocol that leverages self-learning heuristics [64] . First the protocol trains a supervised learning RF classifier . Next , this classifier is applied to the unlabeled dataset and assigns each unlabeled nsSNP to one of the classes . The newly labeled dataset is merged with the originally labeled datasets . Finally , the resulting labeled datasets are used to re-train the supervised RF method . We note that while several RF-based semi-supervised based methods have been recently introduced in pattern recognition and computer vision [65] , [66] , to the best of our knowledge , no RF-based semi-supervised method has been applied in a bioinformatics area . Finally , to further improve the performance on the most difficult 3-class problem , we explore whether the classifier of the 3–class problem can benefit from the other two classifiers addressing one of the 2-class problems . Specifically , for the most accurate classifier of Problem 3 ( selected based on the weighted f-measure ) , we calculate two additional features: the prediction results from the most accurate binary classifiers for Problems 1 and 2 . To obtain these features , we use each of the two binary classifiers to generate the prediction value if it is a positive prediction , or one minus prediction value if it is a negative prediction and scale the value to be from 0 to 1 . The labeled set for a supervised classifier addressing the first 2-class problem includes mutations determined as beneficial as the first class ( strengthening PPI ) and mutations determined as detrimental as the second class ( weakening PPI ) . Another labeled set corresponding to the second 2-class problem includes both beneficial and neutral mutations as the first class ( preserving PPI ) , and detrimental mutations as the second class ( disrupting PPI ) . Mutations in the final labeled set corresponding to the 3-class problem are naturally grouped into beneficial , neutral , and detrimental classes . For each semi-supervised classifier , we use the same labeled data as in the corresponding supervised classifier and the previously described unlabeled set of 17 , 692 nsSNPs ( Table 2 ) . To evaluate all supervised and semi-supervised classifiers for each of the three classification problems , three assessment protocols were implemented . The first protocol was a standard leave-one-out ( LOO ) cross-validation protocol with the goal to compare the methods and select the most accurate classifier for each problem by utilizing each of the labeled datasets for the corresponding problem in both supervised and semi-supervised cases . For each problem , the class-based recall , precision and f-measures are calculated for each class . Next , overall performance of a classifier on the classification problem is assessed by the average accuracy and weighted f-measure scores as following:where NCi , fi , and Ni are the number of correctly identified class members , standard f-measure , and total number of class members in class i , correspondingly . A classifier with the highest weighted f-measure is selected for each problem and included into the SNP-IN tool web-server . In the second protocol , we compare our top performing classifier with the only other published method for predicting the effect of nsSNPs on PPIs , BeAtMuSiC [31] . Unlike our approach , BeAtMuSiC relies on a set of statistical potentials derived from the structures of interacting proteins and does not use a supervised learning and , subsequently , a training set . Coincidentally , for the assessment of this method the authors used the same SKEMPI dataset as was used in SNP-IN tool LOO cross-validation , with a slightly different redundancy removal protocol . Thus , we compared the performances of BeAtMuSiC and SNP-IN tool on the overlapped dataset by calculating the Pearson correlation coefficient between the predicted scores and the experimental data for the latter predictor and comparing with the published score for the former method . The raw classification prediction score of the SNP-IN tool was used . We discuss the validity and potential shortcomings of this assessment protocol further in the paper . In the last protocol , we assess the performance of SNP-IN tool by applying it to the datasets of 26th Critical Assessment of PRediction of Interactions ( CAPRI ) competition [67] . CAPRI is a community-wide competition in computational tasks related to characterization of the molecular structure of protein complexes . Recently , a new type of challenge was introduced with a goal to characterize the effect of mutation on protein-protein complexes . Specifically , there were two challenge targets ( Target 55 and Target 56 ) , each target was a designed influenza inhibitor interacting with hemagglutinin ( HA ) [68] . A comprehensive set of site-directed mutagenesis experiments was done for the residues located next to or inside the interaction interface for each target complex , and the effect of each point mutation on the binding affinity was evaluated by deep sequencing of mutants before and after binding [69] . During the competition , all CAPRI participants were asked to provide a score as the prediction of each mutation's effect on inhibitor-HA interactions . The three types of effects correspond to our 3-class problem and include detrimental , neutral and beneficial mutations . The correlations between predicted scores and experimental evaluations were calculated by using the Kendall's τ rank correlation coefficient ( http://www . ebi . ac . uk/msd-srv/capri/round26/ ) . Here , we apply the CAPRI assessment protocol to predictions of the effect of each point mutation in Targets 55 and 56 obtained by the 3-class classifier from SNP-IN tool . Finally , the SNP-IN tool is applied to analyze nsSNPs in the PPI networks associated with human diseases in two case studies using the following protocol . First , the disease-associated nsSNPs and the corresponding genes are selected from dbSNP database [70] . Second , for each nsSNP , a PPI mediated by the mutated protein is identified , and its structural template is extracted from a recently published dataset by Wang et al [7] . Third , MODELLER [71] is used to build an accurate comparative model for each nsSNP-associated PPI complex . Last , SNP-In tool is used to predict nsSNP-induced loss/preservation of the PPI by characterizing the effect of that nsSNP on the PPI . The SNP-IN tool was implemented as a web-server freely available at http://korkinlab . org/snpintool/ ( Fig . 3 ) . It allows users to upload a pdb file containing the structure of the studied PPI , and provide information about the nsSNP they would like to investigate . The server will then return the effects of the nsSNP predicted by the semi-supervised RF-SL classifiers for both 2- and 3-class problems .
The importance analysis of all 33 features , carried out using InforGainAttributeEval function in Weka [72] , showed that many features ( Table 3 ) were equally important for all three classification problems . These are primarily the energy terms obtained from FoldX and OPUS . On the other hand , some features appeared to be important only for certain classification problems . For instance , Geometric score and Accessible Surface Area ( ASA ) were not important in the interaction disrupting/preserving classification problem , while the Goap energy terms were more important , compared with the other two problems . On the other hand , Electrostatics feature appeared to be more important for the 3-class problem than for the 2-class problems . Interestingly , while relative contribution of the features was different , all features without exception were informative in the vector representation: removing each of the features did not improve the prediction accuracy for any of the supervised methods . The importance analysis , thus , may be used to determine a higher priority when improving the accuracy of certain features , such as the FoldX and OPUS energy terms , which may be beneficial for all three classification problems . To assess performance of the four classifiers , we applied a LOO cross-validation protocol ( Table 4 , Table S1 ) . We started by testing the classifiers on the data for the first classification problem ( strengthening/weakening mutations ) . Interestingly , for all four classifiers , predicting a weakening mutation was significantly more accurate than predicting a strengthening one . In addition , both the SVM supervised classifier and LDS semi-supervised classifier , which relied on transductive SVM ( TSVM ) , performed worse than the RF-based supervised and RF-based semi-supervised learning methods . The top performing RF-based supervised classifier reached 0 . 87 in weighted f-measure and 0 . 89 in average accuracy . The performance gap between the SVM-based and RF-based methods became even more apparent when assessing these methods on the 3-class problem ( Problem 3 ) . Specifically , very low recall and precision when classifying the beneficial nsSNPs made the difference between the weighted f-measures of SVM-based and RF-based methods to be close to 0 . 20 for both supervised and semi-supervised approaches ( Table 4 ) . The top performing method for this classification problem was the RF-based semi-supervised approach , with the weighted f-measure value of 0 . 70 and average accuracy of 0 . 72 . Based on the superior performance of the supervised and semi-supervised RF-based methods for the first 2-class and 3-class problems , we focused on evaluating only those two methods for the second 2-class problem ( disruptive/preserving PPI mutations ) . We found that unlike the previous two classification problems , the performance of both methods on the two classes of this problem was more even ( Table 4 ) . Interestingly , the top performing RF-based semi-supervised approach for this problem ( weighted f-measure is 0 . 78 and average accuracy is also 0 . 78 ) gained ∼0 . 04 in weighted f-measure , compared to the supervised approach . This was not observed in the other two classification problems where the difference between the RF-based supervised and semi-supervised classifiers was at most 0 . 02 . The results of cross-validation allowed us to select the top performing method for each problem , using weighted f-measure ( Table 4 ) . The top classifiers for the more generally applicable second and third classification problems were then integrated into the SNP-IN tool . The overall weighted prediction accuracies ( 0 . 72–0 . 89 ) and f-measures ( 0 . 70–0 . 87 ) , as estimated by the LOO cross-validation protocol , suggest that each of the three problems is feasible when applying a machine learning approach . In addition , we observed that the performance of the classifiers on individual classes varies even in the case of the most accurate methods . To account for that in our evaluation , we calculated the Mathews correlation coefficient ( MCC ) score for the top-performing RF approaches ( Table S1 ) . The overall performance of the methods according to the MCC score was consistent with the performance evaluated based on the weighted f-measure . While the thresholds for employed here are widely used by the community [35] , [37] , [38] , other more conservative definitions for the beneficial/neutral/detrimental mutations exist . For instance , Bogan and Thorn [73] used a threshold of 2 . 0 kcal/mol to identify the residues that contributed to the interaction hot spots . We analyzed and compared the behavior of our top performing supervised and semi-supervised methods by defining beneficial , neutral , and detrimental effects using the more conservative thresholds of ±2 . 0 kcal/mol instead of ±0 . 5 kcal/mol , followed by retraining and evaluation of the methods for each problem ( Table S2 ) . Using the more conservative definition resulted in significantly unbalanced datasets ( beneficial: 48 , neutral: 1388 , detrimental: 518 ) , but the performance of the classifiers was similar , showing that our approach is adaptive to other definitions of interaction effects . Lastly , by including the performance of the two 2-class classifiers as additional two features we were able to get a striking improvement of the most accurate RF self- learning classifier for the 3-class problem ( Table 4 , last row ) . Most significantly , we obtained 82% gain in the recall of classifying beneficial mutations ( from 0 . 22 to 0 . 40 ) , and 25% gain of the MCC score ( from 0 . 49 to 0 . 61 ) . Thus , integrating the intrinsic relationship between classification problems allowed us to significantly improve predictions for the most difficult 3-class problem . We note that there may be other , simpler , 2-level protocols where each of the three classes can be eliminated consecutively ( e . g . , classifying the detrimental nsSNPs vs . the rest at the first level , and classifying the neutral nsSNPs vs . beneficial ones at the second level ) . However , our protocol is less restrictive , since it does not make a classification decision for all three classes until the last level , where the performances of both 2-class classifiers are considered simply as additional numerical features and may or may not influence the final classification . We next compared the performance of our top performing RF-based semi-supervised classifier to BeAtMuSiC , a recently published and the only publicly available tool , to the best of our knowledge [31] . The authors of BeAtMuSiC assessed their method by applying it to the SKEMPI set . Out of 3 , 047 entries in SKEMPI , they removed the redundant entries and entries with multiple mutations . The resulting set of 2 , 007 was used to calculate the predicted values and compare them with the original experimental measurements . Following our preprocessing protocol , we also removed redundant entries and entries with multiple mutations and then successfully predicted 1 , 954 mutations . Finally , comparing our set with the set of 2 , 007 entries used in BeAtMuSiC , we determined 1 , 897 entries shared between the two sets that we used for our comparative assessment . We note that BeAtMuSiC is not a classifier , as it predicts the changes in binding affinity caused by an nsSNP . Therefore , instead of direct classification results , we used the classifier-calculated probability for an nsSNP to be of the preserving type; we expected this probability to correlate well with changes in the binding affinity . We also note that our RF-based classifier and all other classifiers were trained using the SKEMPI set . Therefore , for this comparative assessment we applied a LOO cross-validation protocol to train models and used predictions on the test examples from the same protocol to calculate the Pearson correlation coefficient [31] . As a result , the computed Pearson correlation coefficient between our prediction scores and experimental values from SKEMPI was 0 . 57 , while the authors of BeAtMuSiC reported the correlation coefficient of 0 . 47 . As a final evaluation of our method , we applied the semi-supervised RF-SL classifier of SNP-IN tool to characterize all mutations of both CAPRI Targets , 55 and 56 , and then scaled the probability of each classification to obtain the score of mutation effects on binding . Comparing to other participation groups in 26th round of CAPRI [74] and BeAtMuSiC applied for the same purpose [31] , our RF-SL classifier from SNP-IN tool obtained a Kendall's tau coefficient with experimental results of 0 . 37 on target 55 and 0 . 25 on target 56 . Both results were significantly better than those ones by either a CAPRI predictor or BeAtMuSiC ( Table 5 ) . The validation on the targets of the 26th round of CAPRI demonstrates that our semi-supervised RF-SL classifier is currently the best predictor of the mutation effects on PPIs . The accuracy and computational performance of our approach allowed us to study the mutation-induced rewiring effects of protein-protein interaction networks mediated by disease genes . The rationale of this approach was as follows . All nsSNPs on the surface of a protein could be roughly organized in two groups with respect to their role in a PPI mediated by this protein . The first group included nsSNPs that were located inside the interaction interface , while the second group consisted of nsSNPs that are located outside interface ( but might nevertheless rewire the PPI ) . To demonstrate the applicability of our approach , we used it to study two disease PPI networks centered around the genes critically implicated in two complex genetic diseases , breast cancer and diabetes ( Fig . 4 ) . For each study , we used dbSNP [70] and a recently published INstruct database [75] to ( 1 ) select the disease-associated genes that form a PPI network , ( 2 ) select nsSNPs associated with the disease , and ( 3 ) determine whether any interactions from that network have homologous structural templates . To ensure the accuracy of the PPI data we used HINT database [76] that includes PPIs experimentally supported by one or more publications . We required for each PPI to be supported by at least two references . For each PPI with a known structural template we obtained a homology model ( see Feature representation subsection in Methods ) , mapped known nsSNPs onto the modeled structure of the PPI and grouped them into the two groups discussed above . Finally , we run SNP-IN tool on each structurally resolved PPI and compared the obtained results with the known literature on the effects of those variants .
In this work , we developed a new approach , SNP-IN tool , that characterizes the effects of nsSNPs on protein-protein interactions . We introduced three related nsSNP effect classification problems and applied supervised and semi-supervised machine learning methods leveraging SVM and RF formalisms . The performance assessment of the classifiers allowed us to draw several conclusions regarding the nature of the studied problem and the machine learning methodology addressing it . First , we found that while many of the same nsSNP features play equally important role in all three classification problems , some problems appeared to be more challenging than the others . Second , we concluded that the random forest approach is better suited for this problem than the SVM approach: both RF-based supervised and semi-supervised methods significantly outperformed the corresponding SVM-based methods . Finally , we observed that the semi-supervised learning method did not always significantly outperform the supervised method . The comparative assessment showed the superior performance of SNP-IN tool on the CAPRI targets as well as over the only other published method , BeAtMuSiC . We note , however , that the latter comparison should be treated with caution , as it was done over the SKEMPI dataset that was used in LOO for SNP-IN tool . In contrast , BeAtMuSiC is not a machine learning approach , so it used this dataset exclusively for its assessment . Thus , while none of the assessed examples from SKEMPI were simultaneously used in training ( due to design of LOO cross-validation protocol ) and could not influence the classifiers , further more detailed assessment between these two methods must be done , when another large dataset is available . Semi-supervised learning approaches have received growing attention from the bioinformatics community with their successful applications to several areas of bioinformatics and computational biology [47]–[49] . To the best of our knowledge , none of the currently existing semi-supervised approaches in bioinformatics have utilized random forest classifiers . Our simple RF-based semi-supervised classifier performed remarkably better than state-of-the-art transductive SVM and LDS based semi-supervised classifiers , suggesting that this could be a promising direction for addressing the biological classification problems that involve vector-based representations of highly heterogeneous features . Overall , limitation of the labeled data due to the difficulty of obtaining experimental binding affinities from the site-directed mutagenesis experiments renders semi-supervised approaches a powerful alternative to the supervised methods . A related issue is predicting the effect of a non-synonymous SNP on a function carried by a protein product of the mutant gene , and specifically on a PPI mediated by this protein , has emerged as an important computational challenge . A problem of labeling nsSNPs as detrimental , neutral or beneficial , has been recently introduced for the first time at the 26th round of the CAPRI competition [88] . Considering the 3-class problem as the most comprehensive annotation for nsSNP effects on PPI , we have also introduced two other problems , each involving only 2 classes . While related , the problems are designed to characterize the genetic variation from different perspectives . One two-class problem , where an nsSNP is characterized as disrupting or preserving the associated PPI could be used to study the network rewiring caused by certain mutations , which in turn could be useful in pinpointing the causative SNPs . The other 2-class problem , where an nsSNP is labeled as either strengthening or weakening the interaction , is useful when characterizing molecular mechanisms behind a SNP that has been already linked to a functional change . While an nsSNP occurring inside or in close proximity of an interaction interface will directly modify only one of the two interacting proteins , it is critical that our method takes into account the structural information of the entire interaction , including both binding sites forming the interaction interface . In this manner , the role of the interaction partner and its binding site is taken into consideration . For instance , it is possible that for a protein that competitively binds two other proteins through fully or partially overlapping binding sites , a mutation occurring in the overlapping region of these binding sites would disrupt one interaction but be neutral for another interaction . With hundreds of thousands of available interaction templates [89] and the advancement of comparative modeling , the requirement for structural information of the overall interaction makes an increasingly small impact on the coverage of SNP-IN tool . Understanding functional roles of nsSNPs associated with diseases by studying the disease-centered PPI network has many challenges . Being among the first such methods , SNP-IN tool is yet to deal with some of them . One of the key challenges is accounting for the indirect effects of nsSNPs on the interactions , such as disabling a phosphorylation site that regulates a PPI , altering an allosteric site , or nsSNP-induced structural changes of a protein that affect the interaction . The difficulty of modeling such effects lies in the complexity of indirect mechanisms , as well as in the fact that the effect-causing SNPs may be relatively distant from the protein interaction interface they affect . Another challenge is our ability to infer the functional importance of an nsSNP—and ultimately its contribution to the disease phenotype—from prediction of its effect on a PPI . For instance , the disruptive effect on a PPI predicted for an nsSNP that is either buried inside the interface or lies in its close proximity would indicate the true functional effect of the variation . However , predicting the neutral effect of a surface nsSNP that is in proximity to the interface does not necessarily mean that this genetic variation does not alter a biological function , as it could be a part of another functional site . On the other hand , an nsSNP that is buried inside the protein interaction interface is far less likely to be involved in the other function , e . g . , belong to a DNA- or small ligand–binding site or a site of posttranslational modification . Thus , the predicted neutral effect of such genetic variation would indeed mean that it does not have any functional impact . As a recent work by Wang et al showed [7] , there are thousands of nsSNPs associated with the interaction interfaces , and more SNPs are being identified every year from new high-throughput studies [90] . Combined with the exponential growth of the number of PPI structures being experimentally solved [91] , we expect that the coverage of SNP-IN tool will continue to grow , providing more insights into molecular mechanisms of complex genetic diseases . In addition , with the growing experimental knowledge about the cooperative effects of multiple nsSNPs on PPIs , we plan to expand the SNP-IN tool to multiple mutations as one of the next future steps . Even more challenging is a problem of computational estimation of the values upon structural changes in the protein interaction complex due to genetic variation . The classification of nsSNPs can be considered as a simplified , discretized , version of the latter problem . Based on the success of the current machine learning approach , we anticipate that the supervised and semi-supervised regression approaches will complement the classical biophysical methods to address this challenge . | Many genetic diseases in humans and animals are caused by combinations of single-letter mutations , or SNPs . When these mutations occur in a protein-coding region of a genome , they can have a profound effect on the protein's function and ultimately on a health-related phenotype . Recently , a growing number of evidence suggests that many of SNPs reside on or near the protein regions that are required for the interactions with other proteins . Some of these SNPs could rewire the protein-protein interactions altering the functions of the protein interaction complexes , while other SNPs are neutral to the interactions . Understanding the effect of SNPs on the protein-protein interactions is a challenging problem to solve , both experimentally and computationally . Here , we leverage the machine learning methods by training a computational predictor to tell apart the mutations that are harmful to protein-protein interactions from those ones that are not . We use these tools in two case studies of mutations affecting the protein-protein interaction networks centered around the genes associated with breast cancer and diabetes . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"computer",
"applications",
"algorithms",
"systems",
"biology",
"mutation",
"computer",
"and",
"information",
"sciences",
"computer",
"modeling",
"mathematics",
"genetics",
"applied",
"mathematics",
"biology",
"and",
"life",
"sciences",
"computing",
"methods",
"physical",... | 2014 | Determining Effects of Non-synonymous SNPs on Protein-Protein Interactions using Supervised and Semi-supervised Learning |
Mutations in BBS6 cause two clinically distinct syndromes , Bardet-Biedl syndrome ( BBS ) , a syndrome caused by defects in cilia transport and function , as well as McKusick-Kaufman syndrome , a genetic disorder characterized by congenital heart defects . Congenital heart defects are rare in BBS , and McKusick-Kaufman syndrome patients do not develop retinitis pigmentosa . Therefore , the McKusick-Kaufman syndrome allele may highlight cellular functions of BBS6 distinct from the presently understood functions in the cilia . In support , we find that the McKusick-Kaufman syndrome disease-associated allele , BBS6H84Y; A242S , maintains cilia function . We demonstrate that BBS6 is actively transported between the cytoplasm and nucleus , and that BBS6H84Y; A242S , is defective in this transport . We developed a transgenic zebrafish with inducible bbs6 to identify novel binding partners of BBS6 , and we find interaction with the SWI/SNF chromatin remodeling protein Smarcc1a ( SMARCC1 in humans ) . We demonstrate that through this interaction , BBS6 modulates the sub-cellular localization of SMARCC1 and find , by transcriptional profiling , similar transcriptional changes following smarcc1a and bbs6 manipulation . Our work identifies a new function for BBS6 in nuclear-cytoplasmic transport , and provides insight into the disease mechanism underlying the congenital heart defects in McKusick-Kaufman syndrome patients .
Bardet-Biedl syndrome ( BBS ) is a pleiotropic human genetic disorder belonging to a group of disorders known as ciliopathies [1–3] . Patients with BBS have multi-organ symptoms including: retinitis pigmentosa , polydactyly , obesity , and learning disabilities [4–8] . To date , mutations in 21 genes have been identified in causing BBS [9] . While BBS proteins belong to diverse proteins families , the significant overlap of phenotypes can be attributed , in part , to BBS proteins forming two main complexes: the BBSome and the BBS chaperonin complex . The BBSome , consisting of BBS1 , 2 , 4 , 5 , 7 , 8 , 9 , and 18 , functions in transporting ciliary cargo and , as more recently discovered , cargo destined for the cell’s plasma membrane [2 , 8] . The BBS chaperonin complex , consisting of BBS6 , 10 , and 12 , functions as an early , and transient , scaffold for assembly of the multi-protein BBSome complex [10 , 11] . The clinical features of BBS , and mutations in some BBS genes , are also associated with other syndromes . Mutations in BBS6 , for example , are associated not only with BBS , but also with McKusick-Kaufman syndrome ( MKKS ) [4 , 12] . The gene symbol for BBS6 is also MKKS . For clarity , we use BBS6 for the gene symbol to distinguish from the syndrome . MKKS , the syndrome , is an autosomal , recessive disorder with clinical features including congenital heart defects , female genital anomalies , and postaxial polydactyly . Noteworthy is that some characteristic symptoms of BBS including retinopathy , obesity , and intellectual disabilities are not present in MKKS [12–14] . While MKKS is rare in the general population , it does occur more frequently ( 1:10 , 000 births ) in Old-Order Amish populations . MKKS has one disease-associated allele consisting of two in cis missense mutations in BBS6: Histidine-to-Tyrosine at amino-acid position 84 and an Alanine-to-Serine at position 242 , BBS6H84Y; A242S [14] . MKKS associated heart defects , which are rare in patients with BBS , signal a likely cellular function for BBS6 outside of BBSome assembly . However , the cellular process that is disrupted by BBS6H84Y; A242S is presently unknown . Here we present data demonstrating that BBS6 is actively transported between the cytoplasm and nucleus . We demonstrate that the BBS6H84Y; A242S is defective in nuclear transport , but retains its cilia function . We identify a protein-protein interaction between BBS6 and SMARCC1 , a component of the SWI/SNF family of chromatin remodelers . Moreover , we demonstrate subcellular localization changes and overlapping transcriptional profiles which support that the BBS6 changes are mediated in part through SMARCC1 . This work provides insights into the disease pathophysiology underlying MKKS and reveals a new role for BBS6 in nuclear/cytoplasmic transport directly affecting the regulation of a core chromatin remodeling protein .
Due to the differential phenotypes between MKKS and BBS patients , and the fact that MKKS patients do not display retinitis pigmentosa , we posit that the MKKS disease allele , BBS6H84Y; A242S , may still retain cilia function . To test this , we used CRISPR/Cas9 to knockout Bbs6 in murine inner medullary collecting duct ( IMCD3 ) cells ( S1A Fig ) . We induced ciliation by serum starvation and , as in previous studies , quantified by counting the number of cells able to successfully generate and maintain cilia [2 , 15] . We find a significantly reduced number of ciliated cells in Bbs6 knockout cells compared to wildtype cells ( Fig 1A–1C ) . To functionally evaluate the alleles , the Bbs6 knockout cell line was transfected with tagged human forms of BBS6 and BBS6H84Y; A242S . We find that BBS6 and BBS6H84Y; A242S can both significantly rescue the ciliation defect ( Fig 1C , 1D and 1E ) . These data demonstrate that the MKKS disease allele , BBS6H84Y; A242S , retains the ability to generate and maintain the cilia . We next performed functional analyses in the zebrafish animal model . We previously established the zebrafish as a model system to study BBS proteins in vivo . All BBS genes we have tested to date have shown two characteristic knockdown phenotypes: Kupffer’s vesicle ( KV ) cilia defects and delays in retrograde cellular transport , as measured by transport of the zebrafish melanosomes [2 , 11 , 16–22] . Our previous studies also demonstrate that BBS gene knockdown phenotypes can be suppressed by introduction of exogenous mRNA [9 , 16 , 20–22] . Bbs6 knockdown , using two independent morpholino oligonucleotides ( MO ) , likewise generates a KV defect [18] ( S1C and S1D Fig ) . To functionally evaluate the BBS6H84Y; A242S allele in zebrafish , we performed knockdown and rescue in bbs6 MO-injected embryos ( morphants ) . Embryos first injected with bbs6 MO were split and received a second injection containing either human BBS6 or BBS6H84Y; A242S mRNA; a third group was held back as MO only . Embryos were fixed and stained at the 8–10 somite stage to image cilia in the Kupffer’s vesicles [23] . Max projections of confocal z-stacks were generated and quantified . The cilia placement in the KV allows us to measure the length of the cilia by tracing the fluorescent staining in FIJI/ImageJ . Knockdown of bbs6 in zebrafish causes a significant reduction in the length of the KV cilia ( S1D and S1G Fig ) . Both BBS6 and BBS6H84Y; A242S mRNA significantly suppress the length defect ( S1E–S1G Fig ) . Together , these results support the conclusion that the BBS6H84Y; A242S allele still allows for ciliogenesis and maintenance . Considering these findings , we proceeded to explore cilia-independent functions of BBS6 . BBS proteins are involved in transport not only to the cilia and plasma membrane , but also in retrograde transport in the zebrafish melanocyte [1 , 9 , 18 , 19] . We also noted that tagged BBS6 expressed in HEK 293T cells localizes to both cytoplasmic and nuclear compartments ( Fig 2A and 2E ) , suggesting that BBS6 may be involved in other cellular transport processes . BBS6 contains no identifiable nuclear localization signal , and with a molecular weight of 62kDa , is too large for diffusion across the nuclear membrane . Therefore , movement between the two subcellular compartments must be active transport . To test for active transport of BBS6 , we used an inhibitor of nuclear export , leptomycin B [24] . Leptomycin B ( LMB ) treatment of BBS6 transfected cells results in BBS6 accumulation in the nucleus . This is evident by confocal microscopy ( Fig 2A and 2A’ ) and cellular fractionation ( Fig 2E ) . We quantified these results by measuring fluorescent intensity in ImageJ and calculating the nuclear to cytoplasmic ratio , N/C ( S2 Fig ) . We find LMB treatment results in a significant increase in the amount of nuclear BBS6 compared to untreated cells , indicated by an increase in the N/C ratio ( Fig 2F , WT bars ) . Considering this new observation of BBS-related transport , we decided to evaluate nuclear-cytoplasmic transport dynamics of BBS6 disease alleles . To investigate this , we expressed the myc-tagged MKKS allele , BBS6H84Y; A242S , and two BBS alleles , BBS6Y37C and BBSL454P , in HEK 293T cells and evaluated subcellular distribution in response to LMB treatment [12 , 25] . Similar to wildtype BBS6 , both BBS disease-associated alleles , BBS6Y37C ( Fig 2B and 2B’ ) and BBS6L454P ( Fig 2C and 2C’ ) , show a statistically significant nuclear accumulation upon leptomycin B treatment , indicating that nuclear import is unaffected by these mutations ( Fig 2F ) . In contrast , the MKKS allele , BBS6H84Y; A242S , shows no significant change in nuclear accumulation upon leptomycin B treatment indicating a pronounced nuclear import defect unique to MKKS disease-allele ( Fig 2D , 2D’ and 2F ) . The BBS6H84Y; A242S import defect was also observed by cellular fractionation and western blot ( Fig 2E ) . These results show that active transport of BBS6 between the cytoplasm and nucleus is disrupted in the McKusick-Kaufman syndrome allele BBS6H84Y; A242S , and this defect is not observed in BBS-associated alleles . We next sought to uncover what protein-protein interactions BBS6 may be participating in , that may link this observed phenotype to the MKKS disease symptoms . To identify new interacting partners which may lend insight into nuclear-transport-related functions , we created a zebrafish line with an inducible promoter driving bbs6 expression . We used Tol2-mediated transgenesis to generate a stable zebrafish line expressing GFP-tagged Bbs6 under the control of a heat-shock promoter: Tg ( hsp70:GFP-bbs6 ) , referred to as Tg ( bbs6 ) ( Fig 3A ) . We confirmed expression of the transgene in response to heat-shock by imaging live post-heat-shock embryos under fluorescence and by RT-PCR of transcripts for GFP ( Fig 3B and 3C ) . We also confirmed that there is no transgene expression in non-heat-shocked Tg ( bbs6 ) embryos ( Fig 3C ) . As further characterization of the transgenic line , we tested the function of the GFP-Bbs6 protein in the context of Bbs6 knockdown . Morphant Tg ( bbs6 ) embryos were divided into heat-shock and non-heat shock groups and evaluated for melanosome transport ( Fig 3D ) . Wildtype morphants and non-heat-shocked Tg ( bbs6 ) morphants show delayed transport ( Fig 3E ) consistent with our previously published finding for Bbs6 knockdown [1 , 9 , 18 , 19] . In contrast , the melanosome transport delay is significantly suppressed in sibling Tg ( bbs6 ) morphants with heat-shock induced expression of the GFP-Bbs6 protein ( Fig 3E ) , demonstrating that the GFP-Bbs6 fusion protein is functional in vivo . We used this GFP-Bbs6 fusion protein expressed in the Tg ( bbs6 ) line to screen for new binding partners . Protein lysates were isolated from heat-shocked and non-heat-shocked , 3 days-post-fertilization ( dpf ) , Tg ( bbs6 ) larvae . Antibodies against GFP were used to immunoprecipitate ( IP ) the transgenic protein . The IP samples from heat-shock and non-heat-shock larvae were run on an SDS-PAGE and silver stained . Major bands present in the heat-shock , but absent in the control lane , were isolated for mass-spec analysis ( Fig 4A ) . As expected , the GFP-Bbs6 fusion protein was successfully isolated and identified from the pull down ( Fig 4A , bottom arrow ) . We also identified Smarcc1a , the zebrafish homolog of human SMARCC1 ( SWI/SNF related , matrix associated , actin dependent regulator or chromatin subfamily c , member 1 ) ( Fig 4A , top arrow ) . While defects in the SWI/SNF chromatin remodeling complex have been linked to neuronal development defects and cancer , noteworthy is its association with congenital heart defects [26–30] . Due to the role of the SWI/SNF complex in the heart , we posit that disruption of BBS6-SMARCC1 interaction may contribute to the congenital heart defects observed in MKKS . To explore this , we first cloned and characterized zebrafish smarcc1a . The smarcc1a transcript is maternally supplied and ubiquitously expressed in early development , as determined by in situ hybridization and RT-PCR ( S3 Fig ) . Smarcc1a is also a highly-conserved protein , displaying 74% identity to human SMARCC1 ( S4 Fig ) . For in vivo validation of the Bbs6-Smarcc1a interaction , we verified the interaction between endogenous Smarcc1a and myc-tagged Bbs6 , expressed in zebrafish embryos . Protein lysates from 24 hour-post-fertilization ( hpf ) embryos were subject to IP with a human SMARCC1 antibody that cross-reacts with zebrafish Smarcc1a , blotted , and probed for myc-Bbs6 . We can detect Bbs6 with pull down of Smarcc1a ( Fig 4B , 2nd lane ) , confirming our mass-spec data and the interaction of Smarcc1a and Bbs6 in vivo . We also determined this interaction was conserved in human BBS6 and was still present in BBS6H84Y; A242S ( Fig 4B , 3rd and 4th lanes ) . SMARCC1 homologs have been shown to be present in the cytoplasm , and the cytoplasmic partitioning regulated in response to cellular conditions [31 , 32] . To better understand the interaction between these two proteins we examined the sub-cellular localization of Bbs6 and Smarcc1a in zebrafish embryos . To visualize Bbs6 , zebrafish were injected with myc-bbs6 mRNA , immunostained , and imaged by confocal microscopy . We find predominant cytoplasmic expression of myc-Bbs6 with low-level nuclear staining , matching what we observe in 293T cells , ( S5A Fig and Fig 2A ) . To confirm this nuclear presence in zebrafish , we performed cellular fractionation and find most myc-Bbs6 in the cytoplasmic fraction with a smaller population in the nuclear fraction ( S5B Fig ) , matching the confocal observations of those same zebrafish cells ( S5A Fig ) . In contrast , endogenous Smarcc1a/SMARCC1 shows robust nuclear localization by immunofluorescent imaging in zebrafish and in human cell culture ( Fig 5A and 5B ) . Due to the intense nuclear staining , low level cytoplasmic staining may not be directly evident; therefore , we performed cellular fractionation of zebrafish embryos and HEK 293T cells and found low levels of endogenous Smarcc1a/SMARCC1 in the cytoplasm ( Fig 5C ) . While BBS6 and SMARCC1 are predominantly localized to different subcellular compartments , we find overlap within both the cytoplasm and nucleus , and this overlap presents an opportunity for protein interaction . To determine the subcellular compartment where BBS6 and SMARCC1 are interacting , we transfected HEK 293T cells with myc-tagged BBS6 . These cells were fractionated and lysates subjected to coIP using a SMARCC1 antibody ( Fig 5D ) . The IP fractions were blotted and probed for myc-BBS6 . We find myc-BBS6 in both the nuclear and cytoplasmic IP fractions ( Fig 5E ) . While this demonstrates that BBS6 and SMARCC1 can interact in both cellular compartments , we observe the greatest interaction in the cytoplasmic IP fraction . By normalizing the amount of myc-BBS6 pulled down to amount present in the input ( Fig 5F ) , our data shows that the BBS6-SMARCC1 interaction occurs predominantly in the cytoplasm of the cell . This data suggests that BBS6 may be regulating the transport of SMARCC1 . We next sought to explore the mechanism of this regulation . To investigate the mechanism of this interaction , we determined the extent to which BBS6 modulates SMARCC1 sub-cellular localization . In the first approach , we used BBS6 knockout HEK 293T cell lines , generated by CRISPR/Cas9 ( S1B Fig ) . As previously shown , cellular fractionation of wildtype HEK 293T cells shows SMARCC1 predominantly in the nucleus with some protein in the cytoplasm ( Fig 6A , BBS6+/+ ) . However , in the BBS6 knockout cell line , there is decreased cytoplasmic SMARCC1 ( Fig 6A , BBS6-/- ) . Quantification of confocal images also shows this same pattern ( Fig 6B ) . These results suggest that SMARCC1 subcellular localization is affected by the presence of the BBS6 protein . We next questioned if BBS6 over-expression would be sufficient to alter SMARCC1 cytoplasmic levels . To ensure that we are examining the impact of only our exogenous BBS6 , we transfected myc-tagged BBS6 into BBS6 knockout cells and performed immunostaining for BBS6 and endogenous SMARCC1 ( Fig 6B–6E ) . Nuclear and cytoplasmic intensities from single z-slices of individual cells were quantified and the nuclear/cytoplasmic ( N/C ) ratios were calculated ( S2 Fig ) . Cells transfected with BBS6 show a significant increase in cytoplasmic SMARCC1 , indicated by lower N/C ratios ( Fig 6B ) , and evident by comparing cytoplasmic SMARCC1 staining in transfected vs neighboring untransfected cells ( Fig 6D ) . In agreement with our data showing BBS6H84Y; A242S can still interact with SMARCC1 ( Fig 4C ) , we also see SMARCC1 subcellular localization changes in BBS6H84Y; A242S transfected cells compared to untransfected controls . When comparing the SMARCC1 localization between BBS6 and BBS6H84Y; A242S , we find that BBS6H84Y; A242S causes more cytoplasmic retention than BBS6 ( Fig 6C ) . This is consistent with the hypothesis that the BBS6H84Y; A242S disease allele affects SMARCC1 subcellular localization . These data indicate that BBS6 modulates SMARCC1 sub-cellular localization , likely negatively regulating its import into the nucleus by retaining it in the cytoplasm . This negative regulation of import model predicts that nuclear Smarcc1a would be reduced by both Bbs6 over-expression and Smarcc1a knockdown . Therefore , these conditions should have similar effects on transcriptional regulation . To test this , we performed transcriptional profiling by high throughput RNA sequencing ( RNASeq ) . We utilized CRISPR/Cas9 to generate four independent smarcc1a alleles ( S6A Fig ) . F1 smarcc1a heterozygous embryos show severe cardiovascular defects as well as retinal , fin , and body axis defects ( S6A Fig ) . Our observations are consistent with the early embryonic lethality in mice heterozygous for SMARCC1 mutations [33–36] . smarcc1a knockdown ( using both translation blocking and splice blocking morpholinos ) results in multi-organ defects similar to the smarcc1a heterozygous CRISPR mutants . We titrated gene function by reducing the MO dose and find that the cardiovascular defect is the most penetrant of the phenotypes , which persists as the other phenotypes taper off ( S6B Fig ) . In the low dose context , early embryonic morphological development is normal . Defects become apparent during later heart morphogenesis ( S6C Fig ) . The specificity and efficacy of knockdown was validated by western blot showing reduced endogenous Smarcc1a protein levels in morphants ( S6D Fig ) and by the ability of exogenous smarcc1a mRNA to significantly suppress knockdown defects ( S6E Fig ) . Zebrafish injected with either smarcc1a MO or bbs6 mRNA at the one cell stage were cultured until 48hpf , at which point cardiac enriched tissue was isolated [37] . The enrichment of heart tissue was verified by performing qRT-PCR on pools of heart-enriched RNA and whole embryo RNAs for three different heart markers: myl7 , nkx2 . 5 , and nppa ( Fig 7A ) . For RNASeq approximately 200 hearts were used per sample and four biological replicates per condition: bbs6 over-expression , smarcc1a knockdown , and control embryos . Differential expression analysis of RNASeq data from bbs6 over-expression and smarcc1a knockdown was performed against controls . Using a stringent adjusted p-value of <0 . 0001 as a cut-off , we find 5 , 369 differentially expressed genes in smarcc1a knockdown hearts ( relative to controls ) ( S1 File ) and 449 genes in bbs6 over-expression hearts ( relative to controls ) ( S2 File ) . Comparing both gene sets , we find that 377 genes , or 84% of genes with altered expression in the bbs6 over-expression group , are also differentially expressed in the smarcc1a knockdown group ( Fig 7B ) ( S3 File ) . To determine the probability of getting this substantial overlap of genes by chance we performed a hypergeometric distribution test and get a p-value of approximately 10−133 , indicating it is extremely unlikely to observe this distribution by chance . Comparing the log2-fold-changes ( log2FC ) of those 377 shared genes reveals a strong positive correlation with a spearman correlation of 0 . 66 and a p-value of ( 2 . 4 X 10−48 ) ( Fig 7C ) . That is , any given gene affected by Bbs6 over-expression is likely to be affected in the same way by Smarcc1a knockdown . This strong correlation indicates that Bbs6 over-expression and Smarcc1a knockdown cause similar transcription changes . To look for patterns of genes affecting similar cellular processes in the shared set of 377 genes , we performed a Gene Ontology ( GO ) analysis using the WEB-based Gene Set Analysis Toolkit ( WebGestalt ) [38] . When performing the analysis for the cellular component GO terms we find significant gene sets for the cytoskeleton , SNARE proteins , and membrane proteins ( Fig 7D ) . These gene sets are particularly interesting in respect to BBS6 as there is a direct relationship to cellular transport . Additionally , the top differentially expressed genes ( two-fold or greater change in expression after bbs6 mRNA over-expression and with a concordant change in smarcc1a knockdown ) , we find enrichment for genes with known roles in heart development and cardiac function ( Table 1 ) . Additionally , fgf8a and hoxb1b , which exhibit slightly less than two-fold changes in response to bbs6 over-expression are included because of their significance to heart development . The genes in Table 1 have essential roles in secondary heart field development , cardiac valve formation , and heart morphogenesis—all developmental processes linked to congenital heart disease . With the shared set of co-differentially expressed genes being identified in zebrafish , we sought to determine relevance to human CHD . To explore this we compared our 377 shared genes to a list of de-novo mutations found from whole exome sequencing of parent-offspring trios in which the child has congenital heart disease [58] . Using Ensembl BioMart we matched up our gene list with predicted human homologs and filtered against the list of genes in which de-novo mutations were identified in patients with CHD , resulting in an overlap of 30 genes ( Table 2 ) . While this matched-list does not demonstrate direct clinical relevance , it does provide promising candidates for future studies .
Our findings suggest that the BBS6H84Y; A242S loss of function in nuclear/cytoplasmic transport is the molecular mechanism underlying congenital heart defects in MKKS patients , as is depicted in our model ( Fig 8 ) . We hypothesize that BBS6 binds SMARCC1 in the cytoplasm where it functions in negatively regulating SMARCC1 import . This is supported by our observations of decreased SMARCC1 in the cytoplasm when BBS6 is lost and increased SMARCC1 in the cytoplasm when BBS6 is over-expressed ( Fig 6B–6E ) . Our data also show that BBS6H84Y; A242S can still bind to SMARCC1 , but as demonstrated with leptomycin B treatment , is defective in its ability to enter the nucleus . BBS6H84Y; A242S binding to SMARCC1 , but being unable to enter the nucleus would lead to reduction of SMARCC1 in the nucleus . Knockdown experiments in zebrafish , and attempts to create a CRISPR knockout line , reveal that smarcc1a is extremely sensitive to gene dosage . We find extremely small doses of MO are sufficient to cause severe heart defects in developing zebrafish ( S6C Fig ) . This is consistent with published data demonstrating early embryonic lethality in mice heterozygous for SMARCC1 [59 , 60] . Moreover , bioinformatic analysis of human protein variants support that mutations in SMARCC1 are not tolerated [ExAC Browser; [61]] . In fact , a recent whole exome sequencing study identified a de-novo SMARCC1 missense mutation in a patient with congenital heart disease , supporting the sensitivity of the heart to perturbations of the SMARCC1 [58] . Our data , taken with previous studies demonstrating the role of the SWI/SNF complex in cardiovascular development , suggests a scenario in which small perturbations of the SMARCC1 subcellular distribution , brought on by defective transport of BBS6H84Y; A242S , lead to congenital heart defects in MKKS patients . Altering SMARCC1 levels leads to changes in expression of genes regulated by the SWI/SNF complex . While the complete mechanism linking BBS6 , SMARCC1 , and heart development needs further investigation , this work provides the first demonstration of a possible disease mechanism explaining the phenotypic differences between MKKS and BBS . The heart defects observed in MKKS patients include atrial and ventricular septal defects , atrioventricular canal and valve defects , tetralogy of Fallot , and patent ductus arteriosus [62] . Disruptions of the secondary heart field can lead to atrial and ventricular septation defects and tetralogy of Fallot [63–65] . We find that key genes ( e . g . , isl2b , hoxb1b , prdm1b ) and signaling pathways ( Fgf and retinoic acid ) critical for secondary heart field development are altered in the same manner by bbs6 over-expression and smarcc1a knockdown ( Table 1 and Supplemental Tables ) . Additionally , fgf8 and pcsk6 , are two of the key nodes in a proposed atrial septation network [49] . Finally , the shared list includes genes involved in endocardial cushion/AV valve development ( e . g . wnt9a and versicans b ) or linked to patent ductus arteriosus risk ( sema3e ) [48 , 55] . When we examined early left-right patterning markers which are necessary for correct heart orientation , we found no differences in smarcc1a knockdown compared to wildtype . Furthermore , as previously mentioned , there are not severe heart jogging defects at 30 hpf in smarcc1a knockdown embryos . Taken together , the candidate genes and the smarcc1a knockdown cardiovascular defects support a role for the secondary heart field . In this study , we have demonstrated that BBS6 is being actively transported between the cytoplasm and nucleus . We have shown that the McKusick-Kaufman-syndrome-associated allele , BBS6H84Y; A242S , can still function in cilia-related processes but is defective in its ability to enter the nucleus . We identified a novel BBS6 interacting protein , SMARCC1 , a SWI/SNF chromatin remodeling protein , in vivo and in vitro , by using zebrafish and human cell culture . Additionally , we have demonstrated that BBS6 modulates the sub-cellular localization of SMARCC1 , with reduced cytoplasmic SMARCC1 in BBS6 knockout cell lines , and inversely , increased cytoplasmic SMARCC1with over-expression of BBS6 . And finally , we observe that bbs6 over-expression causes overlapping transcriptional changes to smarcc1a knockdown . Our results provide a candidate mechanism for MKKS and list of affected heart genes with possible relevance to the etiology of the CHDs seen in MKKS—both exciting directions to explore further . On a broad level , this study identifies a new function for BBS6 in nuclear-cytoplasmic transport . Interestingly , recent publications have demonstrated a high degree of similarity between the nuclear-pore complex and the cilia transition zone , both the gatekeepers of their respective cellular compartments , making the possibility of BBS proteins also participating in nuclear transport plausible [66–70] . The assembly of the BBSome is a complex process involving the addition and removal of BBS chaperonin complex members until the final BBSome is completed [71] . BBS proteins may be modular components that can be assembled as needed to form different complexes with different functions at specific time points in development or in a tissue-specific manner . Given the mentioned similarities between the cilia transition zone and the nuclear pore complex , it is plausible that BBS proteins participate in cellular transport pathways in addition to cilia transport . The complete list of BBS proteins involved in this process , is outside of the scope of this paper but does present an exciting path for future work .
All work involving zebrafish was approved by the University of Iowa’s Institutional Animal Care and Use Committee , PHS Assurance No . A3021-01 , under animal protocol No . 5091513 . A construct with N-terminal , GFP-tagged zebrafish Bbs6 under the control of the hsp70 heat-shock promoter , flanked by Tol2 recombination sites , was generated using standard gateway cloning methods . Clutches of embryos were injected with this construct and tol-2 transposase mRNA at the 1–4 cell stage . Founders were outcrossed to wildtype and progeny raised to generate a stable transgenic zebrafish line Tg ( hsp70:GFP-bbs6 ) . Culture plates of zebrafish were transferred to a 50ml conical tube . Excess water was aspirated . Egg water pre-heated to 37°C was added to the tube , and the tube was placed into a 37°C water bath . Heat-shock was performed for the following times for the respective developmental stage: 24hpf larvae: 30 minutes , 48hpf-60hpf: 45 minutes . After the appropriate time , tubes were removed from the water bath and placed into a 28 . 5°C incubator to cool gradually . For melanosome transport rescue assays zebrafish were heat-shocked daily from 1dpf– 5dpf . For protein-interaction identification embryos were heat-shocked daily from 1dpf– 3dpf . Non-heat-shocked control embryos followed the same procedure , except 28 . 5°C water was used in place of 37°C . Following this heat-shock protocol no developmental or morphological defects are observed in heat-shocked embryos . Zebrafish embryos collected from natural mating were pressure injected at the 1–4 cell stage . mRNA for myc-tagged zebrafish Bbs6 was generated by in vitro transcription using the SP6 mMESSAGE mMACHINE kit ( ThermoFisher Scientific ) . 300pg of synthesized mRNA was injected from concentrations of 300 ng/μl . A translation blocking smarcc1a morpholino oligonucleotide ( MO ) ( GeneTools ) was injected at 0 . 5 ng/embryo or lower . bbs6 MO was injected at 2 ng/embryo . Standard control MO from GeneTools was used and was injected at an equal quantity as the experimental MO for each experiment . Microinjection volume was measured in a 1μl capillary tube and calculated using the Microinjection Calculator Android app ( available from Google Play Store ) . To label the developing heart in zebrafish we used a transgenic line driving GFP in differentiated heart tissue Tg ( myl7:EGFP ) . bbs6 MO sequence: 5’-ACTGCACAAACCTTCAGTTCTTCCA-3’ smarcc1a ATG MO sequence: 5’-CAGTCGCCGCTGTCGCCATTGTTTC-3’ smarcc1a splice MO sequence: 5’-CATGAGCAGCAGACCTTCTTATAAT-3’ Melanosome transport assays were performed on dark-adapted , 5 days-post-fertilization zebrafish larvae as previously described [1] . In short , individual zebrafish were placed into a single well of a multi-well plate . Epinephrine was added , and the time required for each fish to retract the melanocytes was measured . Timing was stopped after 6 minutes and any fish not fully contracted were recorded as 360 seconds . Total RNA from 20 , 1dpf zebrafish embryos . This total RNA was reverse transcribed into cDNA using SMART MMLV Reverse Transcriptase ( Clonetech ) primed with oligo-dT primers . This cDNA library was then used to clone a portion of smarcc1a using primers 5’-GGAGGGCCATCTTCCAAGTA-3’ and 5’-AGGGACTTGCGTTCCTTACG-3’ . This product was ligated into a TOPO-TA PCR-II vector ( ThermoFisher Scientific ) following the manufacturers protocol . DIG-labeled RNA-probes ( DIG labeling Mix , Roche ) were synthesized using T7 and SP6 Maxi-script kits ( ThermoFisher Scientific ) following the manufacturers protocol . In situ hybridization was performed as previously described [72] . Human kidney epithelial cells ( HEK 293T , ATCC CRL-3216 ) and mIMCD-3 ( ATCC CRL-2123 ) were cultured in DMEM and DMEM-F12 ( Life Technologies ) supplemented with 10% FBS respectively . For serum starvation mIMCD-3 cells were cultured in DMEM-F12 without FBS supplementation . Transfections were performed with Lipofectamine 2000 using recommended manufacturer conditions . HEK 293T cells were plated onto glass coverslips , transfected , and cultured for 24-hours . The media was replaced with fresh media containing a 20nM concentration of leptomycin B . Cells were incubated with leptomycin B for 2 hours in standard conditions and fixed for 20 minutes in 4% paraformaldehyde at room temperature . Primary antibodies used for staining zebrafish and cell lines are Cell Signaling ( 9B11 ) myc-tag antibody ( 1:3000 dilution ) and a sigma ( T6793 ) monoclonal acetylated-tubulin antibody ( 1:800 dilution ) . An Alexa-488 or 633 conjugated antibodies ( 1:400 dilution ) was used as secondary antibodies . Antibodies were diluted in a blocking buffer made up of 1X PBS + 2 . 5% BSA + 0 . 1% Triton . Washes were performed in 1X PBS + 0 . 1% Triton . Z-stacks were obtained on fixed and stained cells using a Leica SP5 confocal microscope with a 63X oil-objective . For N/C quantification , mean fluorescent intensity values from each region of interest were measured on single z-slices from stacks using FIJI package of Image J . The nucleus and cytoplasm were differentiated by using DAPI staining . A region overlapping the DAPI was used to measure the nuclear region . A region immediately adjacent to , but outside of , the DAPI staining was counted as the cytoplasm . The nuclear/cytoplasmic intensity ratio was calculated for 40–60 individual cells per experimental set and averaged for each group . For cilia length measurements , max projections of z-stacks were generated using FIJI . Cilia were traced by hand and lengths of each tracings were calculated within FIJI . Human cells and zebrafish larvae were fractionated using a BioChain CNM compartmental protein extraction kit ( K3012010-FS ) following the manufacturer’s protocol . Protein lysates were extracted from zebrafish larvae and tissue culture cells using standard protein lysis buffers containing protease inhibitors PMSF and leupeptin . Each protein lysate was brought to a volume of 400 μl with protein lysis buffer , and 2 μl of SMARCC1 antibody ( Abcam ab172638 ) was added . Protein lysates were rotated overnight at 4°C . To each sample 25 μl of suspended Protein A/G agarose beads were added . Samples were rotated for 2 hours at 4°C . Samples were centrifuged to pellet the beads and washed several times before being boiled in SDS-PAGE Buffer + β-mercaptoethanol . Samples were run on the NuPAGE SDS-PAGE system using 4–12% Bis-Tris gels and NuPAGE MOPS running buffer . Gel electrophoresis was carried out at 150V for 1 . 5 hours . For western blotting , the gel was transferred onto PVDF membrane at 30V for 2 hours at room temperature . Blocking was performed in 5% milk in 1X TBS-Tween for 1 hour at room temperature . Membranes were probed with appropriate antibodies . For mass spectrometry and protein identification , after the gel electrophoresis , the SDS-PAGE gel was stained with Silver-Quest silver staining kit following the manufacturer’s protocol . Bands of interest were excised from the silver stained SDS-PAGE gel and submitted to the Carver College of Medicine Roy J . Carver Charitable Trust–supported CCOM Proteomics Facility at the University of Iowa . Primary antibodies: Cell Signaling Technologies 9B11 myc-tag antibody at a 1:10000 dilution , Cell Signaling Technologies ( D4C3 ) SP1 antibody at 1:1000 dilution , Developmental Studies Hybridoma Bank ( AA4 . 3 ) TBA1 antibody at ( 0 . 3μg/mL ) , and Abcam SMARCC1 ( ab172638 ) at a 1:5000 dilution . Secondary antibodies: HRP conjugated goat anti mouse and rabbit antibodies from Jackson ImmunoResearch used at a 1:20000 dilution . Western blots were developed ether using X-ray film or the LiCor C-Digit chemiluminescent scanner . The human codon-optimized Cas9 plasmid pX459 ( Addgene # 48139 ) was obtained from Addgene [73] . sgRNAs were designed and constructed as described previously [73 , 74] . Briefly , a target 20bp sequences starting with guanine and preceding the PAM motif ( 5′-NGG-3′ ) was selected from the gene of interest [73 , 75] . Potential off-target effects of sgRNA candidates were analyzed using the online tool CRISPR Design developed by Zhang's laboratory ( http://crispr . mit . edu/ ) . Briefly , 2 ug of plasmid containing a sgRNA targeting the gene of interest and 6 μl Lipofectamine 2000 were diluted in 100 ul Opti-MEM , mixed 1:1 , and added to cells after 5-minute incubation . After 24 hours , the cells were passed at low density ( 2000 cells/well ) in 10 cm culture dishes and selected with puromycin at 1 . 5 ug/ml for clonal expansion . The media was changed every 2 days until colonies were harvested using cloning cylinders . | To understand how mutations in one gene can cause two distinct human syndromes ( McKusick-Kaufman syndrome and Bardet-Bield syndrome ) , we investigated the cellular functions of the implicated gene BBS6 . We found that BBS6 is actively transported between the cytoplasm and nucleus , and this interaction is disrupted in McKusick-Kaufman syndrome , but not Bardet-Biedl syndrome . We find that by manipulating BBS6 , we can affect another protein , SMARCC1 , which has a direct role in regulating gene expression . When we profiled these changes in gene expression , we find that many genes , which can be directly linked to the symptoms of McKusick-Kaufman syndrome , are affected . Therefore , our data support that the nuclear-cytoplasmic transport defect of BBS6 , through disruption of proteins controlling gene expression , cause the symptoms observed in McKusick-Kaufman syndrome patients . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"medicine",
"and",
"health",
"sciences",
"cardiovascular",
"anatomy",
"293t",
"cells",
"biological",
"cultures",
"vertebrates",
"animals",
"animal",
"models",
"osteichthyes",
"organisms",
"developmental",
"biology",
"cytoplasmic",
"staining",
"model",
"organisms",
"experi... | 2017 | Nuclear/cytoplasmic transport defects in BBS6 underlie congenital heart disease through perturbation of a chromatin remodeling protein |
By the age of 40 , one in five adults without symptoms of cardiovascular disease are at risk for developing congestive heart failure . Within this population , dilated cardiomyopathy ( DCM ) remains one of the leading causes of disease and death , with nearly half of cases genetically determined . Though genetic and high throughput sequencing-based approaches have identified sporadic and inherited mutations in a multitude of genes implicated in cardiomyopathy , how combinations of asymptomatic mutations lead to cardiac failure remains a mystery . Since a number of studies have implicated mutations of the transcription factor TBX20 in congenital heart diseases , we investigated the underlying mechanisms , using an unbiased systems-based screen to identify novel , cardiac-specific binding partners . We demonstrated that TBX20 physically and genetically interacts with the essential transcription factor CASZ1 . This interaction is required for survival , as mice heterozygous for both Tbx20 and Casz1 die post-natally as a result of DCM . A Tbx20 mutation associated with human familial DCM sterically interferes with the TBX20-CASZ1 interaction and provides a physical basis for how this human mutation disrupts normal cardiac function . Finally , we employed quantitative proteomic analyses to define the molecular pathways mis-regulated upon disruption of this novel complex . Collectively , our proteomic , biochemical , genetic , and structural studies suggest that the physical interaction between TBX20 and CASZ1 is required for cardiac homeostasis , and further , that reduction or loss of this critical interaction leads to DCM . This work provides strong evidence that DCM can be inherited through a digenic mechanism .
Heart failure is a major cause of morbidity in the United States with more than 5 million people in the US living with this disease [1] . A major risk factor for developing heart failure is dilated cardiomyopathy ( DCM ) . Clinically recognized as systolic dysfunction accompanied by dilation of one or both ventricles , DCM is a predominating cardiomyopathy and the most common disease requiring heart transplantation in the US [2 , 3]; however , nearly half of DCM cases are of unknown etiology [4] . In efforts to understand the etiology of idiopathic DCM , mutations in over 50 genes including components of the contractile apparatus and cell cytoskeleton , as well as in factors involved in excitation-conduction coupling , have been identified as causative in DCM [5 , 6] . However , few studies have explored the potential for aberrant transcriptional regulation of these factors to contribute to disease pathogenesis . In exception to this , recent studies have identified mutations in the T-box transcription factor TBX20 associated with DCM [7–9] . Results of genetic analysis and protein depletion studies are consistent with an essential role for TBX20 during the early stages of vertebrate heart development [10–17] . Hearts lacking Tbx20 show progressive loss of cardiomyocytes , failure of the heart to undergo looping and chamber formation , and defects in cardiomyocyte maturation [17–21] . In humans , loss-of-function mutations in TBX20 can cause dilated cardiomyopathy , atrial septal defects , or mitral valve disease , while gain-of-function mutations in TBX20 have been reported in patients with Tetralogy of Fallot ( i . e . , pulmonary outflow tract obstruction , ventricular septal defect , overriding aortic root and right ventricular hypertrophy ) [7 , 8 , 22–24] . It has been further demonstrated that ablation of Tbx20 in adult mouse cardiomyocytes leads to the onset of severe cardiomyopathy leading to death within 1–2 weeks after Tbx20 loss [25] . While TBX20 is an essential transcription factor for heart development and its disease relevance is well established , many fundamental questions remain about the mechanism of TBX20 function . Principle among these is how TBX20 mutations associated with DCM circumvent the essential embryonic cardiac requirement for TBX20 . To elucidate the mechanisms by which mutations in TBX20 lead to human adult pathological states , we identified endogenous TBX20 cardiac protein-protein interactions by coupling a tagged endogenous allele of Tbx20 with unbiased proteomic analysis . Results from these studies revealed TBX20 interacts with the essential cardiac transcription factor Castor ( CASZ1 ) , a gene that was also recently linked to DCM [26] . We confirmed that TBX20 and CASZ1 interact biochemically and genetically , and we go on to show that while mice singularly haploinsufficient for Tbx20 or Casz1 are asymptomatic , mice heterozygous for both Tbx20 and Casz1 die , beginning at 4 to 8 weeks post birth , and exhibit cardiomyocyte hypertrophy , interstitial fibrosis , and severe DCM . Interestingly , the human mutant TBX20F256I bypasses the early essential requirement for TBX20 but leads to DCM . We report here that TBX20F256I disrupts the TBX20-CASZ1 interaction , ascribing clinical relevance to this protein complex . Further , by using quantitative proteomics we have identified the molecular pathways altered in TBX20-CASZ1-mediated DCM . Together , these results identify a novel interaction between TBX20 and CASZ1 that is essential for maintaining cardiac homeostasis . These findings imply that DCM can be inherited through a digenic mechanism .
To identify endogenous protein interactions that regulate TBX20 function , we introduced the Avitag , in-frame , to the carboxy terminus of mouse Tbx20 through homologous recombination in mouse embryonic stem cells ( ESCs ) ( Tbx20Avi ) ( S1A and S1B Fig ) . Since the Avi-tag can be biotinylated through recognition of the Avi-tag sequence by the E . coli biotin ligase BirA [27 , 28] , we generated a lentivirus expressing BirA and transduced it into mouse Tbx20Avi/+ ESCs . After hygromycin selection , Tbx20Avi/+ ESCs that stably expressed BirA ( S1C Fig ) were differentiated into induced cardiomyocytes ( iCM ) using a serum-free differentiation method that routinely generates cultures containing >60% cardiomyocytes ( Fig 1A ) [29] . Expression analysis at each day of differentiation confirmed the Tbx20Avi:BirA ESCs recapitulated the wild-type cardiomyocyte differentiation program ( Fig 1B ) . We further showed by Myosin Heavy Chain ( MHC ) expression and time lapse imaging that the Tbx20Avi:BirA ESCs differentiated into beating neonatal cardiomyocytes ( Fig 1C , S1 Movie ) . Our analysis further verifies that Tbx20Avi expression recapitulates endogenous Tbx20 expression with highest levels in immature cardiomyocytes at day 4 and differentiated cardiomyocytes at day 7 ( Fig 1B ) . Published data has demonstrated a requirement for TBX20 in adult mice , with loss of TBX20 leading to abrupt cardiac failure [25] . Recently , mutations in TBX20 in humans were associated with DCM [7–9] . To delineate the mechanisms of how TBX20 DCM-associated mutations circumvent the essential requirements for TBX20 in cardiac development , we isolated and characterized the endogenous TBX20 cardiac protein interactome under physiological conditions from Tbx20Avi; BirA iCMs at day 7 of differentiation . As a control for non-specific interactions , identical affinity isolations were performed from BirA-negative iCMs . Proteins co-isolated from TBX20Avi affinity purifications ( APs ) were analyzed by an SDS-PAGE tandem mass spectrometry-based proteomics approach , as in [30] . TBX20 was detected in the AP from BirA-expressing iCMs , with 19 unique tryptic peptides covering 54% of the TBX20 sequence ( out of a theoretical maximum coverage of ~75% ) ( Fig 1D and 1E; S2A and S2B Fig; S1 Table ) . Identification of candidate high-confidence TBX20 interactions that have the potential to regulate cardiac functions was achieved using a multi-step bioinformatics approach based on the number of identified spectra per protein . First , interacting proteins identified by less than 10 spectra did not meet the identification requirement and were excluded from further analysis . Further , proteins identified in the BirA-expressing isolations were required to have at least a 4-fold increase in identified spectra over isolations from control iCMs . Due to the ascribed function of TBX20 as a critical cardiac transcription factor , we specifically focused on proteins with a nuclear or unknown subcellular localization . Finally , these interaction candidates were ranked by their AP enrichment ( AP abundance versus whole cell abundance , S1 Table ) , which we have previously used to highlight the most prominent associations suitable for functional validation [31 , 32] . Interestingly , the top 50 most enriched proteins in the TBX20 AP were predominately ( 32/50 ) annotated to Chromatin and Transcription gene ontologies ( Fig 2A; S1 Table ) . Functional annotation of these proteins in the STRING database [33] revealed an interconnected network containing components of chromatin remodeling and RNA polymerase transcriptional complexes ( including four components of the INO80 complex—Ino80 , Actr5 , Actr5 , Nfrkb , and five components of the RNA Pol II mediator complex- Med13 , Med14 , Med17 , Med19 , Med27 ) ( Fig 2A ) . These data suggest that TBX20 predominantly acts to regulate transcription in neonatal cardiomyocytes , likely via interactions with the INO80 and RNA Pol II mediator complexes . In addition to identifying components of broadly expressed multiprotein chromatin machines , our analysis revealed the association of TBX20 with the essential cardiac transcription factor CASZ1 in BirA-expressing iCMs ( 25 unique peptides and 21% sequence coverage . As previously reported [34 , 35] , CASZ1 protein runs as three bands on at approximately 191kD , presumably due to post translational processing ) ( Fig 2B , S2C Fig ) . Surprisingly , this was the only developmentally-regulated cardiac transcription factor we found to interact with TBX20 in Day 7 cardiomyocytes . The low estimated cellular abundance of CASZ1 and the relatively high AP enrichment ratio ( S1 Table ) highlighted CASZ1 as a potential in vivo TBX20 interacting protein . We further confirmed the TBX20-CASZ1 interaction through reciprocal immuno-isolation of endogenous CASZ1 in cardiac nuclei from adult mouse hearts ( Fig 2C ) thus , verifying our ESC differentiation-based approach can successfully identify bona fide TBX20 interaction partners under physiological conditions . Since expression analysis and genetic fate mapping studies have shown CASZ1 is expressed only in cardiomyocytes and no other cardiac cell types [36] , and since we were unable to identify this interaction at Day 4 of iCM differentiation , these studies imply the interaction between TBX20 and CASZ1 is temporally regulated and cardiomyocyte-specific . Phylogenic analysis shows that TBX20 and CASZ1 are highly conserved across vertebrate orthologs [34 , 35] , suggesting that the TBX20-CASZ1 interaction may also be evolutionarily conserved . To confirm the interaction and to determine whether it is evolutionarily conserved , we injected X . laevis embryos with the Xenopus orthologous mRNAs of TBX20 and CASZ1 . In parallel , we co-expressed murine versions of tagged TBX20 and CASZ1 proteins in HEK293 cells . Immunoaffinity purification of TBX20 protein complexes from both of these sources , followed by immunoblotting confirms the formation of a TBX20-CASZ1 interaction in human cells and in X . laevis embryos ( S3A and S3B Fig ) . Taken together , our findings are supportive of an evolutionarily conserved role for the formation of a TBX20-CASZ1 protein complex in differentiated cardiomyocytes . To determine the biological relevance of the TBX20-CASZ1 interaction , we tested for genetic interaction between Tbx20 and Casz1 by generating mice with cardiac-specific heterozygous loss of Tbx20 and Casz1 ( Tbx20flox/+; Casz1flox/+; Nkx2 . 5Cre ) [21 , 36] . Compound heterozygous mice , hereafter referred to as Tbx20flox/+; Casz1flox/+ , were born and appeared normal . However , beginning at 4 weeks of age we observed an increased incidence of death among Tbx20flox/+; Casz1flox/+ mice ( 2 . 7% ) compared to the single heterozygotes ( 0% ) ( Table 1 ) . This effect was amplified at later timepoints , with survival rates of 90 . 5% and 62 . 5% at 8 and 16 weeks of age , respectively , compared to a 100% survival rate in the single heterozygotes . Furthermore , we did not observe any overt phenotypes in Tbx20flox/+; Nkx2 . 5Cre or Casz1flox/+; Nkx2 . 5Cre mice . Since we were able to demonstrate that loss of Casz1 does not affect Tbx20 expression in adult heart tissue and that loss of Tbx20 does not affect Casz1 expression ( S4A and S4B Fig ) , these studies are supportive of a genetic requirement for a functional interaction between TBX20 and CASZ1 . To determine the cause of the reduced survival rate we observe in Tbx20flox/+; Casz1flox/+ mice , we performed detailed physiological analysis , using echocardiography , of single and compound heterozygous mice . These studies revealed that cardiac function is significantly compromised in Tbx20; Casz1 compound heterozygotes compared to single heterozygotes . Compound heterozygotes exhibit significantly decreased ejection fraction and fractional shortening , increased left ventricular blood volume , and increased left ventricular diameter ( Fig 3A–3D; Tables 2 and 3; S2 Table; S2–S6 Movies ) . Further , Tbx20flox/+; Casz1flox/+ heterozygous mice display dilated ventricles with a striking decrease in ventricular wall thickness compared to single heterozygotes ( Fig 4A and 4B ) . These findings were observed in both male and female mice ( S3 Table ) . Thus , Tbx20flox/+; Casz1flox/+ mice display defining anatomical features of DCM that progress to cardiac failure . Collectively , these findings suggest that the genetic interaction between Tbx20 and Casz1 is essential for normal cardiac homeostasis , and perturbation of this interaction leads to DCM . One of the defining clinical features of severe DCM is an accumulation of myocardial collagen leading to interstitial fibrosis , a contributing and compounding factor in cardiac dysfunction [37–39] . To confirm that the severe cardiac dysfunction we observe in Tbx20; Casz1 compound heterozygotes is associated with advanced DCM , we examined collagen fibers and found robust collagen deposition in the interstitium of Tbx20flox/+; Casz1flox/+ hearts ( Fig 4B and 4C ) . Despite the interstitial fibrosis and severely impaired systolic function , the fact that over half of these mice survive to adulthood with some degree of cardiac function led us to hypothesize that Tbx20; Casz1 compound heterozygous cardiomyocytes undergo compensatory pathological hypertrophy . To test this hypothesis , we measured cardiomyocyte cross-sectional areas and found that MF20-positive cells in compound heterozygotes were indeed increased in size relative to controls ( Fig 4D and 4E ) . This data suggests that disrupting the TBX20-CASZ1 interaction leads to severe DCM and cardiac fibrosis . In response to this heightened cardiac stress , Tbx20; Casz1 compound heterozygote hearts appear to undergo pathological hypertrophy as an adaptive response . Our data implies the TBX20-CASZ1 interaction is essential for normal cardiac homeostasis . To define the region of TBX20 that mediates interaction with CASZ1 , we conducted immunoisolations with wild-type and deletion mutants in which either the T-Box or the C-terminus of TBX20 has been removed ( Fig 5A ) . Immunopurifications of CASZ1 in the presence of wild-type TBX20 , TBX20ΔT-box , or TBX20ΔC , show that the T-box domain is required for interaction with CASZ1 , but that the C-terminus is dispensable ( Fig 5A ) . In reciprocal studies , we find the four most amino-terminal zinc finger domains of CASZ1 are necessary for interaction with TBX20 ( Fig 5B ) . The CASZ1-interacting region of TBX20 , as well as the TBX20-interacting region of CASZ1 , are highly conserved across species implying functional relevance to these regions ( Fig 5C ) . Recently , human TBX20 mutations have been identified that are associated with DCM; however , only one of these mutations , TBX20F256I , co-segregates in a dominant manner with complete penetrance in a family with DCM [9] . Moreover , DCM was found in all affected family members reported as healthy during health assessments performed when they were juveniles . The functional relevance of the F256I mutation is further underscored by the finding that the amino acid disrupted by this mutation is 100% conserved across all TBX20 orthologs and by the observation that no F256I mutations were identified in 600 control samples [9] . Interestingly , the F256I mutation associated with DCM lies within the TBX20 T-box domain , the region we found essential for interaction with CASZ1 ( Figs 5A and 6A ) . To test if TBX20F256I perturbs the TBX20-CASZ1 interaction , we performed immunoaffinity purifications of CASZ1 in the presence of wild-type TBX20 or TBX20F256I . Interestingly , the F256I mutation significantly reduces the interaction with CASZ1 ( Fig 6A ) . These data imply that the DCM mutation F256I may contribute to the development of cardiac disease by disrupting a critical physical interaction between TBX20 and CASZ1 . To gain a structural understanding of how the F256I mutation disrupts the TBX20-CASZ1 interaction , we conducted molecular modeling of the wild-type and TBX20F256I T-box domain ( Fig 5B and 5C ) . The predicted structures were based on the range of fluctuations in the structure that occur over a period of 100 ns ( S4 Movie ) . Three regions are highlighted which show conformational changes induced by the mutation ( Fig 6B ) . Our models find F256 is not predicted to contact DNA but the conversion of phenylalanine to isoleucine at position 256 leads to steric clashes with the conserved T-box residues E258 and T259 ( Fig 6C ) . The critical functional nature of this region of TBX20 is underscored by the complete conservation of amino acids at residues F256 , E258 , and T259 across 250 members of the T-box gene family ( S5A and S5B Fig ) . Taken together , these findings imply that F256I leads to a conformational change across the surface predicted to interact with CASZ1 , and that disruption of this interaction leads to alteration in DNA binding . To determine the transcriptional consequences of TBX20F256I on the TBX20:CASZ1 interaction , we conducted transcriptional assays with TBX20 , CASZ1 and TBX20F256I alone and in combination . Results demonstrate TBX20 synergistically acts with CASZ1 and that TBX20F256I significantly diminishes transcriptional activation by TBX20:CASZ1 ( S6 Fig ) . These data together with our structural studies provide a mechanistic basis for how F256I disrupts TBX20:CASZ1 function . To identify the molecular pathways altered in DCM haploinsufficient mutant ( Tbx20flox/+; Casz1flox/+; Nkx2-5Cre ) mouse hearts , we used quantitative multiplexed mass spectrometry to identify proteins with altered abundances relative to control hearts . Proteins were extracted from nuclear-enriched mouse cardiac fractions of mutant and control ( Nkx2-5Cre ) mice in duplicate and digested in-solution with trypsin . Peptides from each sample were labeled with different isobaric tandem mass tagging ( TMT ) reagents , pooled , fractionated , and analyzed by reverse phase nanoliquid chromatography coupled to a high resolution quadrupole Orbitrap tandem mass spectrometer . Using this strategy , 3164 proteins were identified and quantified based on their respective sequenced peptides and TMT reporter ions , respectively ( S4 Table ) . To define the TMT ratio threshold for differential relative abundance , protein abundance values were compared between biological duplicates ( S7 Fig ) . For both the control and mutant replicates , the correlation of abundances was high ( R2 = 0 . 99 ) and the majority of proteins had low dispersion from a 1:1 linear curve ( S7A and S7B Fig ) , indicating low biological and technical variation . Curve-fit analysis of TMT abundance ratio histograms for the control and mutant biological duplicates showed that , on average , 90% of the ratios varied less than ±30% ( S7 Fig ) . Based on this result , a relative abundance ratio of at least ±1 . 3-fold between mutant and control mice in both replicates were used to identify a protein as differential . From the total number of quantified proteins , 175 met this criterion , of which 86 and 89 were up and down-regulated , respectively ( Fig 7C; S4 Table ) . Further verifying the role of the TBX20; CASZ1 interaction , 165 of the 175 of the proteins identified by this approach were encoded by a gene previously demonstrated to be a TBX20 target [40] ( S5 Table ) . To generate an initial picture of potentially dysregulated pathways , the known functional connectivity among differential proteins can be determined using databases of annotated pathways and protein-protein interaction . Towards this goal , known functional associations among the 175 differential proteins were scored based on the STRING bioinformatics database [41] , and the relational networks visualized in Cytoscape [42] ( Fig 7D; S8 Fig ) . A high degree of interconnectivity was observed among the differential proteins as 127 of the 175 annotated proteins had at least one other connection and each protein on average was connected to 4 . 6 neighbors . Network clustering was performed to identify subsets of highly connected proteins , which likely share similar functions . Overall , there are 10 functional clusters containing at least 3 proteins , indicated by the color-coding in Fig 6C . To identify the most significant biological processes and pathways that are perturbed in Tbx20; Casz1 hypomorphic DCM hearts , we performed comparative Gene Ontology over-representation analysis of the differentially regulated proteins using ClueGO [43] ( Fig 7E ) . Consistent with studies demonstrating an association between DCM and inflammation [44 , 45] , we found components of the pro-inflammatory response ( i . e . complement activation ) significantly up-regulated . In addition to inflammation-associated proteins , our data further revealed a dysregulation of mitochondrial proteins known to be associated with impaired cardiomyocyte contractile function in DCM [46] . In line with these findings and the observation that reduced contractile force is linked to altered glycogen metabolism and cardiomyopathy [47–49] , we found an over-representation of proteins associated with the glycogen metabolic pathway . We note that these were exclusively down-regulated proteins , represented by glycogen synthase ( Gys1 ) , glycogen phosphorylases ( Pygm/Pygb ) , and phosphorylase kinase gamma 1 ( Phkg1 ) . Interestingly , proteins involved in glycogen regulation and in myosin-dependent muscle contractility were part of the same functional cluster ( Fig 7D , yellow nodes ) ; however , their individual abundances were down- and up-regulated , respectively ( Fig 7D , circle vs . square nodes ) . Taken together , these data confirm at the protein level the DCM pathology in Tbx20; Casz1 hypomorphic DCM mice . In addition to proteins previously reported to be associated with DCM , our analysis identified a distinct set of cell-cell adhesion proteins in Tbx20; Casz1 compound heterozygous hearts that were significantly overrepresented compared to whole genome annotation ( Fig 7E; S8 Fig , yellow ) . These observations highlight the significant changes that are likely occurring in the extracellular and intracellular spaces and raise a key question- what are the signaling mediators that link these processes ? To identify potential key mediators of TBX20-CASZ1-driven DCM , we constructed a gene-linked GO network for the Cellular Component ontology ( S8 and S9 Figs ) . This network highlighted two interesting candidates , bone morphogenic protein 10 ( Bmp10 ) and thrombospondin 1 ( Thbs1 ) , the former being a TGF-beta receptor ligand and the latter having roles in cell-cell adhesion as well as ER stress response [50] . Overall , this systems-level proteome view of DCM provides potential downstream targets and pathways that may be influenced as a result of Tbx20 and Casz1 haploinsufficiency and suggests a role for cell-cell adhesion in mediating DCM .
CASZ1 is a large para-zinc finger protein of unique structure and to date , there have been limited studies on the mechanisms of how CASZ1 regulates transcription [60–62] . These types of studies have been compromised by the lack of high-affinity high-specificity mammalian CASZ1 antibodies , precluding approaches such as ChIP-seq . It further remains unclear if CASZ1 , as a para-zinc finger protein , directly binds DNA or is recruited via other transcription factors . Our structural studies favor a model by which the TBX20-CASZ1 interaction is required for DNA binding . This model predicts that the respective region of TBX20 that binds CASZ1 is near to or contributes to the DNA binding interface and has the potential to impact CASZ1 binding . CASZ1 was first ascribed a role in vertebrate cardiovascular development in Xenopus [34 , 62 , 63] . Subsequent genetic studies in mammals uncovered that like TBX20 , CASZ1 functions in the embryonic heart to control cardiomyocyte proliferation , with loss of CASZ1 leading to cardiac death by E12 . 5 [36 , 64] . Our finding that Tbx20; Casz1 compound heterozygous mice die post-natally implies that CASZ1 has a second and later role in cardiac homeostasis . This model is supported by the recent finding that mutations in CASZ1 , like TBX20 , are associated with human DCM [26] . In these studies , the Nkx2 . 5-Cre driver was used to generate mice null for Casz1 and Tbx20 . In all cases the Nkx2 . 5-Cre driver alone was used , with wild-type mice as a negative control in our physiological studies and in our quantitative proteomic studies . In contrast to previous reports [65] , we could detect no significant changes in any cardiac function in Nkx2 . 5-Cre mice relative to wild-type mice . These finding may be due to genetic background , the sex on which the Cre driver was delivered to the offspring , or environmental variability as reported for other lines [66 , 67] . Regardless of the reason for the variability , we did find the Nkx2 . 5-Cre driver had a high recombination efficiency reducing the levels of TBX20 and CASZ1 by half ( S4 Table ) . Since , reducing CASZ1 expression by half had no detectable alteration in Tbx20 expression and vice versa , our data suggests a biochemical and genetic interaction between TBX20 and CASZ1 . Our data further indicates that disruption of this complex leads to DCM in mice and humans . A previously published model of DCM , the phospholamban R9C transgenic mouse [68] , has also been studied by proteomic analysis [69 , 70] . This model exhibits impaired calcium regulation in cardiomyocytes , accompanied by decreased cardiac contractility and premature mortality [68] . The GO-associated proteome changes that we found in the haploinsufficient mice share similarities with the Phospholamban R9C mice . Specifically , both mouse models show up-regulation of actin-myosin cytoskeletal networks and down-regulation of mitochondria-associated proteins involved in fatty acid oxidation . Interestingly , proteomic analyses performed on ventricular tissues from human patients with inflammatory DCM had similar findings [71] . Yet some functional protein classes in our Tbx20-CASZ1 haploinsufficient DCM mice were distinct , including an up-regulation of the complement system and greater coverage of down-regulated proteins in glycogen metabolic processes . While we found the Tbx20-CASZ1 haploinsufficient mice have evidence of differential regulation in calcium-binding proteins , not surprisingly , the Phospholamban R9C mice have more pervasive effects on calcium-dependent signaling , such as involving ER stress responses , though it is possible that these distinctions may be due to differences in the progression of the fibrosis associated with DCM . One of the hallmarks of DCM is altered cardiomyocyte force transduction that is frequently associated with alteration in the composition or functions of intercalated discs- a cardiac-specific structure at the contact site between cardiomyocytes [72 , 73] . Here , we observed a significant mis-regulation of proteins involved in cell-cell adhesion in heart tissue from Tbx20; Casz1 heterozygous mice . Moreover , these include three proteins which are encoded by genes that when mutated are causative to DCM- TTN , DES , and PDLIM3 . Thus , our findings imply that the TBX20-CASZ1 complex acts , at least in part , to control the electrical and mechanical integration of neighboring cardiomyocytes . The observation that the TBX20F256I mutation leads to a decreased association with CASZ1 , along with the finding that patients heterozygous for a predicted TBX20 null mutation ( TBX20Q195X ) [23] also display DCM suggest that the TBX20F256I mutation may be acting in a haploinsufficient fashion . However , only two of the individuals within a single pedigree with the TBX20Q195X mutation display DCM while other individuals display a wide range of cardiac abnormalities [23] . Moreover , we have screened the Exome Aggregation Consortium ( ExAc ) reference set and have identified four variants in TBX20 in individuals that are asymptomatic . All variants lead to a premature stop codon in one of the TBX20 alleles and all would be predicted to be functionally null ( introduction of stop codons into exons 2 , 4 , 7 , and 8 ) [74] . Together , these findings imply that the function of the TBX20-CASZ1 complex in DCM is not dose dependent . Alternatively , individuals harboring the TBX20F256I mutation have a genetically sensitized background leading to a varying degree of penetrance that is determined by modifying genes that may be carried within the CASZ1 pathway . In cardiovascular disease , genetic mutations often result in varying degrees of penetrance , and in extreme examples , the presence of a disease-causing mutation can be asymptomatic [75–80] . These phenomena have often been explained by the action of genetic modifiers in which one gene mutation is causative to CHD and a second mutation modifies the effect of the first . However , more recent studies suggest an alternate or additional mechanism by which complete penetrance is achieved in human disease states by genetic variation at one or more loci [81] . In digenic inheritance , two genetic mutations are required for the clinical phenotype with either mutations alone being asymptomatic . Our findings provide an example of digenic inheritance in DCM and suggest that mutations in TBX20 or CASZ1 could lead to susceptibility to DCM but in many cases are not in themselves causative . We would envision these findings are not restricted to TBX20 and CASZ1 but rather are applicable to other genes and other forms of congenital heart disease ( CHD ) and DCM , and predict that genome sequencing of familial CHD will ultimately reveal a spectrum of additional CHD susceptibility alleles .
The Tbx20Avi allele was created by introducing the biotin acceptor peptide ( Avi ) targeting cassette , similar to our previous study [82] , in-frame to the terminal exon of Tbx20 in collaboration with the UNC Animal Models Core and the UNC BAC Core ( Chapel Hill ) . The Tbx20Avi; BirA cell line was generated by targeting a sequence containing the Avitag followed by a loxP-flanked neo cassette into the stop codon of exon 8 of a Tbx20a genomic fragment derived from a 129 Sv genomic BAC library . The targeting construct was linearized and electroporated into mouse embryonic stem cells ( ESCs ) of E14TG2a . 4 origin . Targeted ESCs were placed under 250 μg/mL G418 selection for 7–10 days and G418-resistant ESC clones ( n = 384 ) were screened for homologous recombination by Southern blot analysis . Three ESC clones were correctly targeted , and one of these clones was subsequently used to derive the Tbx20Avi/+; BirA cell line . Briefly , Tbx20Avi/+ ESCs were grown to approximately 40% confluence and transduced with 5 MOI Lenti-BirA for 8 hrs . Twenty-four hours following transduction , cells were placed under 200 μg/mL hygromycin selection for 4–5 days . Hygro-resistant Tbx20Avi/+ cells were subsequently used for cardiomyocyte differentiations . Tbx20Avi/+; BirA ESCs were maintained on gelatin-coated dishes in a feeder-free culture system and differentiated [29] in serum-free ( SF ) media according to the Keller protocol . Briefly , ESCs were trypsinized and cultured at 75 , 000 cells/mL on uncoated petri dishes in SF medium without additional growth factors for 48 hrs . Two-day-old aggregated embryoid bodies ( EBs ) were dissociated and the cells reaggregated for 48 hr in SF medium containing 5 ng/mL human Activin A , 0 . 1 ng/mL human BMP4 , and 5 ng/mL human VEGF ( all growth factors purchased from R&D Systems ) . Four-day-old EBs were dissociated and 2 x 106 cells were seeded into individual gelatin-coated wells of a 6-well dish in StemPro-34 SF medium ( Invitrogen ) supplemented with 2 mM L-glutamine , 1 mM ascorbic acid , 5 ng/mL human VEGF , 20 ng/mL human bFGF , and 50 ng/mL human FGF10 ( R&D Systems ) . Cardiomyocyte monolayers were maintained in this media for 4–5 additional days with cells typically beginning to beat 2 days after seeding onto gelatin ( total of 7–8 days of differentiation ) . For immunofluorescence of cardiomyocytes , four-day-old ES cell-derived EBs were dissociated and seeded into 8-well chamber slides precoated with 0 . 1% gelatin . Induced cardiomyocytes were fixed on day 7 of differentiation in 4% paraformaldehyde for 20 min at room temperature , washed ( 3 x 1X PBS ) , permeabilized in 0 . 1% Triton X-100 in 1X PBS for 10 min , and blocked ( 10% fetal bovine serum [FBS] , 0 . 1% Tween 20 in 1X PBS ) for 30 min . Anti-myosin heavy chain ( Abcam ) was applied overnight , followed by PBS washes ( 3 x 1X PBS ) , and incubation with goat anti-mouse Alexa 546 ( Invitrogen ) for 1 hr . Cells were incubated in DAPI ( 200 ng/mL in ethanol ) for 30 min and visualized by confocal microscopy on a Zeiss 710 . Protein preparations , conjugation of magnetic beads and immunoaffinity purification and mass spectrometry were conducted as previously reported [82] . All results are from a minimum of two independent biological replicates . Briefly , immunoisolated proteins were resolved ( ~ 4 cm ) by SDS-PAGE , and visualized by Coomassie blue . Each lane was subjected to in-gel digestion with trypsin and analyzed by nanoliquid chromatography coupled to tandem mass spectrometry as previously reported [83] . Tandem mass spectra were extracted by Proteome Discoverer ( ThermoFisher Scientific , ver 1 . 4 ) , and searched with the SEQUEST algorithm against a theoretical tryptic peptide database generated from the forward or reverse entries of the mouse UniProt-SwissProt protein sequence database ( 2013/08 ) and common contaminants ( total of 43 , 007 sequences ) . SEQUEST search results were analyzed by Scaffold ( version 4 . 6 . 1 , Proteome Software Inc ) using the LFDR scoring scheme to calculate peptide and protein probabilities . Peptide and protein probabilities thresholds were selected to achieve ≤ 1% FDR at the peptide level based on LFDR modeling and at the protein level , based on the number of proteins identified as hits to the reverse database . The spectral counts assigned to proteins that satisfy these criteria and had a minimum of two unique peptides were exported to Excel for data processing . Proteins identified by LC-tandem MS were filtered to exclude non-specific associations . Proteins were retained as specific interaction candidates if the proteins were assigned ( 1 ) at least ten spectral counts in the Tbx20Avi;BirA condition , and ( 2 ) were uniquely identified or had at least a 4-fold spectral count enrichment in the Tbx20Avi; BirA condition versus the control . Next , the subset of candidates assigned a nuclear or unknown UniProt subcellular localization were retained for calculation of enrichment index values , as previously described [31] . Briefly , the relative protein abundance within the affinity purification was calculated using the NSAF approach [84] , then normalized by each protein’s respective cellular abundance estimated in the PAX database [85] ( Mouse—whole organism , SC GPM 2014 ) . Interaction candidates were ranked by their enrichment index and the top 50 proteins were analyzed by STRING [86] for interaction network analysis . Interactions with a combined STRING score of > 0 . 4 ( medium confidence ) were retained , exported , and visualized in Cytoscape ( ver . 3 . 3 ) . Proteins within the network were assigned into broad protein functional classes based on annotations in the UniProtKB database . Western blots were probed with the following primary antibodies overnight at 4°C: mouse anti-V5 ( Invitrogen ) 1:5000; mouse anti-GFP ( JL-8 , Clontech ) 1:10000; mouse anti-HA-HRP ( Cell Signaling #2999 ) 1:1000 , mouse anti-GAPDH ( Millipore ) 1:1000 , goat anti-TBX20 ( Santa Cruz Biotechnology ) 1:600 , rabbit anti-CASZ ( Santa Cruz Biotechnology ) 1:1000 , and chick anti-BirA ( Abcam ) 1:2000 . After being rinsed , blots were rinsed in the following secondary antibodies for 1 hr at room temperature: anti-IgG2a-HRP ( Jackson Immunoresearch ) 1:10000 . Antibody-antigen complexes were visualized using an ECL Western Blotting Analysis System ( Amersham ) . Tbx20flox/+ mice were generously provided by Sylvia Evans ( UCSD ) [21] . The Casz1flox/+ mouse has been previously reported [36] . Histological sectioning and immunohistochemistry were done as reported except as noted [36] . All mice are on a mixed B6/129/SvEv/CD-1 background and all mouse experiments were performed according to the Animal Care Committee at the University of North Carolina , Chapel Hill . Cardiac function was assessed in conscious 8–11 week-old Casz1flox/+; Nkx2 . 5Cre , Tbx20flox/+; Nkx2 . 5Cre , and Tbx20flox/+; Casz1flox/+; Nkx2 . 5Cre mice ( 5–10 mice per genotype ) by thoracic echocardiography using VisualSonics Vevo 770 ultrasound system ( Visual Sonics , Inc . ) . All imaging was done by trained technicians blinded to the genotypes of the animals . Briefly , a topical hair removal agent was used on the chest and abdomen of mice . The mice placed on a warmed table in the supine position for imaging . A 30 MHz pediatric probe used to capture 2-dimensional guided M-mode views of the long and short axes at the level of the papillary muscle . VisualSonics Analytic software was used to determine mean ventricular wall and interventricular septum thickness , as well as the left ventricle diameter from at least 3 consecutive cardiac cycles . Means were used to calculate ejection fraction and fractional shortening . All statistical analysis performed using SAS JMP 10 . Statistical significance between individual groups was calculated using Student’s T-test , while significance between more than 2 groups was calculated using ANOVA . | A molecular understanding of cardiomyocyte development is an essential goal for improving clinical approaches to CHD . While TBX20 is an essential transcription factor for heart development and its disease relevance is well established , many fundamental questions remain about the mechanism of TBX20 function . Principle among these is how TBX20 mutations associated with adult dilated cardiomyopathy circumvent ( DCM ) the essential embryonic requirement for TBX20 in heart development . Here we report using an integrated approach that TBX20 complexes with the cardiac transcription factor CASZ1 in vivo . We confirmed TBX20 and CASZ1 interact biochemically and genetically , and show mice heterozygous for both Tbx20 and Casz1 die , beginning at 4 to 8 weeks post birth , exhibiting hallmarks of DCM . Interestingly , the human mutant TBX20F256I bypasses the early essential requirement for TBX20 but leads to DCM . We report here that TBX20F256I disrupts the TBX20-CASZ1 interaction , ascribing clinical relevance to this protein complex . Further , by using quantitative proteomics we have identified the molecular pathways altered in TBX20-CASZ1-mediated DCM . Together , these results identify a novel interaction between TBX20 and CASZ1 that is essential for maintaining cardiac homeostasis and imply that DCM can be inherited through a digenic mechanism . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"medicine",
"and",
"health",
"sciences",
"cardiovascular",
"physiology",
"cardiovascular",
"anatomy",
"gene",
"regulation",
"regulatory",
"proteins",
"protein",
"interaction",
"networks",
"dna-binding",
"proteins",
"animal",
"models",
"physiological",
"processes",
"model",
... | 2017 | Formation of a TBX20-CASZ1 protein complex is protective against dilated cardiomyopathy and critical for cardiac homeostasis |
The Arp2/3 complex is essential for actin assembly and motility in many cell processes , and a large number of proteins have been found to bind and regulate it in vitro . A critical challenge is to understand the actions of these proteins in cells , especially in settings where multiple regulators are present . In a systematic study of the sequential multicomponent actin assembly processes that accompany endocytosis in yeast , we examined and compared the roles of WASp , two type-I myosins , and two other Arp2/3 activators , along with that of coronin , which is a proposed inhibitor of Arp2/3 . Quantitative analysis of high-speed fluorescence imaging revealed individual functions for the regulators , manifested in part by novel phenotypes . We conclude that Arp2/3 regulators have distinct and overlapping roles in the processes of actin assembly that drive endocytosis in yeast . The formation of the endocytic actin patch , the creation of the endocytic vesicle , and the movement of the vesicle into the cytoplasm display distinct dependencies on different Arp2/3 regulators . Knowledge of these roles provides insight into the in vivo relevance of the dendritic nucleation model for actin assembly .
Dynamic networks of branched actin filaments are frequently found adjacent to membranes and appear to play a role in many cellular process ( reviewed in [1] ) . The dendritic nucleation model provides a framework to understand how networks of branched actin filaments assemble and generate a pushing force [2–4] . A key step in the formation of a branched actin filament network is the activation of the Arp2/3 complex , which nucleates a new actin filament from the side of an existing filament . This new growing filament pushes against the membrane . Arp2/3 normally exists in an inactive state and requires an activator protein to induce a large conformational change , which allows for nucleation of a new actin filament ( reviewed in [1 , 5 , 6] ) . Coronin , an Arp2/3 inhibitor , stabilizes Arp2/3 in the inactive conformation [7] . In some cases , targeting the Arp2/3 complex to a subcellular location—rather than activating it—may be the principal role for an Arp2/3-binding protein ( reviewed in [3 , 8–10] ) . Of note , with yeast actin as a substrate , the yeast Arp2/3 complex nucleated polymerization rather well in the absence of an activator [8 , 11] . Activators enhanced the activity of the Arp2/3 complex in these studies , but by a relatively small amount . Thus , yeast Arp2/3 activators may be critical for spatial control of actin assembly by targeting , rather than activating , Arp2/3 . An important challenge for the field is to understand how the activities of multiple Arp2/3-activating proteins are coordinated in vivo . Do these activators have overlapping functions or does each of these proteins have a unique role in the formation of a proper network ? We addressed this question by investigating the roles of all of the proposed Arp2/3 regulatory proteins in the assembly and movement of cortical actin patches in Saccharomyces cerevisiae . Actin patches contain five proteins with an acidic/DDW motif for binding the Arp2/3 complex: a WASp ( Las17 ) ; two type-I myosins ( Myo3 and Myo5 ) ; an Eps15 homology ( EH ) protein ( Pan1 ) ; and an actin filament–binding protein ( Abp1 ) ( Figure 1A ) [12] . The acidic/DDW region is necessary and sufficient to bind the Arp2/3 complex , and all of these proteins can bind and activate Arp2/3 in vitro [13–20] . Actin patches also contain a coronin , Crn1 , which can inhibit Arp2/3 activation in vitro [21 , 22] . The actin patch provides an excellent system to test the roles of these regulators in a single , complex , multicomponent system . In yeast , actin patch assembly and movement mediates endocytosis [23–27] . Actin patches assemble at the plasma membrane as the endocytic vesicle forms . Patches then move into the cell with the vesicle [24 , 28 , 29] . Patch formation and movement occurs rapidly , on a time scale of seconds to minutes , and depends on Arp2/3 complex and actin polymerization [29–32] . The life cycle of an actin patch is a stereotyped sequence of events that is characterized by changes in protein composition , location , and movement; this can be seen as three phases ( Figure 1B ) . In phase I , actin patches assemble at the cortex and display a limited amount of motion , as if tethered in place . Many proteins are recruited to the actin patch during this phase , including WASp/Las17 and endocytic adaptors Sla1 , Sla2 , and Pan1 . Near the end of phase I , the actin filament network begins to assemble , as indicated by the arrival of the Arp2/3 complex , capping protein ( Cap1/Cap2; CP ) , and the actin filament–binding proteins fimbrin/Sac6 and Abp1 [24 , 33] . Phase II is characterized by slow movement of the patch a short distance into the cell , away from the plasma membrane . Some patch components appear to remain at the membrane during this movement , such as Las17 and Myo5 [24 , 34] , whereas others appear to move into the cell , including Sla1 , Sla2 , Pan1 , Arp2/3 , CP , and Abp1 . At the end of phase II , several proteins are lost from the patch , including Sla1 , Sla2 , and Pan1 . During phase III , patches contain actin filaments and actin-binding proteins , and they undergo more rapid and lengthy movement into the cell before disappearing [24 , 33] . In this study , we used high-speed video microscopy , coupled with computer-assisted patch tracking and quantitative motion analysis , to study the effect of mutations in genes for Arp2/3 regulators on the assembly and movement of actin patches marked by green fluorescent protein ( GFP ) -labeled components . The methodology allowed the study of large numbers of patches , which revealed novel phenotypes in the mutants . These studies provide new evidence about the function of each of the Arp2/3 regulatory proteins and reveal that these regulators have some roles that are distinct from each other , as well as some that are overlapping .
Las17 , the yeast WASp protein , binds to and activates Arp2/3 via an acidic/DDW region at its C terminus ( Figure 1A ) [11 , 19] . To determine the role that this region plays in actin patch motility , a C-terminal truncation that removed the acidic/DDW region was generated . All of the mutations described in this study were made at the endogenous locus and were examined in haploids , where the mutant allele was the only allele of the gene present . The mutation removed the Arp2/3-binding region but not other known domains , including the WH2 domain that binds actin monomer . The truncated Las17 localized to actin patches by GFP tagging ( data not shown ) . We tracked the positions of hundreds of patches over time and then used several forms of quantitative motion analysis to assess the effect of mutations . The methodology is described fully in Materials and Methods . Mean squared displacement ( MSD ) plots were one form of analysis , and examples of how such plots were generated and used to monitor each phase of the patch life cycle are presented in Figure 1C–1E . We first examined the effects of this mutation on phase I and II of patch motility using Sla2-GFP labeling ( Video S1 ) . MSD plots , generated from displacement data of individual patches aligned at the start of their lifetimes and then averaged , indicated a defect in the behavior of las17Δacidic patches during phase I or II . In contrast , MSD plots of the same data , averaged after aligning individual patch curves at the end of their lifetimes , showed no defect in the mutant ( Figure 2A ) . To examine phases I and II more closely , we quantitated directly the frequency and timing of early events . The amount of time that patches spent at their origin , before moving off the membrane , was greatly increased in las17Δacidic mutant cells ( Figure 2C ) , which can account for the observed defect in MSD plots aligned at the start ( Figure 2A ) . The percentage of patches that left the membrane , corresponding to transition from phase I to phase II , was essentially normal ( Figure 2B ) . To determine if the timing of the recruitment of actin was normal in las17Δacidic cells , we determined the time between the appearance of Sla2 and of Abp1 , a marker of actin filaments . The arrival of actin filaments to the patch was delayed in las17Δacidic cells ( Figure S1 ) . To examine patch movement during phase II , we isolated tracking data for patches—after they moved off the membrane—with Sla2-GFP labeling . In this analysis , the las17Δacidic cells were normal ( Figure 2D ) . Taken together , the data support a model where the acidic domain of Las17 is critical for the duration of phase I but not for the ability of the patch to leave the membrane or for its initial movement off the membrane . The effect of the las17Δacidic mutation on the movement of actin patches during phase III was examined using Abp1-GFP ( Video S2 ) . Patch movement was decreased in the mutant and was assessed with MSD plots aligned at the start or the end of patch lifetime ( Figure 2E ) . Decreased movement at the start of Abp1-GFP patch life can result from a prolongation of phase I , as seen with Sla2-GFP . Indeed , as expected , the time that Abp1-GFP patches remained at their origin was prolonged ( Figure 2G ) , and the percentage of patches that moved away from their origin was slightly reduced in las17Δacidic cells ( Figure 2F ) . MSD plots aligned at the end of patch lifetimes showed a decrease in movement , as noted above ( Figure 2E , right ) . We tested the phase III movement directly by isolating tracking data for patches after they moved 200 nm from their membrane origin . Patches of las17Δacidic cells showed decreased movement in MSD plots aligned at the beginning and end of this movement ( Figure 2H ) . In previous studies in yeast , GFP-labeled WASp proteins remained on the membrane , exhibiting phase I behavior , and did not move off the membrane [24 , 33 , 35] These results raised the question of how the Arp2/3 complex and actin assembly might power the movement of endocytic vesicles into and about the cytoplasm . We found similar results with GFP fused to the C terminus of Las17 . However , we find that neither C- nor N-terminal fusions of GFP to Las17 , which were expressed from the endogenous las17 locus , are fully functional for actin patch motility , especially for the late movements of phase III ( Figures S2 and S3 ) . Interestingly , when we overexpressed a novel N-terminal fusion from the GAL1 promoter , approximately one-third of GFP-Las17–labeled patches exhibited substantial movement away from their origin . They moved into and about the cytoplasm , as seen in confocal movies taken at the equator of the cell ( Figure 2I–2J , Videos S3–S4 ) . Although this method is technically challenging , we have observed examples of colocalization of these particles moving into the cytoplasm with Abp1-tdimer2 ( Video S5 ) . However , this GFP fusion protein may also not reveal the normal localization of the Las17 protein . Actin patch motility in strains expressing this fusion was impaired , especially during phase III ( Figure S4 ) , indicating that Las17 function was not normal in these strains . In addition , cells expressing GFP-Las17 from the GAL1 promoter have a distinct population of GFP-Las17 particles moving freely inside the cell ( Figure 2K , Video S6 ) . These particles have a markedly greater lifetime than do actin patches , so the particles may correspond to membranous vesicles moving in the cytoplasm reported in previous studies [33 , 36 , 37] . The C termini of fungal type-I myosins , including Myo3 and Myo5 of budding yeast , have an acidic/DDW sequence that binds Arp2/3 complex ( Figure 1A ) [14 , 17 , 38 , 39] . These type-I myosins are capable of activating Arp2/3 in vitro when they are artificially connected to a WH2 domain or in the presence of the WH2-containing protein , Vrp1 , the yeast WASp-interacting protein ( WIP ) [13 , 15] . To determine the role of type-I myosins in actin patch motility , we examined the effect of the deletion of the myo3 and myo5 genes on Abp1-GFP dynamics . myo3Δ cells were normal , by MSD analysis ( Figure 3E ) , with no defects in the percentage of patches leaving the origin or the time that patches spent at the origin ( Figure 3F–3G ) . In contrast , myo5Δ cells had substantial defects in Abp1-GFP motility ( Figure 3E ) . The median lifetime of Abp1-GFP patches in the absence of Myo5 was found to be increased , and the distance of the movement of Abp1-GFP patches was decreased , in MSD plots with patch tracks aligned at the beginning and at the end of their lifetimes . The time that Abp1-GFP patches spent at the origin was increased in myo5Δ cells , and the percentage that moved away from the origin was decreased ( Figure 3F–3G ) . The effect of the loss of Myo5 on the early phases of actin patch motility was also examined using Sla2-GFP labeling . MSD plots of Sla2-GFP movement , aligned at the beginning and at the end of patch lifetimes , revealed decreased movement with an increase in median lifetime ( Figure 3A ) . This defect was due in part to a decrease in the frequency with which Sla2-GFP patches left the origin and an increase in the time the patches spent at the membrane ( Figure 3B–3C ) . To determine if the phase II movement of patches in myo5Δ cells was normal when they did leave the membrane , we isolated and analyzed movement of Sla2-GFP–labeled patches from the membrane . Phase II patch movement was found to be decreased in the myo5Δ cells by MSD analysis ( Figure 3D ) . Observations of cell growth and endocytosis have indicated that Myo3 and Myo5 have overlapping functions , with apparent redundancy in some cases [40 , 41] . To address this question , we examined Abp1-GFP patch behavior in myo3Δ myo5Δ double mutants . MSD curves showed a severe defect in patch motility , with a nearly complete loss of movement away from the origin and an increase in median patch lifetime , when compared to wild-type or single-mutant cells ( Figure 3E ) . The percentage of Abp1-GFP patches leaving the membrane was very small , and the lifetime of patches at the origin was greatly increased ( Figure 3F–3G ) . The number of patches leaving the membrane was not sufficient to permit a direct analysis of phase III movement . Taken together , these results show that although the function of Myo5 is distinct from that of Myo3 , the proteins do have some functions in common . The results are consistent with the previous observation of endocytosis defects in myo5Δ but not myo3Δ single mutants , and with multiple observations that myo5Δ myo3Δ double mutants are more severely affected in growth and endocytosis than either single mutant alone [40 , 41] . To determine if the actin patch phenotypes of the type-I myosin null mutants , myo3Δ and myo5Δ , result from loss of Arp2/3 interaction , we truncated the acidic/DDW region of each type-I myosin . For Myo5 , loss of the acidic/DDW region had no effect on phase I and II movement except for a small , statistically significant increase in movement in MSD plots of Sla2-GFP patch tracks aligned at the end of their lifetime ( Figure 3H–3J ) . For phase III movement , which was assessed with Abp1-GFP , MSD curves showed a slight leftward shift when tracks were aligned at the start , but no change when aligned at the end ( Figure 3K ) . Abp1-GFP patches remained at the origin for slightly less time in the mutant cells ( Figure 3M; p = 0 . 0003 ) , but the frequency with which patches moved away from the origin was normal ( Figure 3L ) . These results were a surprising contrast to those for the myo5Δ null mutant , which had a strong defect in phase III and significant ones in phases I and II . For Myo3 , a similar truncation of the acidic/DDW region had no effect on phase I or II of actin patch movement , which were examined with Sla2-GFP ( Figure 3H–3J ) , or on phase III , which was examined using Abp1-GFP ( Figure 3K–3M ) . These results were largely similar those for the myo3Δ null mutant and thus not surprising . To determine if the absence of a phenotype in the single-mutant strains was a result of functional redundancy , we examined cells carrying both truncation mutations . Patch movement was again remarkably unaffected . The double mutant produced results very similar to those for the Myo5 single mutant for all three phases of patch life ( Figure 3H–3J ) . During phase III , actin patches of double-mutant cells were slightly different from those of the Myo5 single mutant ( Figure 3K–3M ) , but this difference barely achieved statistical significance . Overall , the Arp2/3-binding regions of Myo3 and Myo5 were dispensable for function in this otherwise wild-type genetic background , in striking contrast to the effect of the complete loss of one or both proteins . Several lines of evidence suggest that the type-I myosins might function as a complex with WASp/Las17 and WIP/Vrp1 , with overlapping function among the three Arp2/3-binding regions [14 , 17 , 42] . To test this possibility , we examined actin patch motility in haploid strains carrying a las17Δacidic mutation in combination with myo3Δacidic and/or myo5Δacidic mutations . First , we examined phases I and II with Sla2-GFP labeling . Patch movement in las17Δacidic myo3Δacidic cells was similar to that in las17Δacidic cells during phases I and II ( Figure 3N ) . Only small differences of marginal statistical significance were observed for the time that patches remained at the origin and for the frequency with which patches left the membrane ( Figure 3P and 3O ) . In contrast , cells lacking the acidic/DDW regions of WASp/Las17 and Myo5 showed substantially greater defects than cells lacking only the acidic/DDW region of WASp/Las17 . Patch movement in the double mutant was decreased , by MSD analysis , compared with that in the single mutant ( Figure 3N ) . This defect resulted in large part from a decrease in the frequency with which patches left the origin to begin phase II movement ( Figure 3O ) . In addition , patches in las17Δacidic myo5Δacidic cells spent more time at their point of origin before disappearing or moving away than did those of las17Δacidic cells ( Figure 3P ) . In triple-mutant cells , which lack the acidic domains of WASp/Las17 , Myo3 , and Myo5 , Sla2-GFP patch motility was decreased compared with double mutants ( Figure 3N ) . The frequency with which triple-mutant patches left the origin was only slightly lower ( Figure 3O ) , and the time that patches remained at the origin was similar ( Figure 3P ) when compared with las17Δacidic myo5Δacidic cells . Phase III patch movement was examined in these mutants using Abp1-GFP . The loss of the acidic domain of Myo3 had little effect on the motility of Abp1-GFP patches in las17Δacidic cells by MSD analysis ( Figure 3Q ) . The frequency with which patches moved away from the origin and the time patches spent at the membrane were very similar ( Figure 3R and 3S ) , similar to the results with Sla2-GFP labeling . If anything , loss of the acidic domain of Myo3 suppressed the phenotype of increased time that Abp1-GFP patches spent at the membrane in las17Δacidic cells ( Figure 3S , p = 0 . 0007 ) . In contrast , las17Δacidic myo5Δacidic double-mutant cells had greater defects in Abp1-GFP patch motility than did las17Δacidic cells ( Figure 3Q ) . The percentage of patches that moved away from their origin was decreased ( Figure 3R; p = 0 . 04 ) , and the time that patches spent at the membrane prior to moving away was increased ( Figure 3S; p = 0 . 0004 ) . Patches in las17Δacidic myo5Δacidic cells still retained a measurable level of dynamics and movement , so we asked if Myo3 was important in this context . Indeed , Abp1-GFP patch motility was decreased in las17Δacidic myo3Δacidic myo5Δacidic triple-mutant cells compared to las17Δacidic myo5Δacidic cells , with a nearly complete loss of motility ( Figure 3Q ) . The percentage of patches that moved away from the origin was less , and the time that patches remained at the origin was greater ( Figure 3R–3S ) . Together , the results indicate that the Arp2/3-binding regions of WASp and the type-I myosins do have a significant level of functional redundancy , consistent with the notion that they may act in a complex in which any and all of the three Arp2/3 complex interactions can be important . In previous studies suggesting the existence of such a complex , WIP , known as verprolin/Vrp1 in yeast , was also found in biochemical association . We found that a WIP null mutant , vrp1Δ , had essentially no patch movement , with Sla2-GFP or Abp1-GFP labeling ( unpublished data ) , in agreement with another study [15] . Actin patches did still form in all of these mutants , including the WIP null mutant and the WASp/type-I myosin triple mutant , showing that actin filaments can still polymerize but not with the dynamic control needed to achieve movement . The endocytic adaptor protein Pan1 has an acidic/DDW region for binding the Arp2/3 complex , along with two EH domains , a coiled-coil region , and a WH2 domain [43] . Pan1 is essential for viability in yeast , and because the loss of endocytosis does not appear to be lethal , this suggests that Pan1 may have other functions . To investigate the possible importance of Pan1′s interaction with the Arp2/3 complex in actin patch dynamics , the acidic/DDW region of the Pan1 protein was removed ( Figure 1A ) . The PAN1 gene was mutated at its endogenous locus in a diploid strain and tetrad dissection produced haploid mutant segregants that grew well ( Protocol S1 ) . In haploid mutant pan1Δacidic cells labeled with Sla2-GFP , the time that patches remained at their origin was slightly increased ( Figure 4C ) , and MSD plots of tracks aligned at the start of their lifetimes showed decreased movement ( Figure 4A ) . The percentage of patches leaving the origin was normal ( Figure 4B ) , as was MSD analysis with curves aligned at the end of their lifetimes ( Figure 4A ) . Examining only the data for movement away from the membrane in phase II , we found that patch movement was also normal in the mutant ( unpublished data ) . Thus , only the earliest stages of patch dynamics were affected by removing the Arp2/3 binding region of Pan1 . WASp / Las17 was also important in the early stages of the actin patch life cycle , as described above , so we combined the pan1Δacidic and las17Δacidic mutations . Double-mutant haploid cells had a more severe defect than did either single mutant , in terms of the time that patches spent at their origin ( Figure 4C ) and MSD analysis with curves aligned at the start ( Figure 4A , left ) . The percentage of patches leaving the origin was decreased slightly ( Figure 4B ) , and MSD analysis with curves aligned at the right was normal , similar to the single mutants ( Figure 4A , right ) . This enhanced phenotype in the double mutant is consistent with synthetic interactions between pan1 and las17 mutations in terms of cell growth [42] . With Abp1-GFP labeling to examine phase III behavior , pan1Δacidic single-mutant cells displayed a defect in MSD plots aligned at the left , no defect in MSD plots aligned at the right ( Figure 4D ) , a decrease in the percentage of patches that moved away from the origin ( Figure 4E ) , and a slight increase in the time that patches remained at the origin before moving away or disappearing ( Figure 4F; p = 0 . 0005 ) . These results are consistent with the phase I and II results for this mutant , based on the Sla2-GFP labeling result above , and they indicate that phase III patch behavior was normal in the pan1Δacidic single mutant . For double-mutant las17Δacidic pan1Δacidic cells labeled with Abp1-GFP , the early phases of actin patch dynamics were similar to those of las17Δacidic cells , based on the early time points in MSD plots aligned at the left . The time that patches spent at the membrane and in the percentage of patches that moved away from the origin was also similar ( Figure 4E and 4F ) . However , double-mutant cells showed an increase in phase III movement , compared to las17Δacidic cells , at later time points in MSD plots , aligned at the start or end of patch life ( Figure 4D ) . This rescue is the result of a combination of an increase in the percentage of patches making movements beyond 200 nm and an increase in the movement during phase III ( unpublished data ) . Thus , in this respect , the pan1Δacidic mutation suppressed the phenotype of the las17Δacidic mutation , suggesting that Pan1 and Las17 have opposing actions on Arp2/3 complex in this later phase of patch lifetime . Abp1 is present at the patch just before movement of the patch away from the membrane to the end of its life . Abp1 has two acidic/DDW regions that bind the Arp2/3 complex ( Figure 1A ) and an ADFH domain that binds F-actin [44] . In early studies , abp1Δ null mutants were nearly normal in many respects , but the abp1Δ mutation has been found to have genetic interactions with sla1Δ , sla2 , and las17Δ mutations [45] . Actin patches of abp1Δ mutants , labeled with Sla1-GFP , showed an increase in the distance that Sla1-GFP moved from the membrane and an increase in Sla1 and Sac6 lifetimes [23] . We found that Sla2-GFP patches of abp1Δ cells remained longer at their origin ( Figure 5C; p = 0 . 0002 ) , and the percentage of patches that left their origin was normal ( Figure 5B ) . By isolating the data for patch movement away from the membrane in phase II , we found that initial movement of patches away from the membrane was normal ( Figure 5D , left ) and that , remarkably , the final extent of movement was greatly increased ( Figure 5D , right ) . Consistent with these observations , MSD analysis of the complete tracks showed an increase in the median lifetime of the patches and greater movement away from the origin ( Figure 5A ) . One explanation for the increase in median lifetime and movement is that Sla2-GFP remains on the patch for a longer time in the absence of Abp1 . Phase III of actin patch movement in abp1Δ cells was examined using Cap1-GFP as a marker; the behavior of Cap1-GFP patches essentially mirrors that of Abp1-GFP patches in wild-type cells [23 , 33] . The time that patches spent at their origin was increased ( Figure 5G ) , and the percentage of patches moving away from the membrane was normal ( Figure 5F , p = 0 . 08 ) , as seen with Sla2-GFP labeling above . MSD plots of the complete lifetime were decreased in abp1Δ cells , for curves aligned at the start of their lifetime , but not at the end ( Figure 5E ) . We isolated the data for patch movement beyond 200 nm , and the results for abp1Δ cells were identical to those for wild-type cells ( Figure 5H ) . The Cap1-GFP results indicate a delay in the initiation of patch movement with normal movement after leaving the membrane for the abp1Δ mutant . Considering all the data , for Sla2-GFP and Cap1-GFP labeled patches , Abp1 appears to be important to promote patch assembly before leaving the membrane , but not for patch movement after leaving the membrane . In addition , Abp1 appears to play a role in the removal of Sla1 and Sla2 from the patch after the initial movement into the cytoplasm . Coronin/Crn1 has been shown in vitro to inhibit the activity of the Arp2/3 complex for nucleation of actin polymerization [21] . To determine if coronin has such a role in actin patch dynamics , we examined crn1Δ null mutant cells . Phases I and II of patch motility , examined with Sla2-GFP , were nearly normal in most respects , with only a slight decrease in the MSD plot , a slight increase in median lifetime ( Figure 6A , left ) , and a slight increase in time at the origin ( Figure 6C; p = 0 . 001 ) . With Abp1-GFP labeling of patches , the time spent at the origin was increased but not the probability of leaving the origin ( Figure 6F and 6G ) , similar to the results with Sla2-GFP . Together , these results suggest a minor delay in the transition from phase I to II when coronin is absent , suggesting that coronin promotes the ability of patches to move off the membrane . Phase III dynamics in crn1Δ mutant cells , examined using Abp1-GFP , showed a rightward shift in the MSD plot with curves aligned at the start for the mutant ( Figure 6E , left ) . The plot extended to greater distances and longer times , suggesting that patches moved farther and longer in the absence of coronin . In support of this idea , when the curves were aligned at the end of patch lifetime , the MSD plot showed increased movement of crn1Δ patches ( Figure 6E , right ) . Isolating the data for movement of Abp1-GFP patches in phase III , crn1Δ patches showed more movement by MSD analysis , with curves aligned at the start or end of their lifetime ( Figure 6H ) . Thus , coronin appears to inhibit the movement of patches after they move away from the membrane and into the cytoplasm .
Early in the life of an actin patch , a series of proteins are recruited to a location on the plasma membrane , assembling the machinery required for endocytosis [24] . During this time , the actin patch appears to be tethered or corralled , undergoing random movements confined to a small area [24 , 33] . Assembly of a dendritic network of actin filaments proceeds [24 , 33] , which appears to be needed for the movement of the actin patch and endocytic vesicle away from the membrane . Our results show that , on some level , essentially all of the Arp2/3 regulatory proteins play a role in these initial phases of the process . Mutations in these proteins alter the amount of time that patches persist at the site of their appearance , undergoing tethered movement , which is termed phase I . For the WASp Las17 and the EH domain protein Pan1 , mutations of the Arp2/3-binding region prolonged the duration of phase I , as did null mutations for the type-I myosin Myo5 , coronin , and Abp1 . Mutations in the Arp2/3-binding motif of Las17 delays the appearance of Abp1-GFP at patches , suggesting that although the other Arp2/3 activators can eventually assemble an actin filament network , they do so less efficiently without Las17 . Combinations of Arp2/3-binding mutations reveal the potential for overlapping function among the regulators . For example , combining the las17Δacidic mutation with the pan1Δacidic , the myo5Δacidic , or the myo5Δacidic myo3Δacidic mutations further prolonged the time for patch assembly in phase I . Arp2/3 regulators are likely required at this stage to ensue the proper targeting and activity level of the Arp2/3 complex to generate a functional actin network for movement . In their absence , the network may not have the proper branch density and filament length required to initiate invagination and/or movement . The first nonrandom movement of an actin patch is a short one away from the plasma membrane into the cytoplasm , which has been hypothesized to represent membrane invagination to form an endocytic cup [24] . Complete deletions of the type-I myosin genes had the greatest effect on this movement . Loss of Myo5 decreased the probability that a patch would undergo phase II movement , and it lessened the extent of movement per se . Loss of both Myo3 and Myo5 resulted in almost no movement away from the cortex , supporting a model that type-I myosins are needed to power this movement prior to vesicle scission [15] . Given the severity of the phenotypes in the type-I myosin null mutants , we were surprised to find that truncation of the type-I myosin Arp2/3-binding regions had little or no effect . Myo3 and Myo5 have been suggested to function in coordination with or in parallel with WASp/Las17 , perhaps as a multisubunit complex [14 , 17 , 42] . When we combined Arp2/3-binding mutations of WASp/Las17 with those of type-I myosins , the phenotype was enhanced , consistent with that view . In las17Δacidic myo5Δacidic and las17Δacidic myo3Δacidic myo5Δacidic cells , actin patches failed to move away from the origin . Verprolin/Vrp1 , the yeast homologue of WIP was suggested to function as a scaffold or central component of such a multisubunit complex [14] . In support of that view , loss of Vrp1 caused a complete loss of actin patch movement [15] . In our hands , the loss of Vrp1 caused an almost complete loss of inward movement in plots of MSD ( unpublished data ) . Many Arp2/3 regulators are dispensable for phase II movement . Deletion of the acidic domain of Las17 or Pan1 , independently or together , had no effect . The loss of neither coronin nor Abp1 affected this movement . Loss of Abp1 did result in Sla2-GFP–labeled patches moving farther away from the cortex than normal , which supports the view that Abp1 may help to remove early components such as Sla1 and Sla2 from the patch [23] . Phase II may not depend on precise regulation of Arp2/3 localization or activation , because no single Arp2/3 binding mutant had a detrimental effect on this movement . However , previous studies of the capping protein suggest that the regulation of the number of free barbed ends is critical for normal movement during phase II . Furthermore , in the absence of the Arp2/3-binding regions of Las17/WASp and the type-I myosins , phase II movement fails to occur . This failure in phase II may reflect a requirement for Arp2/3 activation during phase II movement itself or it may also reflect a failure to establish a proper network during the phase I . In addition , the motor activities of the type-I myosins appear to play some role in the in initial movement of patches away from the membrane [15] . After the actin patch departs the plasma membrane , it moves faster and farther , through the cytoplasm . When the Arp2/3-binding region of WASp/Las17 was truncated , actin patches had a severe defect in phase III movement , which is important in light of the relative normalcy of phase II movement for this mutant . The simplest interpretation of these results is that WASp/Las17 remains bound to the endocytic vesicle as it moves through the cytoplasm , analogous to the situation for the movement of pathogens in cells or beads in cell extracts . GFP-WASp fusions at the N and C termini were not seen to leave the membrane in this manner in previous studies in budding and fission yeast , including some in our lab [24 , 33 , 35] . We observed that tagging WASP/Las17 at either end with GFP results in defects in actin patch motility reminiscent of those seen in las17Δacidic strains ( Figures S2–S4 ) . We also observed that overexpressed Las17 , tagged at the N terminus , did leave the membrane and move about the cytoplasm , in the manner of actin patches and endocytic vesicles ( Figure 2I–2J ) . To be fair , the fluorescence of these foci of GFP-Las17 was weak , and Abp1-tdimer2 patch movement in these cells was defective . Therefore , it remains a possibility that Las17 may not normally leave the membrane with patches . Our results with coronin address this question of how the patch moves through the cytoplasm . The coronin null mutant showed increased patch movement during phase III , which is consistent with the prediction from biochemical studies that coronin inhibits the activity of Arp2/3 complex . Yeast Arp2/3 complex is highly active in the absence of any activator when actin from yeast is used [11] , so an inhibitor , namely coronin , may be quite important in this system . The loss of the capping protein resulted in a specific defect of phase III movement [33] , which is also consistent with the model . Together , these results with WASp/Las17 , coronin , and capping protein provide evidence that Arp2/3-mediated actin assembly powers the movements of phase III , which supports the relevance of the dendritic nucleation model for this process . Several key components of the dendritic nucleation model are known to be present on patches during phase III movement , including actin , the Arp2/3 complex , capping protein , and now WASp/Las17 [23 , 24 , 30 , 33] . The location of actin filament nucleation and the orientation of the actin filaments during actin patch assembly and movement is an important question . When endocytosis is blocked in sla2Δ cells , actin “comet tails” appear at the plasma membrane , with growing barbed ends oriented toward the membrane [24] . However , it is unclear where actin filaments are nucleated when endocytosis is proceeding normally , especially during invagination and during movement after the patch leaves the plasma membrane . No actin connection between the plasma membrane and a phase II or III patch can be seen . If filaments growing at the plasma membrane were responsible for this movement , the newest parts of the network would be at the membrane , and would likely be relatively bright when monitored by fluorescent fusion proteins . This is not the case . Previous results show that actin cables do not drive patch movement in the cytoplasm [33] . We therefore favor a model where actin filaments are being nucleated at the surface of the endocytic vesicle , during and after invagination . We observed some longer-lived GFP-Las17–labeled particles moving in the interior of the cell without first seeing them leave the plasma membrane . For actin patches , this type of observation was very rare . Whether these GFP-Las17 particles represent a later stage of actin patches , endocytic vesicles , or another cellular compartment remains to be determined . The character of the movement of these particles is reminiscent of what has been seen by studying the membrane receptor Ste2 or the lipid dye FM4–64 , which follow endocytic trafficking pathways [33 , 36] . In previous studies , the movement of Ste2-GFP particles was impaired when WASp/Las17 was truncated or when Lsb6 , a Las17-binding protein , was absent [36 , 37] , which also supports the hypothesis that Las17 is needed to drive Ste2 vesicles .
The strains used in this study are listed in Table S1 and their construction is described in Protocol S1 . For each genotype , two or three mutant and wild-type haploid segregants from a heterozygous diploid were tested . For each segregant , movies were collected from eight different cells . Strains were grown overnight in YPD at 25 °C or 30 °C to an optical density at 600 nm ( OD600 ) of 0 . 1–0 . 5 . Cells were harvested by centrifugation at 82g , suspended in SD-complete media , placed on 2% agarose pads made with SD-complete , and covered with a number 1 coverslip as described [33] . GFP fluorescence movies were made using a spinning-disc confocal microscope system consisting of an upright microscope ( BX52 , Olympus ) with a PlanApo100X 1 . 4 NA oil immersion objective , a CSU10 Yokogawa spinning disc head ( Solomere Technology ) , and an intensified charge coupled device ( CCD ) video camera ( XR Mega10 S30 camera , Stanford Photonics ) . Two color images were collected sequentially using a LMM5 Laser Merge Module ( Spectral Applied Research ) , a multipass dichroic mirror , and emission filters in a Lambda 10–3 high-speed filter wheel ( Sutter Instruments ) . The temperature was maintained at 30 °C . Large budded cells were selected for observation . Images were collected from a single focal plane at the equator of the cell . Abp1-GFP , Abp1-tdimer2 , Sla2-GFP , and GFP-Las17 movies were collected at frame rates of 5/s , 5/s , 2/s , and 2/s , respectively . For quantitative motion analysis , 200 consecutive frames were collected . Images were collected using QED In Vivo software ( Media Cybernetics ) , except for the experiments in Figure 4D–4F and Figure 2I–2K , which were collected using Piper Imaging . GFP-labeled actin patches in movies were tracked using previously described software [31] . The experimenter verified every patch track using an ImageJ ( National Institutes of Health ) plug-in that overlaid the position information from the tracking software onto the original movie , and necessary corrections were made by hand . The data were then imported into an Excel spreadsheet ( Microsoft ) . Tracks were retained for analysis if they met the following criteria: ( 1 ) The patch was observed for at least 30 frames ( 6 s for Abp1-GFP and 15 s for Sla2-GFP ) . ( 2 ) The patch originated near the cortex . ( 3 ) The patch disappeared during the movie . ( 4 ) The patch was readily distinguished from other patches . These criteria enrich our sample for patches whose entire lifetime is captured in the movie . Zero time for a track was defined as the time at which the patch was first observed , and the end point was defined as the last time point at which the patch was observed . The character and extent of the motion of GFP-labeled patches was analyzed first with plots of MSD versus time . To calculate MSD , the square of the distance of each patch from its origin was calculated at each time point . The squared displacement values were averaged for all the patches from all the cells of each segregant of a particular genotype . The degree of variation among individual tracks was high , as seen in previous studies , so we averaged the data for 150 to 542 patches for each genotype . To test for statistical significance , MSD curves from each segregant of one genotype were first compared by analysis of variance ( ANOVA ) to ensure that they yielded the same result . A grand MSD versus time plot was then generated , averaging the data for all of the patches from all of the segregants of a given genotype within one experiment . Before averaging , the squared displacement values for each patch were aligned at the beginning or end of the track lifetime . Examples of how a set of data was aligned in the two ways before averaging are shown in Figure 1C . Plots aligned “at the start” or “on the left” provide information about the behavior of patches in the early part of their life . These curves are truncated at the time when 50% of the patches have disappeared , representing the median lifetime . Plots aligned “at the end” or “on the right” allow for a better understanding of the motion of GFP-labeled actin patches later in their lifetime , away from the origin . Comparisons between genotypes were evaluated using Student's t-test . The patch tracking data were analyzed in additional ways . For individual patch track data of X-Y position and time , we determined if a patch left its origin and how long it remained at its origin before it moved away or disappeared . To analyze “phase II only” data after initiation of movement , the data for Sla2-GFP patches after they moved away from the origin were selected . MSD versus time plots were generated with these data . To analyze “phase III only , ” we selected the data for Abp1-GFP labeled patches after they moved more than 200 nm away from the origin . MSD plots were generated . | A branched network of growing actin filaments , pushing against a membrane , provides the force for certain cellular movements . The Arp2/3 complex plays a central role in this process by generating new filaments and branch points . A number of proteins bind to and , in some cases , regulate Arp2/3 . It is important to determine , in the cell , the precise roles of each of the many Arp2/3 regulators in generating actin networks during a complex , multistep , cellular movement . In yeast , endocytosis occurs at the plasma membrane in association with the assembly and movement of cortical actin patches , which contain six Arp2/3 regulators . We have used the actin patch as a model system to determine the specific roles of these regulators during patch assembly and movement . We used high-speed video microscopy , coupled with computer-aided particle tracking , to monitor the movement of fluorescently labeled actin patches in cells with one or more mutations of the Arp2/3 regulators . The sensitivity of this technique allowed us to identify previously unappreciated functions for Arp2/3 regulators and to assign each of the regulators a specific role during actin patch assembly and movement . Our results demonstrate that Arp2/3 regulatory proteins play overlapping roles at certain stages of actin patch movement , but distinct roles at other stages . In addition , our results provide new insight into how the assembly of an actin filament networks powers the movement of endocytic vesicles away from the membrane . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"cell",
"biology"
] | 2008 | Distinct Roles for Arp2/3 Regulators in Actin Assembly and Endocytosis |
Binocular vision requires an exquisite matching of projections from each eye to form a cohesive representation of the visual world . Eye-specific inputs are anatomically segregated , but in register in the visual thalamus , and overlap within the binocular region of primary visual cortex . Here , we show that the transmembrane protein Ten_m3 regulates the alignment of ipsilateral and contralateral projections . It is expressed in a gradient in the developing visual pathway , which is consistently highest in regions that represent dorsal visual field . Mice that lack Ten_m3 show profound abnormalities in mapping of ipsilateral , but not contralateral , projections , and exhibit pronounced deficits when performing visually mediated behavioural tasks . It is likely that the functional deficits arise from the interocular mismatch , because they are reversed by acute monocular inactivation . We conclude that Ten_m3 plays a key regulatory role in the development of aligned binocular maps , which are required for normal vision .
The functional capabilities of the central nervous system are critically dependent on highly specific patterns of neural connectivity that are established during early development . In many parts of the brain , axonal projections form a topographic map whereby relative spatial relationships between afferent and target fields are maintained . This spatial mapping is thought to be important in maintaining the integrity of information at successive levels of processing . Molecular gradients have been shown to play a key role in establishing normal topography in the developing visual system [1–5] . Binocular vision places an additional requirement on map formation , because the projections from each eye need to be both topographic and also aligned with the projections from the other eye . Although there have been substantial advances in our understanding of the mechanisms that control whether retinal ganglion cells ( RGCs ) project ipsilaterally or contralaterally [6–10] , the cues that allow these projections to form aligned binocular maps within their targets have remained elusive , despite attempts to identify them [11] . The only family of axonal guidance molecules with a claim on a role in eye-specific mapping is the EphA family of receptor tyrosine kinases and their ligands , the ephrinAs [12 , 13] . However , EphA–ephrinA interactions also play a major role in the establishment of topography of contralateral visual projections to central targets [14–18] , which complicates the interpretation that these molecules have a role in eye-specific patterning . Deletion of an eye-specific mapping molecule should disrupt the projections of one eye , but not that of the other , thereby causing a mismatch between the two eyes' projections in central targets . The only genetic mutation that causes an interocular mismatch reported to date is that associated with albinism , which has been well characterised functionally in Siamese cats [19–21] . In this case , however , the mismatch is due to an aberrant decussation at the optic chiasm followed by normal retinal mapping within the target , albeit on the inappropriate side of the brain [20] , rather than changes in response to guidance cues in the target itself . This has some parallels with mapping defects in the achiasmatic Belgian sheepdog [22] . We recently identified Ten_m3 in a microarray screen for molecules that are differentially expressed between visual versus nonvisual pathways [23] . The Ten_m ( also known as Odz and teneurin ) molecules are a recently described family of highly conserved type II transmembrane proteins of unknown function . They are the vertebrate homologs [24] of the late-acting Drosophila pair-rule gene Ten_m/Odz [25 , 26] . Recent work indicates potential roles for these molecules in mediating cellular interactions and adhesion [27] . Both Ten_m1 and Ten_m2 have been shown to be expressed in the developing avian visual system [28 , 29] . The aim of the current study was to investigate a potential role for Ten_m3 in the development of the visual pathway . We show that Ten_m3 has an important and previously unsuspected role as an eye-specific mapping molecule: the absence of Ten_m3 causes a dramatic change in the mapping of ipsilateral retinal inputs . Significantly , these occur in the absence of a major change in the mapping of geniculocortical or contralateral retinogeniculate projections . Consequently , there is a mismatch in binocular mapping , and this is associated with major deficits in the performance of visually mediated behavioural tasks . It is likely that these deficits are a result of functional suppression arising from the mismatch , because acutely silencing the inputs from one eye in adult animals restores vision .
We first analysed the normal expression of Ten_m3 in the developing visual pathway using in situ hybridisation . Ten_m3 mRNA was expressed in the innermost layer of the developing retina , corresponding to the developing RGC layer at embryonic day ( E ) 16 ( Figure 1A ) . The differential expression across the dorsoventral axis of the retina was quantified using real-time polymerase chain reaction ( PCR ) at P0 and from sections through the retina at P2 . Ten_m3 mRNA was 3 . 1 ± 0 . 50 ( mean ± standard error ) fold higher in samples from ventral retina compared to dorsal at P0 ( p < 0 . 05 , Pairwise fixed reallocation randomisation test [30] ) . Within the RGC layer , expression of mRNA for Ten_m3 was in a linear , high-ventral to low-dorsal gradient by P2 ( Figure 1D ) . In situ hybridisation showed that Ten_m3 was also expressed in the dorsal lateral geniculate nucleus ( dLGN ) , where it was highest dorsally and lowest ventrally ( Figure 1B ) . Quantification along the long ( dorsomedial to ventrolateral [DM-VL] ) axis of the dLGN revealed that Ten_m3 expression is consistently in a linear gradient ( Figure 1E ) . Interestingly , since ventral retina projects topographically to the dorsal region of the dLGN , these data show Ten_m3 is highest in corresponding regions of the visual pathway , suggesting a potential role in the establishment of retinogeniculate projections . These regions include , although they are not limited to , the sites of origin [31] and termination [32] of ipsilaterally projecting RGCs . Ten_m3 expression was also observed , though at low levels , in the ventral LGN ( vLGN; n . b . : the vLGN is a distinct nucleus of the ventral thalamus as opposed to the ventral region of the dLGN ) . Immunostaining for Ten_m3 showed a similar pattern within the dLGN as seen with in situ hybridisation , with particularly strong reaction product in the dorsal region of the nucleus ( Figure 1C; arrow , and 1F ) . Labelling was also seen superficially in the optic tract , which is the region from which retinal inputs enter the dLGN . Fine fascicles resembling fibres were also observed running through the dLGN and more ventrolaterally in tracts running to/from the cortex . It therefore seems likely that the antibody staining seen in the dLGN reflects the presence of Ten_m3 , not only in cells of the dLGN , as shown by in situ hybridisation ( Figure 1B ) , but also within fibre tracts that enter and leave the nucleus from the retina and cortex . This is consistent with the presence of Ten_m3 immunoreactivity in the white matter below visual cortex [23] . Quantification revealed that the high dorsal–low ventral expression pattern is also seen at the protein level ( Figure 1F ) , although the gradient is not as smoothly linear as for mRNA . This is most likely due to presence of protein on afferent and efferent axons as well as on dLGN cells . The antibody staining also showed reaction product in the vLGN . As for the dLGN , this is likely to represent Ten_m3 expression in the vLGN neurons , as well as on retinal axons . The above data demonstrate that the transmembrane protein Ten_m3 is expressed in both afferent axons and target structures of the developing visual pathway in a gradient that is consistently highest in regions that correspond topographically . This , together with evidence from a previous study that showed that Ten_m3 promotes homophilic interactions between cells and their processes [23] , suggested that this molecule may play an important role in regulating appropriate connectivity of the visual pathway . To investigate its functional role , a Ten_m3 knockout ( KO ) mouse was generated by disruption of exon 4 , which encodes the transmembrane region of the protein ( Figure 2A ) . Ten_m3 is an approximately 3 , 000–amino acid type II transmembrane protein; the transmembrane region is located at around 300 amino acids from the amino terminal [24] . Quantitative PCR showed that although Ten_m3 mRNA is present in the KO , it is significantly down-regulated ( 0 . 24 ± 0 . 05 fold; p < 0 . 001 , Pairwise fixed reallocation randomisation test ) compared to wild type ( WT ) , suggesting that it may be targeted for degradation rather than synthesized into protein . Western blots using an antibody directed against the extracellular domain of Ten_m3 on brain lysates from WT mice showed a single band of approximately 300 kDa corresponding to the Ten_m3 protein , whereas homozygous KO mice lacked this ( Figure 2B ) , confirming the effectiveness of the KO strategy . The presence of a severely truncated form of the protein ( up to a maximum of around 10% of the normal length ) , corresponding to part of the intracellular domain , cannot be excluded at this stage . If present , the truncated protein would be at low levels , however , with respect to WT . Ten_m3 homozygous KOs are viable and survive into adulthood . Their numbers from heterozygote breedings are typically less than 25% , suggesting some embryonic lethality , although the reasons for this are currently unknown . No phenotype was observed in heterozygotes . Ten_m3 KOs often appeared a little smaller than their WT littermates during early postnatal development , although no specific developmental delays were evident . By adulthood , the KO and WT mice are of a similar size and weight ( WT: 21 . 56 ± 1 . 01g , n = 5; KO: 20 . 58 ± 1 . 26 g , n = 5; p > 0 . 5 , t-test ) . Many adult Ten_m3 KO mice have a slightly curly tail and/or a humped back . The appearance of the brain is normal , and histological observations did not reveal any major changes in brain structure or organisation . Most importantly , no differences in size , cellular density , or lamination were apparent in Nissl-stained sections through retina , dLGN , or visual cortex of Ten_m3 KOs compared to the WT mice ( Figure 2C–2H ) . To investigate a role for Ten_m3 in regulating the formation of visual projections , the organisation of ipsilateral and contralateral retinal projections was examined in Ten_m3 KOs during the fourth postnatal week when the projection is adult-like [33] . RGC axons from each eye were labelled with cholera toxin subunit B ( CTB ) conjugated to either a red or green fluorescent dye . In WTs , the ipsilateral projection consistently formed a distinct patch within the dorsomedial quadrant of the dLGN at all rostrocaudal levels of the nucleus ( Figure 3A–3C ) . A dramatic change in the targeting of the ipsilateral projection was observed in Ten_m3 KOs . The difference was least apparent in sections though the caudal region of the nucleus , where the ipsilateral patch was confined to the dorsomedial quadrant , although unlike in WTs , the patch was expanded such that it abutted the dorsomedial border of the dLGN ( Figure 3D ) . More rostrally , the difference was much more marked; the ipsilateral patch was narrower and elongated along the DM-VL axis of the dLGN ( Figure 3E and 3F ) . The shape of the ipsilateral patch often appeared comet-like , with an expanded head dorsomedially and a thinner tail ventrolaterally . The change seen was symmetrical , highly consistent between animals , and maintained into adulthood ( see below ) . In some cases , two distinct patches of label were visible in a given section , although examination of the rostrocaudal series of sections revealed that the two patches are always continuous with each other . The distribution of ipsilateral label was quantified along the DM-VL axis of the dLGN for the entire rostrocaudal extent of the nucleus in five each KOs and WTs ( Figure 3G and 3H ) and illustrates the highly consistent nature of the mapping change . Statistical analysis confirmed that the change is highly significant ( p < 0 . 0001 , Kolmogorov-Smirnov test; n = 5 each for WT and KO ) . In WT mice , the ipsilateral and contralateral projections are anatomically segregated; contralateral projections fill all regions of the dLGN not occupied by ipsilateral terminals , including the ventrolateral region of the nucleus ( Figure 3I–3I" ) . Given the change in the mapping of ipsilateral axons in KOs , it was of interest to determine whether ipsilateral and contralateral axons are still segregated . Comparison of the distribution of the ipsilateral and contralateral projections in any given KO ( Figure 3J–3J" ) indicated that the terminals from the two eyes are clearly segregated from each other , despite the fact that ipsilateral axons now map to regions of the dLGN that would normally be innervated by axons from the contralateral eye . Since activity is believed to be important in segregating terminals from the two eyes in the dLGN [13 , 34] , this suggests that at least some aspects of retinogeniculate activity are normal in Ten_m3 KOs . Analysis of the relative area occupied by ipsilateral terminals across sections showed no change in KO mice ( WT: 14 . 5 ± 1 . 03% , n = 5; KO: 14 . 7% ± 1 . 02 , n = 5; p > 0 . 5 , t-test ) . It thus seems that the ipsilateral zone of the dLGN is elongated rather than expanded per se . In mice , over 95% of RGCs project contralaterally . The adult ipsilateral projection arises from RGCs within the peripheral ventrotemporal retina known as the ventrotemporal crescent ( VTC ) . Within the VTC , approximately 15% of RGCs project ipsilaterally and the remaining RGCs project contralaterally [31] . In normal mice , the ipsilaterally projecting cells map to the dorsal ( binocular ) region of the dLGN . More ventral regions of the dLGN exclusively receive inputs from the contralateral retina . The observed alteration in the mapping of ipsilateral projections in Ten_m3 KOs could , therefore , potentially be accounted for by two distinct mechanisms: ( 1 ) axons from dorsal or nasal retina make an inappropriate choice at the optic chiasm and project ipsilaterally , but then make an appropriate topographic choice within the target itself ( crossing defect , but in reverse direction to an albino ) ; or ( 2 ) axons make an appropriate choice at the chiasm , but make an inappropriate topographic choice within the target nucleus ( mapping defect ) . To discriminate between these possibilities , retrograde tracing studies were performed to determine the point of origin and number of ipsilaterally projecting RGCs in Ten_m3 KOs versus WTs . Injections of wheat-germ agglutinin conjugated to horseradish peroxidase ( WGA-HRP ) were made into the dLGN of four adult WT and four KO mice . Retrogradely labelled cells filled the contralateral retina of both groups of mice , confirming the accuracy of the injections . No difference in the distribution of retrogradely labelled cells across the contralateral or ipsilateral retinas , including the VTC , was apparent between the groups . Most importantly , in the ipsilateral retinas , the vast majority of retrogradely labelled RGCs were located within the VTC in both WTs and KOs ( Figure 4A and 4B ) . Quantification confirmed that there was no difference either in the size of the region containing labelled cells ( WT: 1 . 63 ± 0 . 17 mm2 , n = 3; KO: 1 . 62 ± 0 . 17 mm2 , n = 3; p > 0 . 9 , t-test; n = 3 ) or their density within the VTC ( WT: 548 ± 127 cells/mm2; KO: 612 ± 92 cells/mm2; p > 0 . 7 , t-test ) . These numbers provide an estimate of just under 1 , 000 ipsilaterally projecting RGCs in both WTs and KOs , similar to previously published values in pigmented mice [31] . The lack of a change in the number of ipsilaterally projecting cells is also consistent with the observation that the proportion of the dLGN occupied by ipsilateral terminals is not altered in KOs . A small number of retrogradely labelled cells were seen scattered over other regions of the ipsilateral retina in both KOs and WTs , but no difference in their occurrence was apparent . Retrograde labelling from the superior colliculus in developing animals ( postnatal day [P]3–5 ) , labelled only a subset of RGCs , but produced qualitatively similar results ( unpublished results ) . These data present strong evidence that the change in mapping is due to a mapping defect within the dLGN and not to inappropriate crossing of cells from more dorsal or nasal retina at the optic chiasm . To investigate the topography of retinal projections , we made focal injections of the carbocyanine dye , DiI , into the retina of P11–12 animals ( this age was chosen because it is just after the period when topography becomes refined [34] ) . Following a focal injection of DiI into the VTC of WTs , a single , well-localised terminal zone ( TZ ) was consistently observed contralaterally as a densely labelled region abutting the dorsomedial border of the dLGN ( Figure 5A ) . The ipsilateral TZ was typically larger and positioned slightly more ventrally ( Figure 5B ) , consistent with the normal characteristics of the ipsilateral projection in mice [32 , 35] . In Ten_m3 KOs , a single , well-localised TZ was seen contralaterally , adjacent to the dorsomedial border of the dLGN ( Figure 5C ) , as in WTs . In contrast , ipsilateral axons in Ten_m3 KOs consistently showed abnormalities in their targeting within the dLGN ( Figure 5D ) . Most strikingly , focal injections into the VTC consistently resulted in two distinct TZs within the ipsilateral dLGN: one dorsomedially and the other more ventrolaterally . This was observed in all cases examined ( Figure 5E–5H ) . In cases in which retinal axons were well labelled ( as opposed to predominantly terminal labelling ) , axons were observed to run in a fairly straight line between the two TZs ( inset in Figure 5D ) , suggesting that individual axons may innervate both TZs ( if different axons innervated each TZ , we would expect that retinal axons would pass directly from the optic tract to the dorsomedial TZ , rather than arriving there via the more ventrolateral TZ ) . The possibility that individual RGC axons have multiple terminal foci is also supported by the observation that some axons appeared to branch near the more ventrolateral TZ , sending one branch into this TZ and another dorsomedially towards the other TZ ( inset in Figure 5D ) , although analysis at higher resolution will be needed to confirm this . Occasionally , only one TZ was visible in a given section from a KO , but comparison with an adjacent section revealed two distinct patches of label ( see plot for animal 3 in Figure 5H ) . Although label also spanned multiple sections in WTs , comparison of adjacent sections suggested that they were always part of a single TZ ( Figure 5G ) . To gain an objective measure of the number of TZs per section , a cluster analysis was performed . This analysis determined that there was a single cluster present in 100% of contralateral sections and in ipsilateral sections from WTs ( median = 1 , mode = 1 , mean = 1 . 0 ± 0 ) . The same analysis determined that there were two clusters in almost every labelled ipsilateral section from KOs ( median = 2 , mode = 2 , mean = 1 . 9 ± 0 . 14 ) . This was significantly different from all other groups ( p < 0 . 01 , Wilcoxon rank sum test , comparing ipsilateral sections in KOs versus ipsilateral sections in WT , and contralateral sections in KO and WT ) . The position of the contralateral patch along the DV-ML axis was similar between WTs and KOs , supporting the suggestion that there is no major shift in topography of contralateral projections . As an additional control , focal injections into the dorsal retina were also performed . These experiments revealed no change in the topography of contralateral retinal projections from this region ( Figure S1 ) . No ipsilateral terminals were observed following injections into dorsal retina , consistent with the retrograde tracing studies demonstrating that the absence of Ten_m3 results in a mapping , rather than a decussation , defect . The dLGN receives inputs from the retina and sends outputs to the primary visual cortex ( area 17 ) . We therefore examined the topographic relationship between visual thalamus and cortex in Ten_m3 KOs . This was of particular interest both because Ten_m3 is expressed in geniculocortical neurons and in developing visual cortex [23] , and because compensatory changes in geniculocortical projections have been reported in Siamese cats [19] . For this , injections of biotinylated dextran amine were made into medial , central , and lateral regions of area 17 of adult WTs and Ten_m3 KOs . Labelling patterns correlated well with injection position in all animals ( Figure S2 ) . For example , injections into medial area 17 resulted in a patch of anterograde and retrograde label in ventral dLGN of both WTs and KOs ( Figure S2A and S2B ) . Injections into more lateral regions of area 17 resulted in the normal topographic shift of the transported label to more dorsal regions of the dLGN for all animals ( Figure S2C and S2D ) . The size of the labelled region following similar injections typically appeared slightly larger in KOs , although this difference was not significant ( WT: 5 . 8 ± 1 . 3 × 105 pixels , n = 9; KO: 6 . 7 ± 1 . 2 × 105 pixels , n = 8; p > 0 . 5 , t-test ) . That is , although there may be a subtle decrease in the precision of geniculocortical connectivity , the overall topography of the pathway appeared normal in Ten_m3 KOs . In order to confirm these results with respect to the potential transfer of information from the aberrantly mapped ipsilateral eye to the visual cortex , retrograde tracing from the cortex was combined with anterograde tracing from the retina . In WTs , injections of green CTB into the lateral ( binocular ) zone resulted in a patch of retrogradely labelled cells in the dorsomedial region of the dLGN . Notably , all of the retrogradely labelled cells were contained either within , or immediately adjacent to , the ipsilateral patch ( this was visible either as ipsilateral retinal terminals anterogradely labelled with red CTB as in Figure 6B , or with the gap in the location of labelled terminals in the contralateral dLGN as in Figure 6A ) . The absolute position of the cells retrogradely labelled from lateral area 17 was similar in KOs ( Figure 6C and 6D ) ; however , because the ipsilateral patch is much narrower and extends into the ventrolateral region of the nucleus , the locations of the anterograde and retrograde label did not necessarily correlate with each other . Injections into the medial ( normally monocular ) region of area 17 produced retrograde label in the ventral region of the dLGN in both WTs ( Figure S2A ) and KOs ( Figure S2B ) . The combined anterograde and retrograde labelling confirmed that this region is in close proximity to , and overlaps with , the ipsilateral patch in KOs ( Figure 6E ) . The presence of retrogradely labelled cells within the ipsilateral patch is clearly visible at higher power ( Figure 6F ) . Although only seen here in low numbers , this confirms that some ipsilateral retinal inputs will be represented in medial area 17 in KOs . The small numbers most probably reflect the fact that only around 15% of geniculocortical axons represent ipsilateral inputs ( as opposed to the 85% that represent contralateral inputs ) and the associated difficulty of accurately targeting them without prior physiological mapping of injection sites . A correlation between the position of cells retrogradely labelled from medial area 17 and the ipsilateral patch was not seen in WTs ( unpublished data ) . This indicates that the mismatch of ipsilateral and contralateral retinogeniculate inputs is transferred to the cortex . A schematic diagram summarizing the change in retinogeniculocortical mapping in Ten_m3 KOs is provided in Figure 7 We wished to determine whether the observed retinogeniculate mismatch affects vision in the Ten_m3 KO mice . Three visually mediated behavioural tasks—vertical placement , horizontal placement , and a modified version of the visual cliff test [36]—were performed under different conditions . As a control , the vertical placement test was performed under red light ( a wavelength not detected by the mouse retina ) , and mice were graded by observers who were blind to genotype as to when they reached for the target , a metal bar , using somatosensory cues ( Figure 8A and 8B ) ( see Materials and Methods for explanation of scores ) . Performance of the two groups under these conditions was identical , suggesting that Ten_m3 KOs are both capable of , and motivated to , perform the test when it is mediated by the somatosensory system ( Figure 8C; WT: 1 . 08 ± 0 . 08 , n = 6; KO 1 . 0 ± 0 . 00 , n = 4; p > 0 . 5 , t-test ) . To test the ability of Ten_m3 KOs to perform this task using the visual system , the task was repeated under normal ( ambient ) light with the whiskers trimmed so they could not rely on somatosensory cues . WTs scored significantly better than KOs under these conditions ( Figure 8C; WT: 1 . 1 ± 0 . 1 , n = 10; KO: 0 . 22 ± 0 . 22 , n = 9; p < 0 . 01 , t-test ) . In many cases , KOs did not reach out until their nose touched the bar ( Figure 8B ) . The horizontal placement test was also performed , and although scores were lower for all mice , results for WTs were significantly higher than for KOs who showed no visual response to the bar at all ( WT: 0 . 6 ± 0 . 18 , n = 18; KO: 0 ± 0; n = 6; p < 0 . 01 , t-test ) . The visual cliff test also revealed a clear difference in the behaviour of Ten_m3 KOs ( Figure 8D ) . For this test , animals were placed in the centre of a box with a clear acrylic ( Perspex ) base with a high-contrast grating appended to one half of its lower surface . The box was positioned such that the clear side protruded from the laboratory bench to give the impression of a “cliff” at the edge of the grating . The number of times animals approached the visual cliff , the percentage of these approaches that resulted in the animals crossing the cliff to the clear side of the box , the total amount of time spent in each half of the box , and mean activity levels were analysed . In the majority of cases , WTs approached the border of the grating , inspected the cliff , and then retreated to the patterned side ( Video S1 ) . In contrast , KO mice frequently walked straight over the border region and onto the cliff side without pausing ( Video S2 ) . On average , WT mice spent in excess of 90% of their time on the patterned surface ( 91 . 4 ± 1 . 66%; n = 31 ) , whereas Ten_m3 KO mice spent approximately half their time ( 57 . 4 ± 5 . 0%; n = 18 ) in this region ( Figure 8D ) . This difference is highly significant ( p < 0 . 001 , t-test ) and suggests that , unlike WT mice , KOs do not exhibit a preference for the patterned half of the box . A more detailed analysis of the behaviour of the mice showed that although KOs approached the cliff less often than WTs ( Figure 8E; WT: 18 ± 1 . 5 , n = 33; KO: 10 . 4 ± 1 . 9 , n = 14; p < 0 . 01 , t-test ) , they crossed it significantly more frequently ( Figure 8F; WT: 25 . 4 ± 4 . 1% , n = 33; KO: 61 . 5 ± 4 . 6% of approaches , n = 14; p < 0 . 0001 , t-test ) . In addition , the average time spent on the cliff side of the box per crossing was almost 10-fold higher in KOs ( Figure 8G; WT: 10 . 5 ± 2 . 1 s , n = 33; KO: 93 . 4 ± 35 . 0 s , n = 14; p < 0 . 05 , t-test ) . As a control for possible differences in absolute activity levels , the average linear displacement of a randomly chosen subset of KOs and WTs was also determined; this did not differ between the groups of mice ( WT: 251 ± 43 cm/min , n = 5; KO: 265 ± 21 cm/min , n = 5; p > 0 . 7 , t-test ) . To determine whether the observed differences in the behaviour of WTs and KOs in the visual cliff test are mediated by visual cues , the same test was conducted under dark conditions ( red light ) . The two groups performed almost identically to each other for all parameters examined ( Figure 8D–8F; activity levels: WT: 765 ± 232 cm/min , n = 7; KO: 925 ± 249 cm/min , n = 5 , p > 0 . 5 , t-test; time on patterned side: WT: 55 . 9 ± 2 . 9% , n = 7; KO: 54 . 0 ± 3 . 8% , n = 5; p > 0 . 5; approaches: WT: 17 . 0 ± 2 . 8 , n = 7; KO: 17 . 0 ±1 . 4 , n = 5; p = 1 , t-test; percent crossings: WT: 88 . 7 ± 3 . 4% , n = 7; KO: 90 . 8 ± 0 . 9% , n = 5; p > 0 . 5 , t-test ) . When these values were compared to those for each group under normal light conditions , it was found that mean activity levels were markedly higher for both groups , although this did not reach statistical significance for either WTs or KOs ( p > 0 . 05 , t-test ) . The number of approaches in WTs showed no difference when compared to WTs under light conditions ( p > 0 . 9 , t-test ) , but KOs approached the border significantly more often in the dark compared to light conditions ( Figure 8E; p < 0 . 05 , t-test ) . Both groups also crossed the cliff significantly more often than under normal light ( Figure 8F; p < 0 . 0001 for both groups , t-test ) . Most significantly , both groups of mice spent approximately half their time on the patterned side ( Figure 8D ) . These values are significantly different from those obtained for WTs under ambient light ( p < 0 . 001 , t-test ) , but essentially identical to those for KOs under ambient light . Together , these results suggest that the marked difference in the behaviour of the WTs and KOs in the visual cliff test is indeed mediated by visual cues . They also indicate that the KOs have the capacity to distinguish light from dark , even though they show little behavioural response to the visual cliff . We postulated that , given the relatively normal histological appearance of the visual pathway , the defect in the performance of the visual cliff test may be a direct result of the interocular mismatch in the retinogeniculocortical pathway ( see Figure 7 ) rather than due to more generalised effects of Ten_m3 on neural connectivity . To test this hypothesis , the visual cliff test was performed on mice that had inputs from one eye acutely silenced via an intraocular injection of the sodium channel blocker tetrodotoxin ( TTX ) . The effectiveness of the blockade was confirmed by checking that the pupillary light reflex was absent in the injected eye at the time of the test ( this is normally present in Ten_m3 KOs; C . A . Leamey and A . Sawatari; unpublished data ) . Activity levels were not different from those obtained under normal light conditions ( WT: 330 . 6 ± 51 . 3 , n = 6; p > 0 . 2 , t-test; KO: 273 . 3 ± 40 . 6 , n = 7; p > 0 . 7 , t-test ) , suggesting that the drug did not compromise the mobility of the mice . The number of approaches to the cliff made by WT mice with monocular TTX ( Figure 8E; 15 . 3 ±3 . 2 , n = 6 ) showed no difference to WT mice without TTX ( p > 0 . 4 , t-test ) . The proportion of times they crossed the border ( Figure 8F; 47 . 9 ± 12 . 1 , n = 6; p > 0 . 1 ) and total time spent over the patterned surface ( Figure 8D: 68 . 9 ± 6 . 4% , n = 7; p > 0 . 05 , t-test ) were , however , both noticeably different . Although the differences were not statistically significant , the change in behaviour associated with monocular TTX was of concern . To see whether the decrease in the performance levels of the WTs with TTX was a direct consequence of the loss of part of the visual field , the data were analysed to see whether the behaviour of the mice correlated with the eye that was facing the cliff as they approached the border . For example , the mice often approached the cliff either at an acute angle , or close to one of the side walls . If the active eye was facing away from the cliff , or towards the wall , the cliff would be largely or completely out of the field of view of the active eye . The number of approaches , and associated crosses , were therefore subdivided into those made while the cliff was predominantly within the field of view of the active versus the inactive eye . When the analysis criteria were restricted such that only approaches where a clear bias for one eye versus the other was apparent , we found that only a minority of these ( 36 . 3 ± 13 . 2% , n = 6 ) , were associated with the mouse crossing the border when the cliff was predominantly within the field of view of the active eye . Expanding the criteria to include cases where the cliff was within the field of view of both of the active and inactive eye produced similar results ( 35 . 3 ± 11 . 9% , n = 6 ) . These values are not different from the percent crosses made under normal light conditions ( p > 0 . 45 , t-test ) . In contrast , a very high proportion of approaches made where the active eye was facing away from the cliff were associated with a crossing ( 81 . 9 ± 8 . 2% , n = 6 ) . This value is significantly higher than the percent crosses under normal light ( p < 0 . 001 , t-test ) , and the percent crosses made when approaches were made using the active eye ( p < 0 . 05 , t-test ) or both eyes ( p < 0 . 05 , t-test ) , and is in fact similar to the percent crosses made under dark conditions . These results show that the decrease in the performance of the WTs with monocular TTX is directly associated with the loss of part of the visual field , rather than with other effects of the drug . The time spent on the cliff side of the box per cross did not show a significant difference compared to untreated WTs ( Figure 8G; 17 . 2 ± 5 . 2; p > 0 . 2 , t-test ) . The behaviour of the Ten_m3 KO mice following monocular activity blockade was dramatically altered in comparison to their behaviour without TTX . Under these conditions , the KOs approached the border a similar number of times ( 7 . 3 ± 1 . 2 , n = 7 ) to KOs without TTX ( p > 0 . 1 , t-test ) and appeared to actively investigate the cliff before retreating ( Video S3 ) , a behaviour that resembled that of WTs under normal light conditions . Following TTX administration , the KOs only crossed the border in a minority of cases ( 14 . 2 ± 6 . 8% of approaches , n = 7 ) . This value is significantly less than for KOs without TTX ( p < 0 . 0001 , t-test ) . The KOs with TTX also exhibited a more than 10-fold reduction in the time spent over the cliff per crossing compared to KOs without TTX ( 8 . 4 ± 6 . 6 s , n = 7; p < 0 . 05 ) , a value similar to WTs under normal light . On average , KOs spent 90 . 6 ± 7 . 9% ( n = 6 ) of time on the patterned side during monocular blockade , almost identical to the values obtained for WTs under normal conditions . The difference between this value and that obtained from the KO mice under control conditions is significant ( p < 0 . 01 , t-test ) . As with the WTs , the KO mice crossed the cliff much less frequently when it was in the field of view of the active versus the inactive eye ( only 4 . 8 ± 4 . 8% , n = 7 , of approaches made with the active eye facing the cliff were associated with a crossing compared to 40 . 3 ± 18 . 6% , n = 7 , where the inactive eye was facing the cliff ) . To confirm that this remarkable apparent recovery of visually mediated behaviour was due to the monocular blockade rather than individual variation , two KO mice were tested both before and during the blockade . Both of these mice spent far less time on the cliff side ( reductions of 41% and 22% of total time , respectively ) during the monocular blockade than before the blockade .
Ten_m3 is a transmembrane protein [24] that promotes homophilic interactions between cells that express it [23] . We have shown here that it is expressed in both afferent and target structures of the developing visual system in gradients that are consistently highest in regions that correspond topographically to ventral retina ( dorsal visual field ) . We have also shown that there is an expansion of ipsilateral retinal axon terminals into ventrolateral regions of the dLGN in Ten_m3 KOs . This occurs in the absence of an observable change in the origin of the ipsilateral projection or the morphology and cytoarchitecture of visual structures . Together , these data suggest a role for Ten_m3 in axon guidance , possibly acting in the manner of a chemoaffinity molecule as hypothesised by Sperry [37] to attract axons from ventral retina to dorsomedial dLGN . Interestingly , rather than this being achieved by the expression of distinct ligands and their receptors , in the case of Ten_m3 , our data suggest that this may be achieved by homophilic interactions between molecules expressed in corresponding gradients across the afferent and target fields . Although homophilic interactions between cell adhesion molecules have previously been postulated to play a role in the molecular matching of afferent and target regions [38–42] , our data raise the possibility that they may also play a direct role in controlling the topographic mapping within these regions . It is possible that other , yet to be discovered , receptors/ligands for Ten_m3 exist and that these contribute to the role of Ten_m3 in axon guidance . Although there are some similarities between our data supporting a role for Ten_m3 in topographic mapping and the central tenet of Sperry's chemoaffinity hypothesis [37] , the observation that ipsilateral terminals were shifted , not only ventrolaterally , but also dorsomedially , suggests , however , that the role of Ten_m3 in axon guidance may be more complex than can be accounted for by monotonic chemoaffinity gradients . Recent work has provided strong evidence for attractive and repulsive forces acting to counterbalance each other in the mapping of the retina onto central targets [43–46] . It is likely that the expression of and/or response of axons to multiple attraction/repulsion cues are affected by Ten_m3 . Because repulsive EphA–ephrinA interactions are reported to confine ipsilateral axons to dorsomedial dLGN [13] , the ventrolateral expansion of the ipsilateral terminal region is suggestive of changes in EphA–ephrinA expression . However , as the targeting of contralateral axons is not appreciably altered , and this is known to be affected by EphA–ephrinA interactions [14 , 18] , it is possible that other or additional mechanisms are involved in targeting of eye-specific projections . Mice lacking the β2 subunit of the nicotinic acetylcholine receptor also have an expanded ipsilateral projection which , unlike WTs , is not segregated anatomically from the contralateral projection into a distinct , ipsilateral layer [47 , 48] . This may be attributable to the absence of correlated , spontaneous retinal waves during early postnatal development [49] , although the absence of the receptor in the target structures may also play a role . The fact that ipsilateral and contralateral retinal axons are segregated into distinct regions in Ten_m3 KOs suggests that at least some aspects of activity are normal in Ten_m3 KOs . It also demonstrates that segregation can occur independent of the positioning of the ipsilateral terminals . This is consistent with work in ephrinA2/A3/A5 mutants that also showed that segregation can occur independently of normal topography [13] . One of the most intriguing observations made here is the change in the mapping of ipsilateral projections in the absence of a noticeable change in the topography of contralateral projections . This provides strong evidence that different guidance cues and/or receptors mediate the mapping of ipsilateral and contralateral projections . Such differential responses are clearly necessary to produce maps that are in register for the two eyes . Ten_m3 is the first molecule to be reported that clearly acts as an eye-specific guidance molecule within target nuclei . The interpretation of previous studies that have suggested that ephrinAs have a role in eye-specific mapping is complicated by the fact that they constitute topographic cues for both ipsilateral and contralateral RGCs . Thus , although removing ephrinA gradients leads to an alteration of the ipsilateral eye projection [13] , similar manipulations also cause an expansion of the topographic representation of the contralateral projection [14–18] . Indeed , both ipsilateral and contralateral projections are shifted ventrolaterally in ephrinA2/A5/A3 KOs [13 , 18] . This is consistent with optical imaging [1] experiments in the visual cortex of ephrinA2/A5/A3 KO mice that indicate that the topographic representation of both ipsilateral and contralateral projections is expanded in a matched way . A likely means by which Ten_m3 could affect the guidance of ipsilateral , but not contralateral , RGC axons would be if it was expressed only in the ipsilaterally projecting population . The expression pattern of Ten_m3 , which extends beyond the VTC , is , however , not consistent with this possibility . Differences in expression levels between contralaterally and ipsilaterally projecting RGCs and/or interactions with other molecules that are differentially expressed in ipsilaterally versus contralaterally projecting RGCs [6–10] offer potential explanations for the specificity of Ten_m3′s role in RGC guidance; these possibilities require further investigation . Although not previously demonstrated for central targets , the differential expression of , and response to , guidance cues by ipsilateral versus contralateral axons is well established at an important intermediate guidance point for retinal axons , the optic chiasm . Recent work has shown that the EphB1 receptor plays a key role in guiding ipsilateral projections in this region [7] and that expression of the transcription factor zic2 in the developing retina defines the adult ipsilateral projection [9] . Interestingly , Ten_ms have been shown to undergo regulated intercellular proteolysis in a manner similar to Notch [50 , 51] . Following binding of the extracellular domain , the intracellular domains are cleaved and then translocated to the nucleus where they interact with transcription factors . Ten_m2 has been shown to interact with zic1 [50] . Although binding partners for Ten_m3 have not yet been identified , interactions with zic genes or other molecules expressed in ipsilateral axons , such as EphB1 , are possible . It is therefore likely that Ten_m3 may act both directly , as an adhesion/attractant molecule , and indirectly , via transcription of other molecules , to guide ipsilateral axons within the target . Other recent work has shown that the adhesion molecule NrCAM is required for the late-born cells of the VTC to project contralaterally [8] . Given there is no evidence that contralateral projections are affected in Ten_m3 KOs , interactions with this signalling pathway , which is distinct from the EphB1 pathway [8] , seem less likely . Another major finding of the study is the profound deficit in visually mediated behaviour exhibited by Ten_m3 KOs . The findings suggest active functional suppression of inputs from one eye by the other , rather than merely the loss of stereoscopic depth perception associated with the binocular field , for the following reasons . ( 1 ) The visual cliff test employs the ventral visual field ( dorsal retina ) , which is not part of the binocular visual field in mice [52] . ( 2 ) The performance of Ten_m3 KOs improved to a level almost identical to that of WTs in the control condition following monocular inactivation , and WTs with TTX performed the task effectively when the cliff was within the field of view of their active eye . It is surprising that a change in the targeting of the ipsilateral projection , which accounts for only 2%–3% of the total RGC population in mice [31] , has such a marked effect . The ipsilateral input is greatly amplified , however , both in terms of its area of representation in dLGN [35] and its ability to drive cortical neurons [31] . The increased dLGN representation of ipsilaterally projecting cells is maintained in the Ten_m3 KOs , thus the observed changes could lead to much larger effects than suggested by consideration of RGC number alone . Despite the profound deficits that KOs exhibit when performing behavioural tasks that require patterned vision , they appear to retain an ability to discriminate light from dark; it is possible that this is mediated via subcortical visual pathways . Based on our anatomical data , we suggest that the mistargeting of ipsilateral axons in the dLGN in Ten_m3 KOs results in a more widespread representation of ipsilateral inputs to the visual cortex rather than being confined to the lateral third of area 17 as in WTs ( Figure 7 ) . Cells that would normally receive inputs from nearby regions of the contralateral monocular visual field will , therefore , receive inputs from widely disparate regions of monocular and binocular visual fields . Thus , there will be a mismatch of inputs to the visual cortex , potentially leading to interocular suppression , as reported in visual cortex following strabismus [53 , 54] , in the superior colliculus following monocular deprivation [55] , and in Siamese cats in which , as here , an interocular mismatch in the dLGN is transferred to visual cortex [20 , 21] . The dramatic and rapid apparent recovery of visual function in Ten_m3 KOs following acute monocular blockade is strongly supportive of this hypothesis: the blockade allows one cortical hemisphere to receive appropriately mapped contralateral inputs without the confounding influence of aberrant ipsilateral inputs , relieving the mismatch and thus restoring visual behaviour . Of equal interest , cortical networks upstream of the dLGN seem able to correctly interpret visual input from one eye alone , after blockade of the mismatch , and enable vision in adulthood , even though these upstream structures have never received matched inputs from the two eyes during development .
Embryos were obtained from timed matings of C57/Black6 mice . Mothers were anaesthetised with 4% isofluorane and the embryos delivered by Caesarean section and then decapitated . Neonatal mice were anaesthetised with an overdose of sodium pentobarbital and the brains removed . For PCR analysis , the retinas were subdivided and relevant regions were dissected . RNA was extracted , and quantitative real-time PCR for Ten_m3 and a control gene was performed as described in [23] . For in situ hybridisation , tissue was frozen in isopentane on dry ice . In situ hybridisation was performed using standard procedures on 15 μm–thick fresh-frozen cryostat sections using 200 bp–long dioxigenin-labelled riboprobes to sense and antisense Ten_m3 sequences . The reaction was developed using a fluoroscein-labeled TSA-plus kit ( PerkinElmer , http://www . perkinelmer . com ) . Antibody staining was performed using a rabbit anti-Ten_m3 [56] at 1:25 , followed by a biotinylated goat anti-rabbit secondary antibody , and developed with ABC ( Vector Laboratories , http://www . vectorlabs . com ) and a TSA plus ( fluoroscein ) kit . Ten_m3−/− mice were generated with a targeting construct in which the transmembrane-containing exon 4 of the Ten_m3 gene was disrupted with a neomycin gene ( Figure 2 ) . The targeting vector was electroporated into R1 embryonic stem ( ES ) cells ( passage 13 ) , and four independently targeted ES cell clones were injected into C57Bl/6 blastocysts to generate germline chimeras . The chimeric founders were crossed to C57Bl/6 females to establish heterozygous Ten-m3+/− and subsequently homozygous Ten_m3−/− ( KO ) mice . Third generation backcrosses into C57/Bl6 were performed , but KO mice did not survive well on this background . Consequently , these mice were crossed with Sv129 mice , and the strain was maintained on this background . Animals are of varied pigment . Because alterations in the ipsilateral pathway of pigmented and albino mice have been reported [31] , all quantitatively assessed anatomical tracing experiments reported here on the retinogeniculate pathway were performed on pigmented mice . Qualitatively similar changes in mapping were , however , observed in animals from both pigment groups . Behavioural experiments were performed on both albino and pigmented animals . Initially , data were analysed separately for pigmented and albino animals , but because no differences in visual behaviour relating to pigment were detected , KO and WT data were pooled across pigment groups . Genotyping was performed by Southern blot or PCR using DNA isolated from tail biopsies . For this , expression lysates from the brain of WT and KO mice were reduced , resolved in 6% SDS-PAGE gels , and then transferred to PVDF membranes and incubated with an affinity-purified rabbit anti-mouse Ten_m3 antibody [56] . All surgical manipulations were performed under anaesthesia induced and maintained by inhalation of 2%–4% isofluorane in oxygen . Intraocular injections of 0 . 5–1 μl of 1% CTB conjugated to Alexa Fluor 594 ( red ) or Alexa Fluor 488 ( green ) were made into the right and left eyes of Ten_m3 KO and WT littermates at P21–23 . For combined anterograde and retrograde labelling , injections of green CTB were made into subregions of visual cortex and red CTB was injected into one eye as above . Animals were euthanised 1–3 d later and perfused with 0 . 9% saline followed by 4% paraformaldehyde in 0 . 1 M phosphate buffer . Coronal sections , 60 μm thick , were cut using a freezing microtome . Focal injections of a 10% solution of 1 , 1′-dioctadecyl-3 , 3 , 3′ , 3′-tetramethylindocarbocyanine perchlorate ( DiI ) in dimethyl formamide were made in peripheral VT retina of P11–12 mice under isofluorane anaesthesia; animals were perfused 1–2 d later . This age was chosen because topography is largely adult-like by this stage [45] . Coronal sections , 100 μm thick , were cut on a vibratome . Injections of 10% biotinylated dextran amine ( BDA ) were made into subregions of primary visual cortex ( area 17 ) of adult mice using stereotaxic coordinates . Label was allowed to transport for 7–10 d before animals were euthanised , perfused as above , and reacted as described in [57] . For retrograde labelling from the thalamus , 5% WGA-HRP was injected into the dLGN using stereotaxic coordinates . Animals were perfused 16–20 h later with 0 . 9% saline followed by 2% paraformaldehyde in 0 . 1 M phosphate buffer . Retinas were dissected out and reacted as whole mounts . Coronal sections , 80–100 μm thick , were cut through the dLGN using a vibratome , and a tetramethyl-benzidine reaction was performed using standard techniques . Images were analysed using Image J ( NIH ) . The freehand selection tool was used to outline the perimeter of the dLGN , and the area contained within this region in arbitrary units was determined . Analyses were performed across the entire rostrocaudal extent of the nucleus . Care was taken to exclude the optic tract , intergeniculate leaflet , and vLGN from this analysis . To isolate ipsilateral projections , background was subtracted using the rolling ball function , and the area containing the ipsilateral projections was selected using the wand tool . To determine shifts along the DM-VL axis , the image was cropped at the borders of the dLGN , thresholded as above , and the coordinates of each non-zero pixel were written to a file along with information on the size of the image in arbitrary units . The DM-VL axis was divided into 100 bins , and the proportion of label in each bin , expressed as a percentage of the total ipsilateral label contained within each bin , was plotted using Matlab . Means and standard errors for five KOs and five WTs were calculated and plotted . For the analysis of focal injections , all sections containing label from four KO and four WT animals were used . Thresholded data were binned as above , and Matlab was used to generate a heat map of the distribution of label . For an objective analysis of patch number , the Image analysis toolbox ( Matlab ) was used to calculate the number of clusters of label in images from KO and WT animals from the thresholded images . The area of labelled regions in BDA-injected animals was calculated by thresholding all labelled images in ImageJ . Labelled pixels were written to a text file and summed across all labelled sections for each animal . Means and standard errors are presented unless otherwise stated . For determination of the area of retina containing a high proportion of ipsilaterally projecting RGCs , this region was outlined in low-power photomicrographs of retinal whole mounts using ImageJ . For estimates of cell density , two fields of view within the region containing ipsilaterally projecting RGCs were photographed using the 20× objective for three KO and three WT cases . Images were thresholded to isolate individual RGCs . The number of thresholded patches was calculated using ImageJ software . For the vertical placement test , mice were held by the tail and lowered towards a horizontal metal bar . Scores were given by two independent observers who were blind to genotype , according to when mice made a clear reaching motion for the bar with their forepaws . A score of 2 was given for mice that reached more than 1 cm from the bar and a score of 1 for mice that reached within 1 cm , but before their nose touched the bar . A score of 0 was given for mice that touched the bar with the nose before reaching for the bar . The vertical placement test was performed under 2 conditions: normal ( ambient 55Cm−2 ) light with whiskers trimmed and under red light with whiskers at normal length . For the horizontal placement test , mice were moved horizontally toward the bar and scored as above . For the visual cliff test , mice were placed in a 60-cm × 60-cm box with a clear acrylic ( Perspex ) base . A high-contrast grating was attached to the underside of one half of the box . The box placed such that the clear half of the base protruded from the laboratory bench , revealing a drop to the floor of approximately 90 cm . Mice were placed in the centre of the box and their behaviour monitored for 10 min via a digital video camera mounted 1 m above the floor of the box . The data were analysed , blind to genotype , from video recordings . An approach was defined as moving from the patterned region towards the “cliff” such that the nose was within 5 cm of the midline . Crossing was defined as the animal completely crossing the midline from the patterned to the clear half of the box . Retreating was defined as moving such that the head was within 5 cm of the midline and retreating without the entire body crossing the border . Subsequent approaches were not scored until the mouse had moved outside the 5-cm range . To confirm the role of the visual system in this task , the test was also performed under dark ( red light ) conditions . The effect of monocular blockade of activity was determined by injecting 0 . 5 μl of 1 mM tetrodotoxin into the left eye under isofluorane anaesthesia as above . Some mice appeared lethargic for a few hours postinjection and so were allowed to recover from any systemic effects of the procedure for 2–16 h before testing to ensure that this would not bias the results . Activity levels were monitored to confirm this was the case by calculating the movement of the mice for 1 min . For this , a scaled grid was placed over the video monitor , and the points at which the mouse crossed the grid were plotted and measured . In addition to the other analyses performed , the approaches and crosses during monocular inactivation were subdivided into cases where the mouse approached the cliff such that it was predominantly in the visual field of the active versus the inactive eye . Approaches to the cliff that were made such that the mouse was parallel to and within 10 cm of a side wall , and thus the cliff was largely out of the field of view of the eye facing towards the wall , were included in this analysis . Approaches made at an angle such that the cliff was predominantly within the field of view of one eye rather than the other were also included . Approaches where there was no clear bias for one eye versus the other were scored separately . The effectiveness of the blockade at the time of the test was confirmed by checking that the pupillary light reflex was absent in the injected eye .
The National Center for Biotechnology Information ( NCBI ) ( http://www . ncbi . nlm . nih . gov ) accession number for Ten_m3 is NM_011857 . | The visual world is represented within the brain as a series of maps of visual space . In species with binocular vision , the inputs from the two eyes are aligned to form a cohesive map; little is known about how this organisation is achieved during development . We show that a transmembrane protein , Ten_m3 , plays an important role . Ten_m3 is required for the guidance of uncrossed retinal axons: uncrossed projections from the eye to the brain map aberrantly in mice that lack Ten_m3 , although crossed projections map normally . Consequently , projections from the two eyes are not aligned in these mice . We show that this mismatch has devastating consequences for vision . Mice lacking Ten_m3 perform very poorly in behavioural tests of visual function . The deficits are a direct result of the mismatch , because acutely silencing inputs from one eye restores visual behaviour . This remarkable and rapid recovery suggests the mismatch of the inputs from the two eyes leads to functional suppression in the brain . We conclude that Ten_m3 acts as an eye-specific guidance cue for retinal axons and is required to produce aligned projections from the two eyes , and further , that this is critical for normal visual function . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods",
"Supporting",
"Information"
] | [
"developmental",
"biology",
"mus",
"(mouse)",
"neuroscience"
] | 2007 | Ten_m3 Regulates Eye-Specific Patterning in the Mammalian Visual Pathway and Is Required for Binocular Vision |
Aquaporins are transmembrane proteins that facilitate the flow of water through cellular membranes . An unusual characteristic of yeast aquaporins is that they frequently contain an extended N terminus of unknown function . Here we present the X-ray structure of the yeast aquaporin Aqy1 from Pichia pastoris at 1 . 15 Å resolution . Our crystal structure reveals that the water channel is closed by the N terminus , which arranges as a tightly wound helical bundle , with Tyr31 forming H-bond interactions to a water molecule within the pore and thereby occluding the channel entrance . Nevertheless , functional assays show that Aqy1 has appreciable water transport activity that aids survival during rapid freezing of P . pastoris . These findings establish that Aqy1 is a gated water channel . Mutational studies in combination with molecular dynamics simulations imply that gating may be regulated by a combination of phosphorylation and mechanosensitivity .
Transmembrane water flux is fundamental to the physiology of all living organisms , yet biological membranes display only limited intrinsic water permeability . Cells thus control and maintain water homeostasis using aquaporins , which facilitate the selective movement of water ( orthodox aquaporins ) or other small molecules such as glycerol or urea ( aquaglyceroporins ) [1] . These solutes pass through aquaporins at almost the diffusion rate yet the pore simultaneously prevents the leakage of protons [2]–[4] . Human and plant aquaporins are frequently regulated post-translationally by phosphorylation [5] or by biophysical stimuli such as changes in pH [6] , [7] . Unicellular organisms , which have large surface-to-volume ratios , tightly regulate the flow of water and osmolytes during different stages of growth and under conditions of stress [8] . Yeast aquaporins are suggested to be regulated at the transcription level [9] , and in addition a yeast aquaglyceroporin has been reported to be post-translationally regulated by phosphorylation [10] or osmolarity [11] . In the well studied yeast Saccharomyces cerevisiae , the presence of aquaporins enhance the host's tolerance to rapid freezing [12] and aquaglyceroporins control the cellular osmolyte content following osmotic shock [13] . The physiological relevance of rapid freeze tolerance is explained by the formation of less intracellular ice crystals . During the freezing of cells in an aqueous medium , extracellular water will freeze faster than the intracellular water because of the higher osmolarity of the cellular content . Aquaporins facilitate the rapid out flow of water out of the cell and hence less intracellular ice forms within the organism , which results in less cellular damage [8] . Pichia pastoris is a methylotrophic yeast that was first isolated from chestnut trees [14] but is now widely used as a protein overproduction host [15] . Its genome encodes a single aquaporin , Aqy1 , and no aquaglyceroporin . In comparison , the genome of S . cerevisiae encodes two aquaporins and two aquaglyceroporins . It is common among yeast and filamentous fungi to only have one orthodox aquaporin , but it is rather uncommon to lack aquaglyceroporins [9] . Nevertheless , the human fungal pathogen Candida albicans , which is carried by 80% of the world's population with no harmful effects but can be fatal in immunocompromised patients [16] , also contains only a single aquaporin , which has high sequence similarity ( 67% ) to Aqy1 of P . pastoris . Another striking feature of Aqy1 is that it contains an extended N terminus of unknown function , a characteristic that is frequently encountered among yeast aquaporins and aquaglyceroporins . Specifically , when compared with its closest human homologue aquaporin 1 ( hAQP1 ) , Aqy1 has 34 additional N-terminal residues . There are currently nine reported high-resolution aqua ( glycero ) porin structures ( <3 . 0 Å ) within the protein structure database: three from bacteria [17]–[19] , two from bovine [4] , [20] , one from human [21] , one from sheep [22] , one from spinach [23] , and one from Plasmodium falciparum [24] . All the reported structures share a conserved overall fold with six transmembrane helices and five loops , where loops B and E dip into the channel and together form the seventh pseudo-helix . However , there are no reported structures of aquaporins from yeast or filamentous fungi . To better understand unique structural and functional characteristics of yeast aquaporins , we determined the structure of Aqy1 from P . pastoris at 1 . 15 Å resolution , which is the highest resolution structure reported to date for a membrane protein . At this resolution the anisotropic motions of water molecules and side-chains could be observed within the channel . Furthermore , the structure clearly establishes that the channel is gated , where a unique conformation of the N terminus is observed to cap the water channel entrance and thereby close the channel . This closure mechanism is distinct from that of a gated plant aquaporin , for which an extended D-loop caps the channel from the cytoplasm [23] . Measurements of the water transport activity using both proteoliposomes and spheroplasts of P . pastoris , showed that full length Aqy1 is active , whereas the N-terminal truncated form of Aqy1 has dramatically increased water conductance . From a combination of site-directed mutagenesis data and molecular dynamics simulations a unified picture emerges for the regulation of Aqy1 , whereby the water conductance is suggested to be controlled by a combination of mechanosensitive gating and post-translational regulation by phosphorylation . Thus our findings for Aqy1 combine to establish that this orthodox aquaporin is gated by a unique mechanism intimately associated with the characteristic N terminus extension of the yeast aquaporins .
To shed light on the unique functional attributes of yeast aquaporins we crystallized and solved the X-ray structure of Aqy1 from P . pastoris to 1 . 15 Å resolution ( Table 1 ) . As is common for aquaporins , Aqy1 crystallized as a tetrameric assembly in the I4 space group with one monomer per asymmetric unit , with crystals growing as stacked membrane bilayers having a gap of approximately 22 Å between the tetramer surfaces ( Figure S1 ) . In particular , the N and C termini do not take part in crystal formation and hence the conformation of these termini is not influenced by crystal packing . The tetramer consists of four independent water channels , in which each monomer has six transmembrane helices that adopt the typical “hourglass” fold [4] , [17]–[20] , [22] , [23] ( Figure 1A ) . Loops B and E insert into the membrane and together form a seventh transmembrane pseudo-helix containing the aspargine-proline-alanine ( NPA ) aquaporin signature motif near the center of the water channel . The water pore narrows on the extracellular side near the aromatic/arginine ( ar/R ) constriction region ( Figure 1A ) , which functions as a selectivity filter . The very high resolution of this structure made it possible to include anisotropic temperature factors during structural refinement . The anisotropic thermal motion is described by six anisotropic displacement parameters ( ADPs ) , and the anisotropy of an atom is then defined as the ratio of the minimum and maximum eigenvalues of the 3×3 matrix formed by the ADPs . This definition means that the ratio is 1 . 0 for a perfectly isotropic ( spherical ) motion and decreases with increased anisotropic ( nonspherical ) motion . The anisotropic thermal motions of the atoms in the water pore are represented as ellipsoids in Figure 1B . It is apparent that eight water molecules are present within the channel , and their anisotropic motions align approximately parallel with the direction of water transport . In contrast , the side chain nitrogen atoms of Asn112 and Asn224 , which belong to the dual NPA signature motif , show low anisotropy ( 0 . 6–0 . 8 for the nitrogen atoms compared to 0 . 3 for the closest water ) , and their movement is not synchronized with water molecules to which they hydrogen bond . Thus , at this resolution it is possible to visualize how the H-bond acceptor atoms of the dual NPA motif serve as rigid structural anchors , necessitating a specific orientation of water molecules as they pass and thereby helping to exclude the cotransport of protons [2]–[4] . Similarly , although water molecules near the aromatic/arginine constriction region show exceptionally large anisotropic motions , with anisotropy as low as 0 . 1 , the side-chain atoms of Arg227 adopt a well ordered conformation with anisotropy in the range of 0 . 6–0 . 8 for the side chain ( Figure 1B ) . This lack of flexibility of the side chain of Arg227 does not support the suggestion that dual conformations of the homologous arginine may gate water transport activity in AqpZ [25] . Strikingly , a functional role for the yeast aquaporin N terminus is implied by the structure of Aqy1 , since the water channel is closed on the cytoplasmic side by conserved N-terminal residues ( Figures 2 and S2 ) . Specifically , the N terminus folds such that each Aqy1 protomer is intertwined with its neighbour within the tetramer via a helical bundle , which is stabilized by multiple hydrogen bond interactions ( Figures 2C and S3 ) . The presence of the bundle contrasts with all other reported aquaporin structures where primarily hydrophobic interactions stabilize the tetramer formation . Hydrogen bond from Tyr27 anchors the bundle of Aqy1 to the aquaporin scaffold and Pro29 introduces a kink ( Figure 2A ) allowing Tyr31 to insert into the water channel . The hydroxyl group of Tyr31 partakes in an exceptionally well ordered hydrogen bond network involving two water molecules and the backbone carbonyl oxygen atoms of Gly108 and Gly109 , with residual Fobs−Fcalc electron density revealing the H-bond donor-acceptor relationships between these groups ( Figure 2B ) . A HOLE [26] representation of the pore profile reveals how the water channel narrows to 0 . 8 Å in diameter near Tyr31 , which is too small to allow the passage of water ( Figure 3B ) . Moreover , when Tyr31 of Aqy1 is substituted by an alanine there is a 6-fold increase in water transport activity ( Figure 3A ) . Equilibrium molecular dynamics simulations also establish the existence of an energetic barrier to water permeation of approximately 30 kJ/mol in this region ( Figure 3B ) , confirming the closed nature of the channel . When Tyr31 is substituted with an alanine in silico the pore widens to a diameter larger than 2 Å , which reduces the free energy barrier to water permeation to less than 13 kJ/mol ( Figure 3B ) , and water molecules enter the channel and fill the space left by Tyr31 . The blocking of the Aqy1 channel with a tyrosine residue is reminiscent of the situation found in mammalian AQP0 ( Figure S4 ) , which helps form thin junctions in the lens core of mammalian eyes . AQP0 has a water transport activity an order of magnitude lower than that typically found for other human aquaporins [27] and its structure has been solved in both a nonjunctional ( open ) form as well as in junctional ( closed ) form [20] , [22] , [28] . The closing of the pore is believed to be caused by proteolytic cleavage of AQP0 and also ( to some extent ) by the addition of calcium or changes in pH [29] . It has been suggested that subtle changes in the conformation of Tyr149 on the extracellular side may provide a mechanism whereby the water flux through the channel can be regulated [28] , [30] , [31] . A hallmark for yeast aquaporins is an extended N terminus . Whereas the N terminus of the aquaglyceroporin Fps1 from S . cerevisiae has been suggested to regulate the opening of the channel upon osmotic changes [11] , the N terminus of the orthodox yeast aquaporins has not yet been assigned any regulatory function . To investigate the functional role of the N terminus of Aqy1 , water conduction through Aqy1 was assayed using P . pastoris spheroplasts . Water transport activity measurements showed that the presence of wild-type Aqy1 in the P . pastoris membrane increases water permittivity . In particular , the transport rates for spheroplasts overexpressing Aqy1 were eight times faster than for the aqy1Δ-strain of P . pastoris ( an engineered strain with the gene encoding Aqy1 disrupted ) . Thus , although the crystal structure of Aqy1 is closed , it is an active aquaporin . However , when water transport activities were compared for strains overexpressing wild-type Aqy1 and overexpressing Aqy1 with a truncated N terminus ( Aqy1ΔN36 , 36 N-terminal residues deleted ) , it was apparent that deleting the N terminus increased the water transport activity by a factor of six ( Figure 4A ) . Since full-length Aqy1 is functionally active yet structurally closed , and its activity is increased by deleting the N terminus , we conclude that it is gated by its N terminus . Moreover , the high degree of conservation of Pro29 , Tyr31 , Gly108 , and Gly109 implies that the proposed gating mechanism is prevalent across other yeast species ( Figure S2 ) . Aquaporin expression in yeast , particularly in S . cerevisiae , is strongly correlated with freeze tolerance [12] . Deletion of the gene encoding Aqy1 in P . pastoris ( aqy1Δ ) also leads to a markedly reduced tolerance of multiple cycles of rapid freezing in liquid nitrogen followed by thawing ( Figure 4B ) . The ability of aqy1Δ to survive repeated freeze/thaw cycles is recovered when either Aqy1 or a truncated form Aqy1ΔN36 is overproduced ( Figure 4B ) . Surprisingly , there is no obvious advantage when reintroducing the AQY1 gene over introducing the AQY1ΔN36 gene . Nevertheless , overproduction of either protein ( Aqy1 and Aqy1ΔN36 ) in the wild-type strain , for which endogenous Aqy1 is already present , did show detrimental effects on freeze/thaw survival for the permanently open Aqy1ΔN36 constructs ( Figure 4C ) . These observations may be reconciled by accepting that ( in this context ) the survival benefits resulting from having at least one aquaporin present in the membrane masks the potential disadvantages of this aquaporin being permanently open . Taken together , these findings establish that water conduction through Aqy1 greatly enhances the survival of P . pastoris during rapid freezing , yet the effect is delicately balanced as the introduction of a permanently open aquaporin actually reduces the chances of survival . Thus , Aqy1 gating appears to mediate a compromise by balancing these conflicting effects by increasing water flux through the cell membrane in some contexts , and reducing it in others . Channel gating facilitates a rapid response to external stimuli when other regulatory mechanisms , such as transcriptional regulation or trafficking , are too slow . Since the gated plant plasma membrane aquaporin SoPIP2;1 is believed to be regulated by phosphorylation [32] , we investigated the effects of mutation of candidate Aqy1 phosphorylation sites . Water transport assays in spheroplasts reveal a significant increase in water transport activity when Ser107 of Aqy1 was substituted by aspartate ( where the aspartate mimics a putative phosphorylation event ) , yet water transport rates comparable to the wild type were recovered when it was substituted by an alanine ( Figure 3A ) . Ser107 lies within a consensus phosphorylation site ( according to NetPhos 2 . 0 Server [33] ) situated near the pore channel and is involved in an important network of hydrogen bonds involving Tyr31 ( Figure 5 ) . Molecular dynamics simulations of S107D Aqy1 mutant also show a widening of the pore near these residues , which increases to a diameter larger than 2 Å , and a corresponding reduction in the free energy barrier to water permeation to less than 13 kJ/mol ( Figure 3B ) . In these simulations water molecules establish a single-file water column between Pro29 , Tyr31 , Tyr104 , Leu189 , Ala190 , and Val191 , after a local rearrangement of the latter three residues , which are located in the lower part of helix 4 towards loop D ( Figure 6A ) . In contrast , simulations of the Y31A Aqy1 mutant allowed water molecules enter the channel , filling the space left by Tyr31 . Thus both molecular dynamics simulations and functional data support the suggestion that Ser107 is a putative phosphorylation site , and that it induces an opening of the pore upon phosphorylation . A putative mechanism of Aqy1 phosphorylation-regulated gating could explain the apparently low water transport activity of wild-type Aqy1 in P . pastoris spheroplasts ( Figure 4A ) . Nevertheless , this idea does not explain why Aqy1 appears to have higher water-transport activity when purified and reconstituted into proteoliposomes than it has in its native membrane . In particular , when both Aqy1 and Aqy1ΔN36 ( N-terminal truncated form ) are reconstituted into proteoliposomes their water transport activities are almost indistinguishable ( measured rate constants of 12 . 9 s−1 and 10 . 9 s−1 , respectively; Figure S5 ) , yet Aqy1 has a water transport activity of only one-sixth of that found for Aqy1ΔN36 in P . pastoris spheroplasts ( Figure 4A ) . A clue hinting at an explanation for this apparent paradox is provided by the unique global topology of the Aqy1 tetramer , which has the N termini intertwined in a helical bundle ( Figure 2C ) in a manner that is reminiscent of the arrangement found for the mechanosensitive gated ion channel MscL from Mycobacterium tuberculosis ( Figure S6 ) [34] . Mechanosensitivity has also been demonstrated for aquaporins found in plant roots [35] . Should Aqy1 be a mechanosensitive channel then this could accommodate the observation that it is more active in proteoliposomes than in its native membrane , because these proteoliposome vesicles are highly curved ( 120–130 nm in diameter; compared to spheroplasts with a diameter of 1–5 µm ) . Increased membrane curvature can activate mechanosensitive channels , as could destabilisation effects owing to the removal of native lipids . To test this hypothesis , nonequilibrium molecular dynamics simulations of Aqy1 were performed in a solvated lipid bilayer being subject to external mechanical stress , either by increasing the lateral pressure up to 10 bar , or by bending the membrane towards the cytoplasmic side ( Figure S7 ) . Spontaneous opening events of one monomer in both simulations were observed ( Figure 6A ) with the pore diameter near Tyr31 widening from 0 . 8 Å in the crystal conformation to values larger than 2 Å in both simulations ( Figure 6B ) . In addition , the energetic barrier for the water permeation dropped substantially compared to the control simulations in which no opening was observed ( Figure 6B ) .
Since all unicellular life forms have a high surface-to-volume ratio , the intrinsic rate of water flow across their cell membranes was not likely to be rate limiting . Thus , the discovery of aquaporins in yeast was unexpected . In this study we demonstrate at least one physiological role for the aquaporin Aqy1 from P . pastoris , which imparts freeze tolerance to the host . Specifically , deletion of the gene causes a freeze-sensitive phenotype that can be remedied by reintroduction of the P . pastoris AQY1 gene or the N-terminal truncated form of the AQY1 gene ( Figure 4A ) . In contrast , overproduction of the N-terminal truncated form of Aqy1 in the wild-type strain of P . pastoris results in a detrimental effect on freeze/thaw survival . Thus , the possibility to open and close the N terminus seems to be optimal for the survival of cells in freezing environments . Although yeast aquaporins have been extensively studied , there are no reports concerning the gating of orthodox yeast aquaporins . Nevertheless , Soveral et al . suggest that the water transport activity of Aqy1 from S . cerevisiae can be regulated by membrane tension [36] , implicating gating . It has also been established that the aquaglyceroporin Fps1 of S . cerevisiae is rapidly regulated by osmolarity [13] , which also implicates gating . The most conclusive functional evidence for aquaporin gating concerns the plant plasma membrane aquaporins [37] . In that case the water channel is closed by a unique conformation of loop D , which folds over the cytoplasmic entrance and blocks the channel ( Figure 7 ) [23] . In contrast , our crystal structure of Aqy1 , captured in a closed conformational state , shows that a tyrosine residue from the N terminus enters the pore ( Figure 7 ) and blocks the opening of the channel by hydrogen bonding to a water molecule within the pore . According to sequence alignment , the N termini of aquaporins from yeast are well conserved ( Figure S2 ) , indicating that other yeast aquaporins may also share this N-terminal gating mechanism . In particular , the aquaporin from the fungal pathogen of humans C . albicans has all important amino acids conserved , suggesting that C . albicans aquaporin is gated in a similar manner as Aqy1 from P . pastoris . Gated ion channels [38] participate in rapid physiological responses such as signal propagation . Likewise , the gated plant aquaporin SoPIP2;1 responds rapidly to cellular acidification as a result of flooding [6] . Our results show that Aqy1 is essential for yeast survival during rapid freezing and thawing , and for organisms of 1 to 10 µm in diameter this can occur within microseconds . Thus the survival benefits inferred by Aqy1 imply that channel opening must either be extremely rapid or that there are constitutively open channels in vivo . Since both the Aqy1S107D and N-terminal deleted Aqy1 ( Aqy1ΔN36 ) mutants show considerably higher water transport activity than wild-type Aqy1 ( Figures 3A and 4A ) , it is strikingly inefficient to express large quantities of Aqy1 in vivo , which has relatively poor water transport activity . From this perspective we argue that the observed host survival benefits arise from Aqy1 being quickly opened in response to rapid freezing . pH-triggered opening , which is known to close SoPIP2;1 [6] , can be excluded as a possible mechanism for Aqy1 since high-resolution structures recovered at both acidic and basic pH are identical , with a root mean square deviation of only 0 . 08 Å for 248 Cα atoms ( Figure S8 ) . Thus mechanosensitive gating emerges as a plausible mechanism . This possibility is supported by molecular dynamic simulations , which show how Aqy1 can be regulated by both the surface tension and membrane curvature ( Figure 6 ) . An intriguing question for the mechanism of mechanosensitivity is how a mechanical signal is transmitted from the membrane to the gate of the channel in the cytoplasmic portion of helix 4 , which is not in direct contact with the membrane . To address this question , a principal component analysis of the backbone atoms of the cytoplasmic halves of helices four , five , six , and loop D was carried out . Projections of both membrane-mediated stress trajectories onto the principal eigenvector show an abrupt transition for the monomers in which an opening takes place ( Figure 6C ) , suggesting that external forces triggering gating are transmitted from the lipid membrane to Leu189 , Ala190 , and Val191 via coupled movements of the helices four , five , and six , the latter being in direct contact with the membrane . To investigate the causal relation between this global conformational change and channel opening , the position along this principal eigenvector was artificially driven from the “closed” to the “open” conformation in an additional set of simulations , leading to reproducible channel opening in 14 of 16 cases ( Figure S9 ) . It is intriguing to consider how the two putative modes of regulation , by both phosphorylation of a serine residue or changes in the surface tension and curvature of the membrane ( Figures 3A and 6B ) , may relate to each other . Strikingly , the opening transition for the Aqy1S107D simulation , which mimics a phosphorylated S107 , involves similar movements of residues Leu189 , Ala190 , and Val191 as in the simulations with the membrane being subject to external mechanical stress ( Figure 6A ) . These similarities are revealed in a principal component analysis of the Aqy1S107D trajectory , where the projection onto the first eigenvector , upon opening , drops in a manner similar to simulations with induced membrane stress ( Figure 6C ) . These findings disclose that both phosphorylation and external membrane-mediated mechanical stress may induce the same opening mechanism , and suggest how mechanical stimuli are transmitted from the membrane to the gate of the channel . Post-translational regulation of Aqy1 by phosphorylation may also be exploited in other physiological contexts , when less rapid changes in water transport activity are required . Hence , phosphorylation may fine-tune the water flux during normal conditions of growth , whereas mechanosensitive gating could provide a rapid pressure valve in response to unexpected shocks . Rapid freezing or thawing and sudden osmotic changes are frequently encountered by micro-organisms [8] . For example , actions of warm blooded animals in cold environments , such as breathing , coughing , walking , and foraging , can expose micro-organisms to large temperature shocks . Likewise , dramatic changes in osmolarity arise when microbes encounter ripe fruit or rainwater . Indeed , the aquaglyceroporin Fps1 of S . cerevisiae is suggested to be rapidly opened when the host is exposed to sudden osmotic changes [11] . Thus the evolution of gated aquaporins and aquaglyceroporins would provide an economic solution to numerous stresses associated with rapidly changing environments , aiding the organism's quest to adapt and survive .
The AQY1-deficient strain was created by insertion of HIS4 into AQY1 , disrupting the AQY1 ORF . The S . cerevisiae HIS4-promoter and ORF were amplified by PCR ( primers A and B [Table S1] ) . PstI ( New England BioLabs ) linearized pPICZαB-AQY1-vector and PstI-digested PCR product were ligated and transformed into Escherichia coli . The created AQY1-HIS4-AQY1 deletion construct was amplified by PCR ( primers C and D ) ( Table S1 ) and transformed into P . pastoris GS115-his4 according to Invitrogen's manual [39] . Transformants were selected for histidine prototrophy and disruption of AQY1 was confirmed by PCR . His6-tagged constructs of Aqy1 and Aqy1ΔN36 , i . e . Aqy1 lacking the soluble part of the N terminus ( first 36 residues ) , were prepared from genomic DNA by PCR using primers E and F ( Table S1 ) , respectively . Antisense-primer G was used for fusing a C-terminal His6-tags to the ORFs . PCR-products and pPICZαB-vector ( Invitrogen ) were digested with EcoRI/XbaI ( New England BioLabs ) and ligated , creating pPICZαB-AQY1-His6 and pPICZαB-AQY1ΔN36-His6 , respectively . The vectors were linearized using PmeI and transformed into wild-type and AQY1-deficient P . pastoris strains . Transformants were selected for on Zeocin-containing plates . Point mutations of Aqy1 ( Y31A , S107D , and S107A ) were created by site-directed mutagenesis using the Quikchange II- and Quikchange Lightning-kits ( Stratagene ) . All cultures were grown in a 3 l bioreactor and the protein expression was induced by methanol according to standard procedure ( Invitrogen ) . Cells were lysed using X-press equipment ( AB Biox ) and resuspended in Breaking Buffer ( 50 mM KH2PO4/K2HPO4-buffer , [pH 7 . 5] , 5% glycerol ) . After removal of cell debris by centrifugation at 15 , 000g , 20 min , 4°C , the membrane fraction was collected by centrifugation at 138 , 000g , 90 min , 4°C and washed with breaking buffer . The membranes were solubilized with beta-octylglucopyranoside ( 5% β-OG , 20 mM Tris [pH 8 . 0] , 100 mM NaCl , 20% glycerol , 0 . 5 mM EDTA ) . Unsolubilized material was removed by centrifugation at 15 , 000g , 20 min , 4°C . The His-tagged protein constructs were purified for the liposome assay ( Figure S5 ) by immobilized metal ion affinity chromatography ( IMAC ) using Ni-NTA resin ( Qiagen ) and eluted with 500 mM imidazole in 20 mM Tris , ( pH 8 . 0 ) , 100 mM NaCl , 20% glycerol , 1% β-OG . Further purification was performed by size-exclusion chromatography using a Superdex200 column ( GE Healthcare ) . The purity of the samples was verified by SDS-PAGE and MS . For crystallization , nontagged , endogenously expressed Aqy1 was used . The protein was purified using ion exchange column ResourceQ ( GE Healthcare ) and size exclusion chromatography on Superdex200 ( GE Healthcare ) according to Nyblom et al . [40] . The protein was concentrated to 9 mg/ml and crystallization trials were set up using the hanging drop vapour diffusion method at 4°C , with drops containing equal volumes of protein- and reservoir-solution . Crystals appeared in two different conditions , at different pH . Crystal1 ( Figure S10 ) grew in 28% PEG400 , 100 mM Na-citrate ( pH 3 . 5 ) , 200 mM Li2SO4 , while Crystal2 ( Figure S10 ) grew in 33% PEG400 , 100 mM Tris ( pH 8 . 0 ) , 100 mM NaCl , 100 mM CdCl2 . The crystals were frozen in liquid nitrogen and diffraction data were collected under cryoconditions at 0 . 873 Å on beamline ID23-2 at the European Synchrotron Radiation Facilities ( ESRF , Grenoble , France ) and at 1 . 04 Å at MaxLAB ( Lund , Sweden ) . High- and low-resolution data for Crystal1 ( 1 . 15 Å ) were processed and scaled using XDS [41] , and a molecular replacement solution was found using Phaser [42] with bovine AQP1 ( Protein Data Bank [PDB , http://www . rcsb . org/pdb] accession code , 1J4N ) as search model . Initial model building was carried out with Arp/Warp [43] , and further model building was performed in COOT [44] . To verify the solution , unbiased composite omit maps were calculated using CNS [45] . Initial refinement was made using Refmac5 [46] , while SHELXL [47] was used in the last stage of refinement . The model was refined anisotropically ( with exception of β-OG , which was refined as rigid body ) and hydrogen-atoms were added to protein atoms as a riding model . No hydrogen atoms were added along the water-channel to allow for the correct observation of difference density . In the Ramachandran [48] plot 91 . 8% and 8 . 2% of the residues fell in favored and allowed regions respectively . Data from Crystal2 ( 1 . 4 Å ) were processed using MOSFLM [49] . Phaser was used to find a molecular replacement solution using bovine AQP1 ( PDB accession code , 1J4N ) . Model building was carried out using Arp/Warp and COOT . The refinement , including simulated annealing with omit map calculations and restrained anisotropic refinement , was performed using CNS and Refmac5 . In the Ramachandran plot 92 . 7% and 7 . 3% of the residues fell in favored and allowed regions respectively . Data collection and refinement statistics are shown in Table 1 . For freezing and thawing experiments P . pastoris cells were grown in BMMY ( Invitrogen ) , starting at OD600 = 1 , for 24 h at 30°C for induction of protein production . In total , five cycles of freezing aliquots of 1 ml , OD600 = 5×10−4 , in liquid nitrogen and subsequential thawing in a water bath at 30°C were followed by spreading aliquots on YPD plates . The spheroplast water transport assays is based on the shrinkage of P . pastoris cells with a destabilized cell wall after subjection to a hyperosmolal solution . P . pastoris spheroplasts [50] were created in biological triplicates by inducing recombinant protein production through growth in BMMY for 22 h , starting at OD600 = 1 . The cells were harvested and incubated with TE-buffer ( 100 mM Tris [pH 8 . 0] , 1 mM EDTA ) and 0 . 5% β-mercaptoethanol for 1 . 5 h to destabilize the cell wall . Subsequently cells were washed with 1 . 2 M Sorbitol , 20 mM MES ( pH 6 . 5 ) , and resuspended to final OD600 = 5 . The shrinkage upon mixture was observed by light scattering at 90° angle in a stopped-flow apparatus ( Biologic Instruments Inc . ) at 435 nm . The water transport assay is based on the shrinkage of proteoliposomes after subjection to a hyperosmolal solution . Liposomes were prepared according to Nyblom et al . [40] . In brief , liposomes were created from E . coli polar lipid extract ( Avanti ) by sonication and the protein was reconstituted using β-octylglucopyranoside , which was subsequently removed by dilution . The shrinkage was observed by measuring light scattering at a 90° angle in a stopped-flow apparatus ( Biologic Instruments Inc . ) at 480 nm . Rate constants were obtained by curve-fitting to 1 s of raw data using the equation: y = A1×exp ( −k× ( t−td ) ) +A2 ( A1 = amplitude , k = rate constant , A2 = vertical offset , td = constant time delay [s] ) . Western blot analysis of the spheroplasts used for the water transport assay was performed according to ECL Plus manual ( GE Healthcare ) with His6 monoclonal primary antibody ( Clontech ) on the membrane fraction of the constructs . The cells were lysed with glass beads in cell resuspension buffer CRB ( 20 mM Tris [pH 8 . 0] , 100 mM NaCl , 0 . 5 mM EDTA , 5% glycerol ) using a FastPrep-24 ( M . P . Biomedicals ) . Cell debris was removed by centrifugation at 10 , 000g , 30 min , 4°C , and the membrane fraction was collected from the supernatant by centrifugation at 100 , 000g , 90 min at 4°C . The membranes were dissolved in CRB containing 5% β-OG and the concentration was adjusted to equal Abs260 nm values with CRB+β-OG to ensure same amounts of membrane were loaded ( Figure S11 ) . Molecular dynamics simulations ( Text S1 ) were carried out using the GROMACS simulation software [51] , [52] , with the Aqy1 tetramer embedded in a solvated POPE bilayer ( Figure S7 ) . Five different simulations were carried out: first , under equilibrium conditions , without surface tension ( I ) ; second , mutating Ser107 into Asp , mimicking a putative phosphorylated state ( II ) ; third , mutating Tyr31 into Ala ( III ) ; fourth , inducing a surface tension onto the membrane ( IV ) , and fifth , bending the membrane towards the cytoplasmic side ( V ) . The simulation boxes contain the protein [53] , [54] , POPE lipids [55] , SPC water molecules [56] , and chloride ions to neutralize the system . Both the temperature and pressure were kept constant by coupling the system to an external bath at a temperature of 300 K and a pressure of 1 bar [57] , respectively . To induce a surface tension in simulation IV , the pressure in the xy plane ( parallel to the membrane surface ) was increased to 10 bar . To bend the membrane in simulation V , lipids at distances larger than 5 nm from the center of the tetramer were pushed along the z coordinate , exerting an external force that resulted in an acceleration of 0 . 01 nmps−2 . A force of equal magnitude and in the opposite direction was exerted on the tetramer for compensation . The system was equilibrated for 1 ns by maintaining the coordinates of the protein harmonically restrained . The simulation length was 100 ns and 10 ns for the simulations I to IV and V , respectively . Pore diameter profiles were obtained with the HOLE software [26] . Free energy profiles were computed using the formula G ( z ) = −kBTln〈n ( z ) 〉 , where kB is the Boltzmann constant , T is the temperature , and 〈n ( z ) 〉 is the water density as a function of the pore coordinate . A principal component analysis , consisting of the calculation and diagonalization of the covariance matrix for the coordinates of the backbone atoms of the lower part of helices four , five and six , and loop D [58] , was carried using GROMACS . Protein Data Bank ( http://www . rcsb . org/pdb ) coordinates for Aqy1 ( low and high pH structures ) have been deposited with accession codes 2W2E and 2W1P . | All living organisms must regulate precisely the flow of water into and out of cells in order to maintain cell shape and integrity . Proteins of one family , the aquaporins , are found in virtually every living organism and play a major role in maintaining water homeostasis by acting as regulated water channels . Here we describe the first crystal structure of a yeast aquaporin , Aqy1 , at 1 . 15 Å resolution , which represents the highest resolution structural data obtained to date for a membrane protein . Using this structural information , we address an outstanding biological question surrounding yeast aquaporins: what is the functional role of the amino-terminal extension that is characteristic of yeast aquaporins ? Our structural data show that the amino terminus of Aqy1 fulfills a novel gate-like function by folding to form a cytoplasmic helical bundle with a tyrosine residue entering the water channel and occluding the cytoplasmic entrance . Molecular dynamics simulations and functional studies in combination with site-directed mutagenesis suggest that water flow is regulated through a combination of mechanosensitive gating and post-translational modifications such as phosphorylation . Our study therefore provides insight into a unique mechanism for the regulation of water flux in yeast . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"biochemistry",
"biophysics/macromolecular",
"assemblies",
"and",
"machines",
"cell",
"biology/cellular",
"death",
"and",
"stress",
"responses",
"biophysics/theory",
"and",
"simulation",
"biochemistry/macromolecular",
"chemistry",
"molecular",
"biology/bioinformatics",
"biochemis... | 2009 | Crystal Structure of a Yeast Aquaporin at 1.15 Å Reveals a Novel Gating Mechanism |
Megacystis-microcolon-intestinal hypoperistalsis syndrome ( MMIHS ) is a rare disorder of enteric smooth muscle function affecting the intestine and bladder . Patients with this severe phenotype are dependent on total parenteral nutrition and urinary catheterization . The cause of this syndrome has remained a mystery since Berdon's initial description in 1976 . No genes have been clearly linked to MMIHS . We used whole-exome sequencing for gene discovery followed by targeted Sanger sequencing in a cohort of patients with MMIHS and intestinal pseudo-obstruction . We identified heterozygous ACTG2 missense variants in 15 unrelated subjects , ten being apparent de novo mutations . Ten unique variants were detected , of which six affected CpG dinucleotides and resulted in missense mutations at arginine residues , perhaps related to biased usage of CpG containing codons within actin genes . We also found some of the same heterozygous mutations that we observed as apparent de novo mutations in MMIHS segregating in families with intestinal pseudo-obstruction , suggesting that ACTG2 is responsible for a spectrum of smooth muscle disease . ACTG2 encodes γ2 enteric actin and is the first gene to be clearly associated with MMIHS , suggesting an important role for contractile proteins in enteric smooth muscle disease .
Berdon first described patients with a severe phenotype characterized by smooth muscle functional failure in 1976 , at a time when total parenteral nutrition ( TPN ) was becoming common clinical practice [1] . Berdon noted that because the functional intestinal defect could not be corrected , finding the cause of the disorder described as megacystis-microcolon-intestinal hypoperistalsis syndrome ( MMIHS , OMIM 249210 ) would be necessary to avoid keeping patients with the disorder as “prisoners” of TPN without otherwise effective treatments . For thirty years , genetic , pathologic , endocrine , and physiologic studies have sought to determine the cause of MMIHS without success , and the clinical history of patients with these disorders often fulfills Berdon's prediction as patients remain on long-term TPN , and the underlying etiology remains unknown . Clinically , MMIHS is characterized by prenatal bladder enlargement , neonatal functional gastrointestinal obstruction , and chronic dependence on total parenteral nutrition ( TPN ) and urinary catheterization [2]–[4] . Patients undergo repeated abdominal surgeries , suffer hepatic complications from TPN , and are susceptible to poor nutrition , as well as infectious complications of ileostomies and intravenous and urinary catheters . The first challenge in understanding the genetics of MMIHS has been in characterizing the clinical phenotype . MMIHS is part of a phenotypic spectrum that includes intestinal pseudo-obstruction [5] ( OMIM 155310 , 609629 ) , hollow visceral myopathy [6] , [7] ( OMIM 609629 ) , pseudo-Hirschsprung disease [8] , and irritable bowel syndrome [9] . Functional gastrointestinal obstruction is also frequently observed associated with other abnormalities such as prune-belly syndrome ( OMIM 100100 ) , external ophthalmoplegia ( OMIM 277320 ) , and Barrett esophagus ( OMIM 611376 ) . However , there is uncertainty about the extent to which locus heterogeneity and variation in expression underlie this clinical variability [9] . In addition , a number of single case reports have proposed an association of MMIHS with other disorders such as trisomy 18 [10] , cardiac rhabdomyomas [11] , and deletion of 15q11 . 2 [12] . However , in these cases it is unclear whether these genetic disorders are related to MMIHS or are coincidental findings . Autosomal recessive inheritance of MMIHS ( OMIM 249210 ) has been suggested in numerous cases based on the presence of two affected siblings [3] , [13] , [14] , consanguinity [15] or both [16]–[19] , but no genes have been identified to date , although in retrospect a report of a dominant mutation in the ACTG2 enteric actin gene in a Finnish family with adult onset visceral myopathy has proved to be relevant [20] . Pathologic studies of intestine in MMIHS have similarly been inconclusive [4] . Some studies demonstrate abnormalities of the circular and longitudinal layers of the muscularis propria [21] , [22] , while others focus on abnormalities of ganglion cells , including reduced [4] , increased [4] , hypertrophic [23] , immature , or dysplastic ganglia [4] , [19] . An imbalance between intestinal hormones in cases of MMIHS has also been noted [24] . Finally , the intrinsic pacemaker cells of the gastrointestinal tract , the interstitial cells of Cajal , were noted to display abnormalities in MMIHS [25] , [26] . Given this broad range of findings , there has been controversy over which pathological changes in the gastrointestinal tract are primary versus secondary [4] . Additional insight into the genetic basis of MMIHS appeared to come from a mouse model of the disease . Mice lacking expression of the α3 subunit of the neuronal nicotinic acetylcholine receptor encoded by the Chrna3 gene and mice lacking both the β2 and β4 subunits encoded by the Chrnb2 and Chrnb4 genes , respectively , displayed megacystis , failure of bladder strips to contract in response to nicotine , widely dilated ocular pupils , growth failure , and perinatal mortality [27] , [28] . These subunits are expressed in various sympathetic and parasympathetic ganglia , and lack of transmission at these ganglia could explain the lack of contraction of involuntary smooth muscle . A role for the α3 subunit was further suggested when reduced mRNA levels were measured by in situ hybridization , and reduced immunostaining for protein was possibly found in tissues from MMIHS patients [29] . However , antibodies against the neuronal nicotinic receptor subunits are notoriously unreliable [30] and a specific search for mutations in CHRNA3 and CHRNB4 in many of the patients studied herein did not identify any potential disease-causing mutations [31] .
Since the findings in mice harboring mutations in Chrna3 or in Chrnb2 and Chrnb4 combined caused MMIHS-like phenotypes , we have conducted a study of MMIHS aimed at identifying the genetic cause . We collected samples from patients with MMIHS and related phenotypes , some of whom have been previously reported [31] . Our cohort of 34 families to date includes 27 DNA samples from probands including individuals diagnosed with MMIHS ( 20 probands ) as well as intestinal pseudo-obstruction ( 4 probands ) , prune belly syndrome ( 2 probands ) , and hollow visceral myopathy ( 1 proband ) . Examples of radiologic findings are shown in Figure 1 . Study recruitment has taken place over a period of 14 years . We undertook whole-exome sequencing in 11 unrelated probands . The exome sequencing characteristics are summarized in Table 1 . Of the 11 probands , eight were diagnosed with MMIHS and three diagnosed with intestinal pseudo-obstruction . We identified heterozygous missense variants in the ACTG2 gene encoding γ2 enteric actin in six of the 11 individuals . We reasoned that ACTG2 was an excellent candidate for MMIHS as a thin filament protein in the sarcomere involved in muscular contraction . We therefore undertook Sanger sequencing of all the exons and intron-exon boundaries of ACTG2 in 16 additional probands in our cohort . The results for all the heterozygous ACTG2 variants in our cohort are summarized in Table 2 . All of these variants were unique to our cohort , as none of the ACTG2 variants found in our patients , were present neither within the 1000 Genomes project data ( http://browser . 1000genomes . org/index . html ) nor within the NHLBI Exome Sequencing Project ( http://evs . gs . washington . edu/EVS/ ) . In addition , within our internal data , excluding the cases presented here , none of these variants were found in approximately 1900 other samples analyzed by the Baylor-Hopkins Center for Mendelian Genomics ( http://www . mendelian . org/ ) nor within 1200 clinical samples analyzed in the BCM clinical laboratory . We did identify a number of other novel heterozygous variants , but these were all distinct from the variants seen in our MMIHS cohort ( Table S1 ) . Within our group , 15 probands had mutations in ACTG2 in comparison to 12 probands in the cohort without mutations in ACTG2 . Of note , we observed ten apparently de novo events ( Figure 2 ) . We observed 6 novel C>T transition mutations at CpG dinucleotides affecting arginine amino acid residues including a recurrent mutation ( c . 769C>T; p . R257C ) seen in 3 de novo cases ( Fam4-1 , Fam30-1 , Fam25-1 ) . The clinical characteristics of the patients with apparent de novo mutations are summarized in Table 3 and in Text S1 . The age at the time of follow-up was from less than one year to 25 years . In the ten apparently de novo cases , seven patients were diagnosed with megacystis prenatally , and two of these underwent fetal surgery . The three individuals without megacystis were nonetheless dependent on catheterization of the bladder long-term . Prune-belly syndrome was observed in one of the cases ( Fam16-1 ) . The gastrointestinal manifestations were similarly severe . Of the ten apparent de novo cases , seven had bilious vomiting as a neonate , and eight were diagnosed with intestinal malrotation . All ten patients had multiple abdominal surgeries ( Table 4 ) . Long-term dependence on TPN was a consistent feature , but did not extend throughout life for all the patients . Two patients had very intermittent TPN requirements , usually during surgical recovery . Another patient ( Fam26-1 ) had an interval of improvement at age four years followed by reinitiation of TPN at six years . Interestingly , of the ten apparent de novo patients , three reported partial but significant clinical improvement on cisapride , a serotonin 5-HT4 receptor agonist and gastroprokinetic agent ( see Text S1 ) . Recently this drug was removed from the market for cardiac side effects , but one patient continued on the drug as an FDA-approved case of compassionate use , and the two others indicated strong desires to remain on the drug despite the risks . Two other families found that the same drug did not have a significant effect . Clinical outcomes in our cohort differ from the 19 . 7% survival rate reported in the literature [2] . Of the ten apparent de novo cases , nine were alive at the time of last follow-up; while one individual died at age 11 after multiple episodes of pancreatitis . One individual had undergone an intestinal transplant , and one individual was wait-listed for combined intestinal and liver transplant . Of the nine surviving individuals , eight had ileostomies . The oldest survivor amongst the apparent de novo cases was 25 years old , while the oldest previously reported case of MMIHS was 24 years [2] . While this suggests improved survival in our cohort , we observed frequent abdominal surgery , and dependence on TPN and chronic catheterization , suggesting that improvements in supportive care over time rather than a milder phenotype associated with ACTG2 mutations are responsible for this improved survival . In addition to these apparent de novo cases , five other probands were heterozygous for ACTG2 variants; in three of these cases the mutation was inherited from one of the parents . In one case the inheritance remains unknown as parental samples are not available . In an additional family ( Fam19 ) , the proband and an affected sibling both carry a heterozygous mutation that affects an alternate exon 4 ( c . 330C>A; p . F110L ) of a predicted ACTG2 short isoform . The proband ( Fam19-1 ) had multiple abdominal surgeries for obstruction and long-standing hypomotility . She has been on TPN intermittently since age 17 years , but has not required bladder catheterization . The sibling ( Fam19-4 ) suffered years of intestinal symptoms and underwent an endoscopy suggesting gastroparesis . However she had not had any abdominal surgery . No parental clinical information was available . The mother does not carry this mutation , and a paternal sample is not available . The data from this family suggests but perhaps do not prove entirely that the alternative exon 4 which would result in a very short protein isoform is functionally important . In all three inherited cases , we observed mutations identical to those identified in the apparent de novo cases ( c . 119G>A; p . R40H , c . 533G>A; p . R178H and c . 769C>T; p . R257C ) . The clinical findings in the familial cases and in the affected parents were notable for milder disease , more frequently classified as intestinal pseudo-obstruction . In one family ( Fam34 ) , the proband inherited the mutation ( c . 119G>A; p . R40H ) from the father , who had no history of megacystis or neonatal abdominal surgery but had two abdominal surgeries in adulthood for episodes of gastrointestinal obstruction . Multiple paternal relatives over four generations carried the same mutation and were noted to have bowel and bladder dysfunction segregating as an autosomal dominant mutation ( Figure 2 ) . In another family , the proband ( Fam 13-1 ) inherited the mutation ( c . 769C>T; p . R257C ) from the mother ( Figure 2 ) . The proband had prenatal megacystis but had a period of normal bowel and bladder function in the first years of life . She eventually experienced progressive pseudo-obstruction and ultimately died at age 13 years . Medical records were not available for her mother , but a history of gastrointestinal disease and a diagnosis of irritable bowel syndrome were reported . The mutations we observed extended from exon 2 to exon 7 of the transcript ( Figure 3A ) . Alignment of all six human actin proteins , which are highly conserved , revealed identity among all the amino acids in which we observed substitutions . All of these genes are implicated in human disease . De novo missense mutations in ACTB and ACTG1 underlie the brain malformation syndrome Baraitser-Winter ( OMIM 243310 ) [32] , [33] . Mutations in ACTC1 are implicated in a range of cardiac phenotypes including cardiomyopathy ( OMIM 613424 ) [34] , cases of nemaline myopathy ( OMIM 161800 ) are due to mutations in ACTA1 which can be dominantly or recessively inherited depending on the mutation [35] , and ACTA2 mutations are associated with incompletely penetrant dominantly inherited aneurysm and dissections ( OMIM 611788 ) . We compared the ACTG2 mutations in our cohort to mutations from these disorders in the identical amino acid position along the actin filament ( Figure 3B ) . There was clearly alignment between multiple mutations observed in our cohort and those implicated in Baraitser-Winter , nemaline myopathy , and thoracic aortic aneurysms and dissections . We also explored the question of the phenotypic consequences of haploinsufficiency at ACTG2 . We identified an incidental nonsense mutation ( c . C187T:p . R63X ) in ACTG2 in our internal exome sequencing database from a group of approximately 1900 individuals in the Center for Mendelian Genomics at Baylor College of Medicine ( Table S1 ) . The individual was an unaffected parent from a study of an unrelated disorder . This individual reported mild intermittent constipation , not requiring medications , and no history of abdominal surgery , or bladder dysfunction . In aggregate , these data suggest that ACTG2 haploinsufficiency is not clinically significant , although a mild phenotype with incomplete penetrance cannot be excluded . For all of the genotype-phenotype relationships shown in Figure 3B , virtually all of the reported mutations are missense with very few or no examples of frameshift or nonsense mutations . One reason for this bias might be that heterozygous loss-of-function mutations are benign or cause a different phenotype for these actin loci . Out of the 10 variants in ACTG2 identified in 14 unrelated probands , six result from C>T transitions at CpG dinucleotides altering an arginine codon ( Figure 4A ) . We observed that the CGC codon encodes 33% of the arginine residues in the γ-actin protein compared to 18% of arginine residues genome wide ( Figure 4B ) [36] . One explanation for the pattern of codon usage could relate to the expression of actin genes , as more highly expressed genes have been observed to have significantly skewed codon usage [37] . Given the presence of multiple CpG dinucleotides due to this pattern of codon usage , we surveyed paternal age in our de novo cases . We observed an average paternal age of 32 . 7 years amongst the families with de novo mutations with a standard deviation of 6 . 7 years which is not sufficient to conclude statistically whether these ten apparently de novo mutations may be associated with advanced paternal age .
Identification of ACTG2 mutations underlying a significant proportion of MMIHS and intestinal pseudo-obstruction has significance for three major reasons . First , autosomal dominant rather than autosomal recessive mutations now are known to be present in the majority of families ( 15 of 26 probands reported in this study ) . Many cases in the literature as well as the Online Mendelian Inheritance in Man database suggest autosomal recessive inheritance . While other loci exhibiting recessive inheritance are possible , nearly half of our cases of MMIHS appear to follow a dominant or sporadic pattern of inheritance with heterozygous mutations segregating with the phenotype . Second , the phenotypic spectrum for disease causing mutations in ACTG2 can now be relatively well defined . All of the apparent de novo cases had clear indications of severe smooth muscle disease with prenatal or neonatal onset , urinary catheterization , and dependence on TPN . However , phenotypic variability and distinct complications existed such that prune belly syndrome , MMIHS , hollow visceral myopathy , and intestinal pseudo-obstruction were all diagnosed . Reorganizing these clinical entities into a spectrum of ACTG2 related disorders will be of great benefit in understanding the natural course of these diseases . Third , a better understanding of the pathophysiology could lead to treatment opportunities . It was notable that three individuals reported here with MMIHS due to apparent de novo mutations in ACTG2 reported clinical improvement on cisapride , a serotonergic 5HT4 agonist . Serotonergic activity at the 5HT4 receptor has been noted to be a therapeutic target for constipation and irritable bowel syndrome [38] . Cisapride was proposed as a therapy for MMIHS in 1991 [14] . Subsequent case reports have noted failure of cisapride to produce clinical improvement in MMIHS [39] , [40] . However in our cohort three patients out of five who had been prescribed cisapride reported improved motility by subjective clinical improvement . Two of the three patients were taken off this drug for cardiac side effects . Given that new 5HT4 agonist agents with safer cardiac profiles are under study [41] , patients with ACTG2 mutations may be candidates for this therapeutic strategy . Also , the mutational data suggest the feasibility of selective knock-down of the mutant transcript using antisense oligonucleotides , offering some therapeutic hope to affected individuals and families [42] . Finally our study effectively refocuses the study of MMIHS back to the contractile apparatus of the smooth muscle in an analogous way to how cardiomyopathy and myopathy tie to muscle contractile genes . Mendelian disorders of skeletal and cardiac muscle function have historically underscored the essential role of the sarcomere and its contractile apparatus in human health and disease [43] , [44] . Since Huxley formulated the sliding filament model , all muscle contraction has been understood as a product of the interaction of two polymers , the thin filament actins and thick filament myosins [45] . Mendelian disorders largely conform to Huxley's fundamental insight as numerous disorders are now attributed to mutations in actins , myosins , and related proteins [46] , [47] . Sarcomere proteins had not been previously explored in MMIHS perhaps because the role of sarcomeric proteins in smooth muscle disease is less clear than in skeletal and cardiac disorders , and smooth muscle lacks the rigid alignment of the sarcomeres seen in cardiac and skeletal muscle . However , vascular smooth muscle disease has also been attributed to mutations in actin and myosin genes with the discovery of mutations in ACTA2 [48] and MYH11 [49] in thoracic aortic aneurysms and dissections . There are also reports of a specific mutation in ACTA2 associated with vascular aneurysms and hypomotility of the gastrointestinal tract ( OMIM 613834 ) [50] and also with prune belly sequence [51] . Additionally , as mentioned above , adult onset visceral myopathy was recently associated with a dominant mutation in the ACTG2 enteric actin gene in a Finnish family [20] . These findings provide the context for our data demonstrating a role for ACTG2 in MMIHS . While the mouse model for CHRNA3 generated promising insight into ganglion cell neurotransmission and smooth muscle function [27] , the smooth muscle itself is clearly also involved in MMIHS . These results strongly suggest that there are other genes that are mutated in MMIHS , and candidate genes can be envisioned based on the combined data from mouse and human mutations . MMIHS can be considered the most severe Mendelian enteric smooth muscle myopathy . After the submission of this manuscript , a report was published detailing the identification of de novo ACTG2 mutations by exome sequencing in two children with MMIHS [52] .
Informed consent was obtained prior to participation from all subjects or parents of recruited subjects under one of two Institutional Review Board approved protocols at Baylor College of Medicine . Whenever possible , our clinicians assessed study subjects by direct history , physical examination , and family history analysis . In some cases , clinicians referred subjects from centers around the world , and in those cases clinical information in the form of chart records and notes from the referring physicians were reviewed . Interviews with these subjects were also conducted by telephone . Families were asked prenatal history , and dates and nature of abdominal surgeries . Whenever available , reports from prenatal ultrasound , operative reports , manometry , or radiologic studies were reviewed . Methods utilized for whole-exome sequencing have been previously described in detail [53] . In summary , 1 µg of genomic DNA was fragmented by sonication in a Covaris plate ( Covaris , Inc . Woburn , MA ) . Genomic DNA samples were constructed into Illumina paired-end libraries as described [53] . Pre-capture libraries were pooled together and hybridized in solution to the BCM-HGSC CORE exome capture design [54] ( 52 Mb , NimbleGen ) . Captured DNA fragments were sequenced on an Illumina HiSeq 2000 platform producing 9–10 Gb per sample and achieving an average of 90% of the targeted exome bases covered to an minimal depth of 20× or greater . Produced sequence reads were mapped and aligned to the GRCh37 ( hg19 ) human genome reference assembly using the HGSC Mercury analysis pipeline ( http://www . tinyurl . com/HGSC-Mercury/ ) . Variants were determined and called using the Atlas2 [55] suite to produce a variant call file ( VCF ) [56] . High-quality variants were annotated using an in-house developed suite of annotation tools [57] . Primers were designed to encompass all the exons and intron-exon boundaries of the ACTG2 gene using ExonPrimer ( Tim Strom , http://ihg . gsf . de/ihg/ExonPrimer . html ) and Primer3 [58] . Sanger reads were analyzed using LASERGENE Seqman software [59] . Multiple sequence alignments were performed using Clustal Omega [60] and depicted using Boxshade . Arginine codon usage was determined using the Codon Usage Database and the countcodon program of Yazukazu Nakamura . | In 1976 , a radiologist , Walter Berdon described a group of patients with a rare intestinal and bladder disorder in which the smooth muscle of those organs failed to contract . These patients are unable to digest food , require multiple abdominal surgeries and are diagnosed with megacystis-microcolon-intestinal hypoperistalsis syndrome ( MMIHS ) . Since the description of MMIHS , the genes that cause it have remained a mystery . We followed and obtained DNA from patients with this disorder over a period of over 14 years and assembled a large group of cases . We used whole-exome sequencing , a powerful tool used to identify disease genes , and found mutations in ACTG2 , a visceral actin gene . Actins are components of muscle contractile units , and one Finnish family has been previously found with less severe gastrointestinal problems due to mutations in this gene . In our patients , we find de novo mutations in the majority of cases of MMIHS . However , we also find families with the disease over several generations due to these same mutations . This work provides the first disease gene for MMIHS , and suggests new treatment options . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"medicine",
"and",
"health",
"sciences",
"pediatric",
"gastroenterology",
"gastrointestinal",
"motility",
"disorders",
"gastroenterology",
"and",
"hepatology",
"genetics",
"biology",
"and",
"life",
"sciences",
"genomics",
"clinical",
"genetics"
] | 2014 | Heterozygous De Novo and Inherited Mutations in the Smooth Muscle Actin (ACTG2) Gene Underlie Megacystis-Microcolon-Intestinal Hypoperistalsis Syndrome |
Armigeres subalbatus is a natural vector of the filarial worm Brugia pahangi , but it kills Brugia malayi microfilariae by melanotic encapsulation . Because B . malayi and B . pahangi are morphologically and biologically similar , comparing Ar . subalbatus-B . pahangi susceptibility and Ar . subalbatus-B . malayi refractoriness could provide significant insight into recognition mechanisms required to mount an effective anti-filarial worm immune response in the mosquito , as well as provide considerable detail into the molecular components involved in vector competence . Previously , we assessed the transcriptional response of Ar . subalbatus to B . malayi , and now we report transcriptome profiling studies of Ar . subalbatus in relation to filarial worm infection to provide information on the molecular components involved in B . pahangi susceptibility . Utilizing microarrays , comparisons were made between mosquitoes exposed to B . pahangi , B . malayi , and uninfected bloodmeals . The time course chosen facilitated an examination of key events in the development of the parasite , beginning with the very start of filarial worm infection and spanning to well after parasites had developed to the infective stage in the mosquito . At 1 , 3 , 6 , 12 , 24 h post infection and 2–3 , 5–6 , 8–9 , and 13–14 days post challenge there were 31 , 75 , 113 , 76 , 54 , 5 , 3 , 13 , and 2 detectable transcripts , respectively , with significant differences in transcript abundance ( increase or decrease ) as a result of parasite development . Herein , we demonstrate that filarial worm susceptibility in a laboratory strain of the natural vector Ar . subalbatus involves many factors of both known and unknown function that most likely are associated with filarial worm penetration through the midgut , invasion into thoracic muscle cells , and maintenance of homeostasis in the hemolymph environment . The data show that there are distinct and separate transcriptional patterns associated with filarial worm susceptibility as compared to refractoriness , and that an infection response in Ar . subalbatus can differ significantly from that observed in Ae . aegypti , a common laboratory model .
Human lymphatic filariasis ( LF ) is caused by several species of mosquito-borne filarial nematodes , including Brugia malayi , Brugia timori , and Wuchereria bancrofti . Lymphatic filariasis is not a disease that causes mortality but it is by no means trivial . It is estimated that 120 million people in the world have LF , with ∼1 . 1 billion at risk of becoming infected . Although LF is rarely fatal , severe morbidity ( including adverse economic and psychosexual effects ) involves disfigurement of the limbs and male genitalia ( elephantiasis and hydrocele , respectively ) [1]–[3] . Mosquitoes belonging to a number of different genera , including Culex , Aedes , Mansonia , Anopheles , and Armigeres , can serve as competent vectors . Following ingestion of an infective bloodmeal , microfilariae ( mf ) penetrate the mosquito midgut epithelium and migrate to the thoracic musculature . Migration to the head region occurs after a series of larval molts to the infective third stage ( L3 ) , which occurs in thoracic muscle cells . The L3 is then passed to a vertebrate host when the infected mosquito takes a bloodmeal . In these mosquito-parasite systems , the coevolutionary history of parasite and vector in one geographic region can differ from vector-parasite relationships in another area [4] . The ability of vector mosquitoes to ingest mf of filarial worm parasites and to support their development after ingestion is an important determinant of filariasis transmission . For a plethora of reasons , often times the number of L3s developing from ingested mf is not constant [5] . Resistance to mf in the mosquito mainly involves melanotic encapsulation ( e . g . , Ar . subalbatus and B . malayi , Aedes trivitattus and Dirofilaria immitis , Anopheles quadrimaculatus and B . pahangi ) , but permissiveness of the midgut for parasite penetration , and physiological suitability of the thoracic muscle cells also may be important determinants of vector competence [6]–[9] . When melanization occurs , mf are rapidly melanized in the hemocoel once they have penetrated the midgut . As soon as 10 minutes following a bloodmeal , melanin deposition is evident on the mf cuticle . At 12–16 hours post feeding , melanization is well underway and pathological effects on the mf are evident . At 24 to 48 hours post feeding mf begin to die , and by 72 hours post feeding , the response is all but complete [7] , [8] . Armigeres subalbatus is a natural vector of the filarial worm Brugia pahangi , but it kills B . malayi mf by melanotic encapsulation [6] . Because B . malayi and B . pahangi are morphologically and biologically similar , this mosquito-parasite system serves as a valuable model for studying resistance mechanisms in mosquito vectors [10] . Previously , we assessed the transcriptional response of Ar . subalbatus to B . malayi , which revealed the possible involvement of a number of unknown and conserved unknown gene products , cytoskeletal and structural components , and stress and immune responsive factors in the mosquito's anti-filarial worm response . The data showed that the anti-filarial worm immune response by Ar . subalbatus is a highly complex , tissue-specific process involving varied effector responses working in concert with blood cell-mediated melanization [11] . Therefore , comparing Ar . subalbatus-B . pahangi susceptibility and Ar . subalbatus-B . malayi refractoriness could provide significant insight into recognition mechanisms required to mount an effective anti-filarial worm immune response in the mosquito , as well as provide considerable detail into the molecular components involved in vector competence . Accordingly , we initiated transcriptome profiling studies of Ar . subalbatus in relation to filarial worm infection to provide information on the molecular components involved in B . pahangi susceptibility for comparison with our earlier studies on B . malayi refractoriness [11] . In addition , these studies also provide information on the infection response of a natural vector , i . e . , the overall transcriptional and physiological change that occurs in the mosquito as a result of parasite infection , for comparison with our previous studies that employed a highly susceptible laboratory model , Aedes aegypti [12] . The time course chosen facilitated an examination of key events in the development of the parasite , beginning with the very start of filarial worm infection and spanning to well after infective-stage parasites had completed development in the mosquito . The data presented herein provide us with a cadre of information to design wet lab experiments and select candidates for further study to more fully dissect the nature of the anti-filarial worm immune response in this mosquito-parasite system . And with no genome sequence available for Ar . subalbatus , these data sets , in conjunction with data generated from our previous work ( see [11] ) , represent the most complete set of transcriptomic information available to date for this mosquito species , which should be of interest to numerous laboratories investigating vector biology and innate immunity in general .
Ar . subalbatus used in this study were maintained at the University of Wisconsin-Madison as previously described [7] . Ar . subalbatus supports the development of B . pahangi mf to L3 [6] , [13] . Three- to four-day-old female mosquitoes were used for bloodfeeding and sucrose starved for 14 to 16 hours prior to this event . All animals and animal facilities are under the control of the School of Veterinary Medicine with oversight from the University of Wisconsin Research Animal Resource Center , and their use in experimentation was approved by the University of Wisconsin Animal Care and Use Committee . Mosquitoes were exposed to B . pahangi and B . malayi ( originally obtained from the University of Georgia NIH/NIAD Filariasis Research Reagent Repository Center ) by feeding on ketamine/xylazine anesthetized gerbils , Meriones unguiculatus . The same animals were used for all three biological replicates . Microfilaremias were determined , using blood from orbital punctures , immediately before each feeding and ranged from 15–60 mf/20 µl blood . Control mosquitoes were exposed to anesthetized , uninfected gerbils . Mosquitoes that fed to repletion were separated into cartons and maintained on 0 . 3 M sucrose in an environmental chamber at 26 . 5°±1°C , 75±10% RH , and with a 16 hr photoperiod with a 90 minute crepuscular period . Nine sample groups were created from thirteen timepoints to study mosquito transcriptome changes in response to B . pahangi development . In each sample group , comparisons were made between mosquitoes exposed to an infective bloodmeal containing B . pahangi mf and those exposed to a bloodmeal without parasites . In addition , a separate set of microarray analyses directly compared transcriptome profiles between mosquitoes in which filarial worms develop to infective stage larvae ( B . pahangi ) and mosquitoes in which an anti-filarial worm immune response had been initiated ( B . malayi ) . This direct comparison contained groups of mosquitoes exposed to an infective bloodmeal containing B . pahangi mf or exposed to a bloodmeal containing B . malayi mf . Sample collection followed the same time course chosen for the DNA microarray experiments investigating B . malayi refractoriness previously published [11] , but included three new biological replicates of mosquitoes exposed to B . malayi , done concurrently with the B . pahangi exposures . This direct comparison was advantageous because indirect comparisons of separate experimental data sets that employ a control that is assumed to be the same treatment for each could yield false information . For example , indirect comparisons made between DNA microarrays that compared B . pahangi vs . blood in this study compared to B . malayi vs . blood in [11] could be problematic , because the variation in the separate control groups are reflected in these comparisons , and because of the variation inherent in indirect comparisons [14] . The sample groups were defined by the time post ingestion ( PI ) of the bloodmeal and represent significantly different stages of parasite development . Twenty mosquitoes , exposed to either B . pahangi or uninfected blood meals , were collected at 1 , 3 , 6 , 12 , and 24 h PI and ten mosquitoes at 2 , 3 , 5 , 6 , 8 , 9 , 13 and 14 d PI . Twenty mosquitoes , exposed to a bloodmeal containing B . malayi mf , were collected at 1 , 6 , 12 , and 24 h PI and ten mosquitoes at 2 and 3 d PI . Mosquitoes were pooled ( 2–5 mosquitoes/tube ) , RNA was immediately extracted , and then stored at −80°C until cDNA synthesis . At each time point , an additional 5 mosquitoes were dissected , and the head , thorax , midgut and abdomen were examined microscopically to verify filarial worm infection and to determine the stage of parasite development . Briefly , at 1 , 3 , 6 , 12 and 24 h PI mf penetrate the mosquito midgut , migrate through the hemocoel , and penetrate thoracic muscle cells . Group 6 was collected at 2–3 d PI , a time when mf differentiate into intracellular first-stage larvae ( L1 ) . At 5–6 d PI , B . pahangi complete the molt to second-stage larvae ( L2 ) and actively feed on mosquito muscle tissue ( Group 7 ) . In Group 8 , collected at 8–9 d PI , parasite development is complete with a second molt to the L3 , which breaks out of the thoracic muscles and migrates to the head and proboscis . The final sample collection ( Group 9 ) was made at 13–14 d PI , when the majority of parasites have completed their migration to the head and proboscis [12] , [13] , [15] . Please note that the terminology used to define the components of the DNA microarray are derived from the original DNA microarray paper [16]; therefore , the target is that which is tethered to the DNA microarray substrate and the probe is the labeled material in solution that hybridizes to the target . Microarrays used in this study were designed as previously described [11] , [17] and printed at the University of Wisconsin Gene Expression Center using a Genomic Solutions GeneMachines Omnigrid arrayer and SMP3 pins from TeleChem International following established printing protocols . RNA was collected for all microarray analyses from whole female Ar . subalbatus and processed as described previously [11] . Briefly , three biological replicates , each with two technical replicates , done as dye swapped pairs ( Cy5 experimental vs . Cy3 control ) , were performed for each experimental set in an effort to eliminate dye bias [18] , [19] . Each time point for each biological replicate consisted of twenty pooled mosquitoes , and each biological replicate consisted of mosquitoes from distinct generations to take into account stochastic variations . RNA integrity was verified via gel electrophoresis or via Bioanalyzer ( Agilent , Santa Clara , CA ) and only quality intact RNA was used for cDNA synthesis . cDNA synthesis was done according to the Chipshot™ Indirect Labeling System with modification ( use of an anchored oligo ( dT ) primer ) ( TM261 , Promega ) . Priming with anchored oligo ( dT ) directed the start of synthesis from the 5′ end of the poly-A tail . Twenty µg of total RNA were used as a template for the synthesis of amino allyl-modified cDNA . Purified cDNA from each synthesis reaction was coupled to Cy3 or Cy5 according to manufacturers' instructions . The CyDye ( GE Healthcare ) probes were purified using the ChipShot™ Membrane Clean-Up system ( TM261 , Promega ) . Purified , dye-coupled cDNA was measured at 260 nm ( Cy5 @ 650 nm and Cy3 @ 550 nm ) to calculate yield . Probes ( 10 pmol/dye/slide ) were dried down using a speedvac , resuspended at room temperature in 45 µl Pronto ! ™ hybridization buffer , incubated at 95°C for 5 minutes , and applied to the DNA microarrays . DNA microarrays were hybridized overnight at 42°C . Hybridized DNA microarrays were processed using the Pronto ! microarray hybridization kit ( Corning ) according to manufacturer's specifications . The microarray data were prepared according to “minimum information about a microarray experiment” ( MIAME ) recommendations , deposited in the Gene Expression Omnibus ( GEO ) database , and can be accessed via the web ( accession number GSE20205 ) . All transcript and EST data for this project are publicly accessible in ASAP ( A Systematic Annotation Package for community analysis of genomes ) [20] as the complete collapsed set ( ARALL v2 ) or through NCBI's GenBank database ( accession numbers EU204979-EU212998 ) via the web . Microarray scanning and analysis was conducted as described previously [11] . Briefly , signal intensities were normalized using GeneSpring GX 7 . 3 . 1 software . All slides were normalized using a global linear regression ( Lowess ) curve fit to the log-intensity vs . log-ratio plot , and 20% of the data were used to calculate the Lowess fit at each point . All data were averaged for replicate spots upon a slide , and then further averaged across slides . Minimum and maximum values were recorded and t-test p-values generated for all replicate sets . Genes showing differential expression over controls were isolated using volcano plots at a 95% confidence interval over 2-fold values . Tests were parametric , but all variances were considered equal . Transcript levels of seven selected genes were measured using SYBR dye technology and quantitative polymerase chain reaction ( qPCR ) to validate microarray results as previously described [11] . Group 1 included three transcripts at 1 hour post challenge: a glycine-rich secreted salivary gland protein ( Genbank:EU207085 ) and two unknowns ( Genbank:EU211627 ) and ( Genbank:EU209094 ) . Group 2 included one transcript at 3 hours post challenge: a potassium: amino acid symporter ( Genbank:EU210583 ) . Group 3 included four transcripts at 6 hours post challenge: a mucin-like peritrophin ( Genbank:EU206650 ) , a glycine-rich secreted salivary gland protein ( Genbank:EU207085 ) , a potassium: amino acid symporter ( Genbank:EU210583 ) , and a serine protease ( Genbank:EU205658 ) . Finally , Group 4 included one transcript at 12 hours post challenge: a chitinase ( Genbank:EU205713 ) . All primer sequences used in microarray validation are presented in Table S1 . Thoraces from mosquitoes exposed to B . pahangi-infected and uninfected bloodmeals were separated from whole bodies at 6 , 9 , and 14 days after blood feeding by making transverse cuts along the cervical membrane and the first abdominal segment . The legs and thoracic ventral cuticle were partially removed by making a coronal cut along the mesosternum and tissues were fixed by immersion in 4% formaldehyde in 0 . 1 M phosphate buffer ( pH 7 . 0 ) . Thoraces then were prepared for light microscopy as described [21] . Briefly , aldehyde-fixed thoraces were dehydrated through a graded ethanol series , infiltrated in JB4-Plus resin ( Electron Microscopy Sciences , Hatfield , PA ) , and anaerobically embedded in polyethylene molding trays . Coronal and transverse sections of 2 . 5 µm thickness were cut , stained with Gill's hematoxylin and Eosin Y , and mounted on glass slides using Poly-Mount ( Polysciences Inc . , Warrington , PA ) . Tissues were then imaged using differential interference contrast ( DIC ) optics on a Nikon Eclipse 90i compound microscope connected to a Nikon DS-Fi1 high-definition color CCD camera ( Nikon Corp . , Tokyo , Japan ) .
The development of B . pahangi was observed microscopically in Ar . subalbatus at 1 h to 14 d ( a total of 11 time points ) post ingestion ( PI ) of a microfilaremic bloodmeal for the first biological replicate , and at 1 h and 14 d PI for the subsequent two biological replicates . The results are summarized in Table 1 . Parasites were recovered from 70 of the 75 mosquitoes examined for an overall infection prevalence of 93 . 3% . Microfilariae were recovered from 1 h to 24 h PI , and constituted the majority of the total parasites through this time period . From 24 h to 3 d PI , almost all parasites had differentiated into intracellular L1s . L1s molted to L2s in the thoracic musculature by 5 d PI , and L2s were the only developmental stage identified between 5 and 6 d PI . The transformation from L2 to L3 occurred at 8–9 d PI . At 8 d PI , L2s and L3s were recovered from the thoracic musculature with only 2% of the total worms located in the head and proboscis . By 9 d PI all worms had molted to the L3 stage . At this time 43% of the L3s were observed in the thorax and 57% were located in the head and proboscis . By 13–14 d PI , all L3s were observed in the head and proboscis . The prevalence of L3s ( for all three biological replicates on 14 d PI ) was 87% ( n = 15 ) and the mean intensity was 11 . 3±9 . 6 L3s . Direct comparisons were made at each time point within DNA microarrays hybridized with probes made from RNA of whole female Ar . subalbatus exposed to either a B . pahangi-infected or uninfected bloodmeal . Volcano plots were used to create working gene lists to identify differentially expressed genes at each time point . At 1 , 3 , 6 , 12 , 24 h PI and 2–3 , 5–6 , 8–9 , and 13–14 d PI there were 31 , 75 , 103 , 76 , 54 , 5 , 3 , 13 , and 2 detectable transcripts of the 6 , 143 features on the DNA microarray , respectively , with significantly different transcriptional patterns ( increased or decreased transcript abundance at a 95% confidence interval over two-fold values ) as a direct or indirect result of parasite development ( Figure 1 ) . Between each time point there was very little overlap ( i . e . , less than 27 transcripts shared between any two timepoints ) in the transcripts that showed significantly different transcriptional patterns ( Table 2 ) . The vast majority of changes in the mosquito transcriptome occurred within the first 24 h of infection , when mf were penetrating through the midgut and invading thoracic muscle cells . In contrast , relatively minor changes in the mosquito's transcriptome were noted at later times during parasite development . This is somewhat surprising , considering that at 5–6 d PI parasites were actively feeding on mosquito tissue , and by 13–14 d PI parasites had grown considerably in size and migrated to the mosquito's head and proboscis . Of the transcripts that showed significantly different transcriptional patterns over this experiment , only 10% had putative immune functions ( Table 3 ) . In addition , of the 364 transcripts that showed significantly different transcriptional patterns over the time course of the B . pahangi vs . blood experiment , 193 ( 53% ) of those transcripts were unknowns or conserved unknowns , i . e . , they have no previously described function ( Figure 2 ) . This suggests that the function of many factors involved in the infection response to developing filarial worms ( i . e . , the factors involved in helping the mosquito maintain homeostasis during a persistent infection ) are not known . Table S2 provides a full representation of all transcripts showing a detectable increase or decrease in abundance at all time points . Direct comparisons were made at each time point within DNA microarrays hybridized with probes made from RNA of whole female Ar . subalbatus exposed to either a B . pahangi- or B . malayi-infective bloodmeal . Volcano plots were used to create working gene lists to identify genes with a significant fold difference to begin to better understand the transcriptional profiles associated with exposure to each species of parasite . At 1 , 6 , 12 , 24 h PI and 2–3 d post challenge there were 10 , 14 , 15 , 27 , and 4 detectable transcripts , respectively , more associated with B . pahangi infection relative to B . malayi resistance . Following the same time course , there were 63 , 20 , 57 , 81 , and 6 detectable transcripts , respectively , more associated with B . malayi resistance relative to B . pahangi infection ( Figure 3 ) . The majority of changes that occurred in the mosquito transcriptome were associated with B . malayi resistance ( 76% ) ; however , there may be some bias in these data because there was no B . pahangi-exposed mosquito cDNA included in the EST libraries used to create our Armigeres microarray [17] . Of the transcripts that showed significantly different transcriptional patterns as a result of B . pahangi infection , only 14% had putative immune functions ( Table 4 ) and 49% were unknowns or conserved unknowns . Of the transcripts that showed significantly different transcriptional patterns as a result of B . malayi resistance , only 12% had putative immune functions ( Table 4 ) and 67% were unknowns or conserved unknowns . These results suggest that although the biosynthetic pathway of melanization is well understood , many factors involved in the anti-filarial worm immune response are not known . Filarial worm susceptibility in Ar . subalbatus is much more complicated than merely an absence of the melanization immune response ( Figure 4 ) ; likewise , filarial worm resistance in Ar . subalbatus is much more complicated than the presence of melanization and likely includes factors involved in recovery and maintenance of homeostasis . For example , there were few transcripts involved in the metabolism of reactive intermediates that showed significantly different transcriptional patterns as a result of B . pahangi infection vs . uninfected blood ( Table 3 ) . This is in contrast to what was observed in Ar . subalbatus in response to B . malayi [11] , and probably has to do with the fact that mosquitoes infected with B . pahangi do not have to protect themselves from the damaging effects of the oxidative stressors produced during melanization reactions [22] . Table S3 provides a full representation of all transcripts showing a detectable difference in abundance between B . pahangi- and B . malayi-infected mosquitoes at all time points . Microarray data were confirmed using both in silico analyses of known transcriptional information in the literature and laboratory-based analyses via qPCR [18] , [23] , [24] . The transcriptional activity of a number of different response genes , induced by bacteria or filarial worms , has been characterized previously in several mosquito species . Based on this information , and the fact that the RNA used to screen the DNA microarrays in these studies result from the exposure to filarial worms , it was expected that a number of parasite responsive genes on the DNA microarrays would show significant transcriptional patterns [11] , [25]–[30] . Comparisons made between our data and parasite responsive genes in the literature corroborated our findings and provided validation of our DNA microarray results . In addition , a number of “house-keeping” genes ( ex . Ribosomal genes , actin , cytochrome C oxidase , etc . ) included on the DNA microarray showed no detectable change in transcript abundance throughout experimentation ( data not shown ) , thereby providing further validation of the expression patterns detected . In conjunction with in silico validation of DNA microarray results , qPCR provided independent , experimental verification of transcript abundance from the same total RNA used in the initial DNA microarray experiment . Because the corroboration of all microarray data was impractical , a subset of seven genes was chosen at random from our lists of significant genes for confirmatory studies . Transcriptional activity of three selected genes was verified at 1 h , one selected gene at 3 h , four selected genes at 6 h , and one selected gene at 12 h post challenge , and eight of the nine conditions tested corroborated with transcriptional patterns detected on the DNA microarray ( Text S1 ) . The elimination of one gene from our dataset may reflect the differential sensitivities of the techniques used or sample variation . This still validated that the DNA microarray was working as expected , showing all three conditions ( increase , decrease , and no detectable change in transcript abundance ) . These results , in combination with in silico and laboratory-based validation , provided confidence that the transcriptional profiles are an accurate depiction of the biological phenomena under study . Histological analyses of thoraces from infected mosquitoes confirmed the infection timeline observed using whole body dissections . At 6 d PI worms were developing in the thoracic musculature and , with rare exceptions , were oriented parallel to the muscle fibrils . At 9 d PI the worms were in the process of exiting the muscle fibers and migrating to the head , where they generally remain for the lifetime of the mosquito ( Figure 5 ) . Because the microarray data showed remarkably few mosquito transcriptional changes associated with parasite maturation and migration in the hemocoel , we undertook a comparative histological analysis of the thoracic musculature from B . pahangi-infected and uninfected mosquitoes , specifically surveying for obvious morphological changes associated with worm infection . As indicators of tissue damage , particular attention was paid to signs of muscle fiber degradation , the structure of myofibrils that were in direct contact with the worms , the morphology of host muscle nuclei , and signs of tissue infiltration by fat body . To assess changes at different points of the filarial nematode's life cycle , these analyses were performed at 6 , 9 , and 14 d post exposure to infective and normal bloodmeals , because these represent times when the worms are actively feeding , migrating , and infective to the vertebrate host , respectively . Overall , examination of hematoxylin and eosin-stained semi-thin sections did not reveal any obvious tissue damage associated with B . pahangi infection ( Figure 6 ) . At 6 d PI , greater than 10 worms were commonly observed in individual sections , indicating that infection intensities were high . Worms developed parallel to the myofibrils with little separation between the nematode cuticle and the mosquito myofibrils . As compared to non-infective bloodmeal controls , no obvious morphological change was observed in the infected muscle fibers , with the exception of what appeared to be occasional host-derived tissue pooling between the worm and the myofibrils . This tissue pooling , consisting of eosinophilic granules that we hypothesize are mitochondria , is likely the result of active feeding by the worm but surprisingly did not damage the integrity of the myofibers: sarcomere patterns were morphologically similar to those in uninfected mosquitoes . At 9 d PI , worms were observed migrating toward the head by laterally crossing the myofibers before exiting into the periphery of the thorax . Besides the physical breaks caused by worm migration , little additional damage was visually detected along the remaining portions of the invaded myofibers and the sarcomeres flanking these breaks displayed normal morphological patterns relative to adjacent myofibers and non-infected blood meal controls . By 14 d PI worms could only be detected near the head and were not intracellularly associated with the thoracic musculature . From the examination of the thoracic musculature we could not conclusively determine the precise location of earlier worm development , indicating that parasites do not leave behind obvious voids near their developmental areas and that much of the displacement caused by their presence is either repaired or filled by adjacent muscle fibers or fibrils . At all timepoints assayed , both infected and uninfected mosquitoes contained small numbers of myofibers that were atrophied or in the process of degradation . However , at none of the timepoints assayed did we detect higher numbers of abnormal myofibers in infected as compared to uninfected mosquitoes , suggesting that this fiber atrophy is either related to normal biological turnover or an artifact of the tissue manipulation process ( embedding , sectioning , etc . ) . Furthermore , fat body was observed near the periphery of the thorax and at times extended between muscle fibers . This also appeared to be a normal biological occurrence as a similar pattern of fat body distribution was observed in uninfected mosquitoes . Lastly , comparison of myofiber nuclear morphology in infected and uninfected mosquitoes revealed no obvious differences in nuclear size or chromatin condensation , further suggesting that infection does not result in significant muscle atrophy .
Interactions between mosquitoes and filarial worms can range from an almost benign symbiotic relationship between organisms , to a fatal competition resulting in the death of the host , or to a fatal competition resulting in the death of the parasite . In Ar . subalbatus both tolerance and resistance strategies can occur depending on the species of filarial worm infecting the mosquito , but high mosquito mortality , caused by ingestion of too many mf , can occur with either scenario . It is important to distinguish between tolerance and resistance because their relative importance will have substantial consequences for the ecology and evolution of host-parasite interactions [31] . Traditionally , most studies have examined the anti-filarial worm immune response in Ar . subalbatus ( e . g . , [11] , [25] , [29] ) , or explored innate immune responses to other pathogens in other mosquito species ( e . g . , [4] , [32] , [33] ) , but few have examined pathogen susceptibility in a natural mosquito vector ( in this case , a laboratory strain ) . Studies like this one , aimed at understanding the molecular basis of parasite-host interactions in a compatible system , will help identify the components involved in host resistance vs . parasite tolerance [4] . It is important to note that resistance and tolerance can be mutually exclusive , interchangeable , or complementary components of a mixed strategy of defense [34] , and a particular gene may be involved in both tolerance and resistance , depending on the pathogens involved [35] . Furthermore , tolerance may involve immunological mechanisms directed at damage or other harmful substances resulting from infection with the parasite , or may even reflect the parasite's ability to persistently evade the host's defenses to remain inside the host to achieve eventual transmission [36] , which will complicate the elucidation of the factors determining filarial worm susceptibility in this species of mosquito . For example , a number of transcripts implicated in innate immunity showed significantly different transcriptional behavior as a result of B . pahangi infection vs . uninfected blood ( e . g . , serine proteases , pattern recognition molecules , etc ) in the current study , and this may be an example of these transcripts being directed at damage or other harmful substances resulting from infection with the parasite . Apoptosis also could explain the activity of a number of the immune responsive transcripts present in this study , because cell death in vertebrates has been shown to trigger both innate and adaptive immune responses [37] , [38] . It also could explain the activity of those transcripts implicated in phagocytosis , because phagocytosis could be functioning to clean-up apoptotic cells [37]–[39] that were destroyed by mf penetrating the midgut . This is consistent with previous reports of apoptosis in the midgut of parasite and/or virus infected mosquitoes [40]–[46] , and it has been shown previously that basal and apical plasma membranes are destroyed by mf penetrating the midgut , likely resulting in cell death [47] . A number of antimicrobial peptides ( AMPs ) showed significantly different transcriptional behavior in the current study as well . Anti-microbial peptides are small , immune-related molecules named for their in vitro activity against bacteria and are detectable in the fat body , hemocytes , midgut , and epithelial tissues of mosquitoes . Although considered a primary defense mechanism against bacteria in mosquitoes , AMP transcription has been associated with responses to B . malayi infection in Armigeres and Aedes , Plasmodium infection in Anopheles , and fungal infection in other mosquito species [11] , [32] , [48] , [49] . Despite these associations , our knowledge of the molecular mechanisms and the true role of these peptides in mosquito innate immunity remain limited , and perhaps , the anti-microbial activity of AMPs might be an ancillary property . Recently , it has been shown that antimicrobial compounds function primarily to protect insects against bacteria that persist within the body , rather than to clear the infection . It has been proposed that AMPs may act as response elements helping insects deal with infection when pathogens in the hemolymph exceed the phagocytic or melanotic capacity of the hemocytes [50] , [51] . The results of the current study are consistent with this hypothesis , i . e . , the fact that AMP induction is evident as a result of B . pahangi infection in addition to B . malayi resistance seems to support the hypothesis that AMPs may play an alternative role , perhaps helping to maintain homeostasis of the hemolymph environment during a persistent infection or to clean up after a successful immune response , but are transcribed regardless of the type of pathogen present . In the vertebrate host , filarial nematodes apply successful strategies to evade the host's immune response [52] , [53] . Their strategy has evolved to be assimilation , defusing aggressive immune reactions , and inducing forms of immunological tolerance to permit their long-term survival [53] . In the Armigeres-Brugia system , B . pahangi may employ similar strategies; whereas , B . malayi is not successful in evading the mosquito's immune response . The specificity of resistance in this vector-parasite relationship warrants further exploration , and we postulate it to occur at the level of recognition , possibly involving well defined motifs ( e . g . , β 1 , 3-glucan [27] ) recognized by receptors that could be the products of resistance genes . Immune defenses ( e . g . , melanization ) would then be triggered only if recognition of the parasite occurs [4] . Therefore , the general hypothesis is that Ar . subalbatus recognizes a surface component ( s ) of B . malayi mf but not B . pahangi mf , i . e . , there is a fundamental biologic difference in the surface components of the two filarial worm species . Consistent with this hypothesis is the fact that a number of transcripts associated with targeting/initiating an immune response changed as a result of B . malayi infection relative to B . pahangi infection: e . g . , CD-36-protein ( GenBank: EU208970 ) , scavenger receptor ( GenBank: EU208273 ) , toll-like receptor 7 ( GenBank: EU208239 ) , 2 C-type lectins ( GenBank: EU206257 and EU208542 ) , and a mannose-binding lectin associated serine protease ( GenBank: EU206540 ) . If vector-parasite specificity is not achieved at the recognition level , we postulate that specificity could be explained by direct or indirect interactions between the products of immune suppressive genes of the parasite and the products of host resistance genes [4] , i . e . , B . pahangi mf can actively suppress the anti-filarial worm response in Ar . subalbatus but B . malayi mf cannot . Consistent with this hypothesis is the fact that the majority of changes that occurred in the mosquito's transcriptome were associated with B . malayi resistance ( 76% ) relative to B . pahangi susceptibility . In addition , indirect comparisons between B . malayi resistance [11] and B . pahangi susceptibility showed considerably more changes associated with resistance ( 761 vs . 346 ) over the course of 72 h . Recently , we examined the infection response of Ae . aegypti to B . malayi at the transcriptional level [12] . In this experimental model system , essentially all ingested parasites successfully develop to L3s . This study revealed very few changes in the Ae . aegypti transcriptome until L2s were present , and the most profound transcriptional changes were observed in mosquitoes that harbored infective-stage parasites . In comparison , the vast majority of changes in the Ar . subalbatus transcriptome as a result of B . pahangi infection were observed between 1 and 24 hours PI when mf were penetrating the midgut and invading thoracic muscle cells , and very few transcriptional changes were observed in mosquitoes that harbored L2s or infective-stage parasites . It is important to note that the array platforms used to conduct these two studies were different ( Ar . subalbatus-B . pahangi = EST-based DNA microarray; Ae . aegypti-B . malayi = whole genome DNA microarray ) because there is no genome sequence available for Ar . subalbatus , and some of the differences observed between the two studies may be the result of the inherent differences in these two approaches . But the different transcriptional changes observed in these two mosquito species that support parasite development also can provide clues to the molecular mechanisms that determine compatibility in natural mosquito-filarial worm associations . By comparing the two infection responses we can begin to better understand the intricacies involved in susceptibility vs . refractoriness , i . e . , by identifying transcripts that are shared between both compatible mosquito-parasite systems and from the refractory condition , we can begin to rule out their involvement in anti-filarial worm immunity . Such comparisons have been made previously between other mosquito genera ( e . g . , [54] ) , but this study is the first to make a comparison between the susceptible vs . refractory state in the same species of mosquito , and may provide a better representation of the genes required to deter filarial worm infection in an incompatible system . These results also illustrate the fact that not all mosquitoes respond the same way to filarial worm infection and not all filarial worm species will elicit the same response in a host . Host-parasite interactions represent coevolved adaptations of significant complexity , and these relationships depend on the relative capacities of the host and pathogen to adapt to and maintain this unique relationship . Furthermore , earlier studies assessing B . malayi and B . pahangi infection of Ae . aegypti and Mansonia uniformis found that nematode migration and development causes minor and severe damage to the thoracic indirect flight muscles [55] , [56] . Later , it was proposed that this damage was not pathogen specific because similar damage was observed in mosquitoes physically traumatized by intrathoracic insertion of a metal probe [57] . In contrast to those studies , the experiments we report here failed to detect significant pathology in the thoracic musculature of Ar . subalbatus infected with B . pahangi , with the exception of breaks in myofibers that appear to be the direct result of nematode exit from the indirect flight muscles . This is consistent with other studies assessing B . pahangi infection in mosquitoes , where infection of the natural vector Aedes togoi resulted in undetectable flight muscle damage ( suggesting complete myofiber repair following migration to the head ) , whereas infection of the artificial vector Ae . aegypti resulted in severe degeneration of the flight muscles [58] , [59] . The histological and transcriptomic data reported in this study further verify the need to work with natural mosquito-parasite systems , because it is evident that extrapolating what is learned from a laboratory model of a parasite-vector relationship to the natural model can be problematic . Aside from myofiber breaks , the primary difference between the thoracic musculature from infected versus normal bloodfed Ar . subalbatus was the accumulation of eosinophilic granules between the nematode cuticle and the host myofibers . We presume that these granules are host mitochondria , and if so , these data would be in accord with electron microscopic data showing similar accumulations following B . pahangi infection of Ae . aegypti [60] . In that study , pooled mitochondria showed no evidence of damage and it is not clear if any deleterious effect is associated with their accumulation . However , because similar structures were observed in the nematode gut [60] , it is possible that filarial nematodes subsist by ingesting mitochondria . Our data and that of others [60] have provided no evidence that filarial nematodes ingest the contractile components of the thoracic musculature . Although our data did not reveal any obvious pathological consequence to infection , pathology may still exist . The techniques used in this study only allowed for the examination of structural damage at the light microscopic level and did not molecularly assess cell death , or explore pathology at the ultrastructural level . Significant mortality has been associated with filarial nematode invasion of and exit from the indirect flight muscles of Ar . subalbatus ( Aliota et al . , unpublished; and [61] ) and it is possible that this mortality is a consequence of damage that impairs flight , resulting in decreased feeding and other essential biological processes . Finally , it is important to consider that host and parasite genotypes share control of epidemiological parameters of their relationship . Most models of the evolutionary processes in host–parasite systems assume that the evolution of attack or defense strategies is governed by the balance of their evolutionary costs and benefits from the point of view of either the parasite or the host and , thus , hold the other partner constant . In other words , they consider that the traits of the relationship are determined by the genotype either of the host or of the parasite , but not an interaction between the two [62] . Future studies that explore the Interactome- the whole set of molecular interactions of both organisms in a symbiotic relationship- of a host-pathogen system should be extremely valuable in determining the evolutionary basis for tolerance vs . resistance and help to elucidate the underlying components of vector competence . | In general , organisms can use two different strategies when confronted with pathogens , tolerance and/or resistance . Resistance reduces the fitness of the invading pathogen , whereas tolerance reduces the damage caused by the pathogen to the host . Mosquitoes that transmit the parasites that cause human lymphatic filariasis generally are tolerant to the parasite , whereas those that do not transmit the parasite are resistant . We examined the effects of filarial worm tolerance and resistance on Armigeres subalbatus by analyzing changes in mosquito gene expression at key stages of parasite development and destruction . Because the gene expression data showed few mosquito transcriptional changes associated with parasite development , we morphologically examined mosquito flight muscle to see if we could identify damage associated with parasite infection . The research described in this manuscript provides a better understanding of the molecular components involved in compatible and incompatible relationships between mosquitoes and the filarial worm parasites that they transmit; and it will provide new insights into the complex biology of vector competence and the origins of host defense , and possibly lead to the functional characterization of previously unknown gene products involved in vector competence . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"genetics",
"and",
"genomics/functional",
"genomics",
"pathology/histopathology",
"immunology/innate",
"immunity",
"molecular",
"biology/bioinformatics",
"cell",
"biology/gene",
"expression"
] | 2010 | Mosquito Transcriptome Profiles and Filarial Worm Susceptibility in Armigeres subalbatus |
The metabolic capabilities and regulatory networks of bacteria have been optimized by evolution in response to selective pressures present in each species' native ecological niche . In a new environment , however , the same bacteria may grow poorly due to regulatory constraints or biochemical deficiencies . Adaptation to such conditions can proceed through the acquisition of new cellular functionality due to gain of function mutations or via modulation of cellular networks . Using selection experiments on transposon-mutagenized libraries of bacteria , we illustrate that even under conditions of extreme nutrient limitation , substantial adaptation can be achieved solely through loss of function mutations , which rewire the metabolism of the cell without gain of enzymatic or sensory function . A systematic analysis of similar experiments under more than 100 conditions reveals that adaptive loss of function mutations exist for many environmental challenges . Drawing on a wealth of examples from published articles , we detail the range of mechanisms through which loss-of-function mutations can generate such beneficial regulatory changes , without the need for rare , specific mutations to fine-tune enzymatic activities or network connections . The high rate at which loss-of-function mutations occur suggests that null mutations play an underappreciated role in the early stages of adaption of bacterial populations to new environments .
Bacteria evolve to exploit the temporal and spatial structure of their native environments , mapping commonly occurring patterns of stimuli to high-fitness responses [1] , [2] . Adaptation occurs through both the acquisition of requisite biochemical and biophysical functions , such as enzymatic capabilities and membrane properties , and evolution of a regulatory network that responds to the environment by deploying the organism's phenotypic capacities in a context-appropriate fashion . In principle , bacteria may grow poorly in a new environment because they lack necessary biochemical capabilities and biophysical properties , or because they express these existing capacities inappropriately . In the former case , mutations that tinker with coding regions to refine existing functions [3]–[5] , horizontal gene transfers that introduce novel functions [6] , and gene amplifications that enable subsequent neofunctionalization [7] could generate the missing functionality . In the latter case , a bacterium's genome encodes the requisite biochemical and biophysical functions , but the organism's sensory and regulatory networks do not express the functions in a context-appropriate fashion ( Fig . 1A ) . While rare mutations that modulate specific network connections can engender the appropriate regulatory capacity ( for example , the hijacking of an aerobic promoter to enable aerobic citrate metabolism in Escherichia coli during a long term evolution experiment [8] ) , comparatively common loss-of-function ( null ) mutations [9] that produce less specific perturbations could also generate advantageous network adjustments . The maladaptive properties of null mutations , including their contributions to genome decay are well known [10] . Unlike the rare , specific changes associated with gain-of-function mutations , however , the loss-of-function mutational space can be explored rapidly by an evolving population due to the large number and variety of sequence-level mutations that can give rise to such changes . Although adaptive null mutations have been observed in bacterial laboratory evolution experiments ( e . g . , [11] , [12] ) , the general potential for null mutations to shape the path of bacterial evolution has not been systematically investigated , despite their potential to enhance fitness by re-deploying the existing capabilities of cells ( Fig . 1B ) . In the discussion below , we refer to any effect in which a single mutation alters cellular fitness by causing non-local changes in information flow or metabolite flux as ‘rewiring’; by definition , any beneficial null mutation which does not exert its impact by removing an actively deleterious reaction must be acting through rewiring . As we show below , the reconfiguration of cellular metabolism triggered by even one or two such changes often yields improvements in fitness . Rather than provide the cells with qualitatively new capabilities , these mutations improve the cells' application of existing metabolic capabilities to the selective conditions that they are experiencing . While a series of null mutations is unlikely to yield optimal deployment of a cell's constituent genes under novel conditions , loss-of-function mutations can allow the survival and growth of partially adapted individuals that might then further evolve and adapt to the new surroundings ( Fig . 1C ) . Null mutations can also provide access to alternate evolutionary trajectories via different epistatic interactions [13] , further expanding the range of phenotypes accessible to an evolving population . Over the past several decades , numerous detailed studies of specific individual mutants , as well as high-throughput studies of deletion libraries in both bacteria [14] and yeast [15] , have identified diverse examples of null mutations that provide a fitness advantage under a wide range of natural and artificial conditions ( specific examples in bacteria are listed in Table S1 ) . Such beneficial loss-of-function mutations can have varied functional consequences ( summarized below in the context of the highly schematized cellular network depicted in Figure 2 ) . The most obvious mechanism for a beneficial null mutation is to remove a protein or enzyme directly detrimental in the environment of interest ( S or E3 in Fig . 2 ) . For example , deletion of ompF reduces tetracycline entry into the cell , increasing tetracycline tolerance [16] . Similarly , deletion of the peptidoglycan-recycling enzyme slt enhances ethanol tolerance by altering cell wall structure [17] . Many gene products whose deletion is beneficial , however , act multiple steps away from the key cellular property that their deletion modulates . For example , deletion of an enzyme ( E2 in Fig . 2 ) or an upstream regulator ( R1 in Fig . 2 ) may modify metabolic flux to better fit the studied environment . This is illustrated by the combined deletion of fnr , arcA , and cafA , which enhances ethanol tolerance in E . coli by increasing ethanol breakdown and subsequent assimilation [17] . Similarly , removal of proteins involved in catabolism or oxidative respiration increases resistance to bactericidal antibiotics by ultimately reducing the production of harmful hydroxyl radicals [18]–[20] . A cellular network's underlying modularity often enables a single regulatory deletion ( R2 or R3 in Fig . 2 ) to alter the levels of multiple components coherently . For example , mutations in many signaling pathways feeding into flhDC , the master regulator of flagellar biogenesis in E . coli , can modulate flagella-based cellular motility , such as deletion of ompR or envZ enhancing motility in high-salt conditions [21] . Similarly , the high connectivity of housekeeping genes ( H in Fig . 2 ) in the cellular network can allow their removal to trigger beneficial phenotypes under diverse environments , such as the deletion of Lon protease conferring an advantage in the presence of A22 , β-lactams , and ammonium chloride [14] , [18] . Additionally , null or hypomorphic alleles of a housekeeping gene can move a cell to a radically different part of the fitness landscape , where epistatic effects can allow accumulation of favorable secondary mutations [13] , [22] , [23] . The key thread uniting these examples , and the broader array of cases presented in Table S1 , is that by altering gene expression and the flow of metabolites , loss-of-function mutations trigger far reaching changes in the cell's regulation and metabolism . As detailed in the meta-analysis presented below , these changes frequently prove adaptive under novel environments . The relative abundance of null mutations coupled with their adaptive potential suggests that specific null mutations likely represent common early steps in the evolution of bacterial populations encountering a new environment . Here , we examine comprehensively the potential of loss-of-function mutations for adaptation to novel environments . We first use a meta-analysis of genome-wide fitness data from transposon-insertion and in-frame deletion mutations across 144 conditions from 7 studies ( including new findings described below ) to show that adaptive null mutations are extremely abundant and disproportionately affect enzymatic and regulatory pathways . We then take as a case study the fitness profile of populations of E . coli transposon-insertion mutants in a set of unusual , nutrient-limited environments . The transposon insertions provide a convenient method to generate tagged null mutations that can be easily identified on a genome-wide scale and are likely to reflect phenotypes arising from common indels and point mutations that result in loss-of-function . In our media challenges , single loss-of-function mutations are sufficient to increase the growth rate up to twofold , demonstrating the suboptimal utilization of existing capacities by the wild-type strain and the ease with which null mutations can enhance fitness through metabolic and regulatory network rewiring .
Cases of beneficial null mutations have been noted previously in a wide variety of studies of both laboratory-evolved and wild strains; many of the best-characterized examples are summarized in Table S1 . Any such list , however , is biased by the limited set of conditions and mutants that have been characterized in detail . The increasing availability of quantitative fitness data from genome-wide screens of loss-of-function mutants in a wide variety of conditions allowed us to systematically study the adaptive potential of null mutations at a much more comprehensive scale . We performed a meta-analysis of null mutation fitness data from a total of 144 conditions from 7 studies in E . coli MG1655 and BW25113 ( including new data described below ) . For each condition , we identified genes for which null mutations gave significant increases or decreases in fitness and then examined the complete data set for evidence of over-representation of specific biological functions ( see Materials and Methods for details on the data sets , which included experiments from both in-frame deletions and transposon-mutagenized libraries , and statistical processing ) . While the relative portions of each functional class showing significant fitness effects ( positive or negative ) upon deletion varied greatly among the conditions ( see Fig . S1 ) , some clear trends were present . Overall , at least one beneficial null mutation was identified in all but five of the 144 conditions considered . In particular , we found adaptive ( and maladaptive ) deletions of regulatory proteins and enzymes in over half of the experimental conditions assayed ( Fig . 3A , B ) , while significant contributions from other classes were generally less frequent . For a more quantitative assessment , we used a resampling approach to determine the significance of each category's contribution to the observed fitness changes relative to its size ( Fig . 3C , D and Table S2 ) . Only enzymes and regulatory proteins showed enrichments of null mutations that both raise and lower fitness . Structural proteins and RNA genes also contained significant numbers of beneficial deletions , mainly due to a large number of beneficial deletions from those classes present in a small number of experimental conditions ( we could not , however , detect a notable unifying factor in the conditions under which null mutations in these classes of genes were beneficial ) . The RNA case in particular is dominated by two conditions under which transposon insertions in ribosomal RNAs were beneficial , and thus must be viewed with some caution . On the other side , membrane proteins , lipoproteins , and cell process proteins also contributed higher than expected numbers of deleterious null mutations , although their contributions were still lower than the frequency for enzymes or regulatory proteins . The abundance of adaptive regulatory null mutations was perhaps our most striking finding; the fact that purely regulatory mutations can allow bacteria to adapt to a wide variety of extreme conditions illustrates the extent to which the physiological capabilities of microbes exceed their regulatory logic , and the relative ease with which knockouts of appropriate regulators can rapidly rewire a maladaptive regulatory network . It was also instructive to consider the breadth of conditions under which a given null mutation could be adaptive; the set of genes for which null mutations were beneficial in at least 10 conditions in our meta-analysis is shown in Table S3 . Consistent with the above findings , 6 out of 7 such genes coded for either enzymes or regulatory proteins , and housekeeping genes ( lon , dnaJ ) played a particularly prominent role . It is likely that these null mutations , as well as loss of function of the nucleoid-associated protein fis , exert their widespread beneficial effects by globally altering expression of other genes , similar to the mechanism of action of a recently characterized rho hypomorph that proved beneficial in more than ten different conditions [23] . The previously published data sets analyzed above consist primarily ( although not entirely ) of chemical or physical hazards added to otherwise standard growth media . An equally realistic scenario for a microbe is to encounter nutrients that the organism's metabolism is poorly equipped to utilize . The relative roles of regulatory rewiring vs . acquisition of new functions in adaptation to such conditions and the potential for adaptive null mutations in these cases remain largely unexplored . To further understand the potential for null mutations to alter fitness in the face of a metabolically challenging environment and to explore the mechanisms employed , we propagated a library of E . coli MG1655 transposon-insertion mutants [21] in four media conditions where the parental strain grew poorly ( defined M9 media with alanine , glutamine , aspartic acid , or asparagine as the sole carbon source; see Fig . 4A ) . In addition to including the data in our meta-analysis , we identified the 809 insertion locations that caused the greatest increases and decreases in fitness ( Fig . 4B ) ( see Materials and Methods and Dataset S1 ) . The use of a transposon library , containing ∼106 disruptive perturbations , allowed us to explore the space of possible adaptive null mutations more rapidly and comprehensively than evolutionary approaches . Although such mutations are unlikely to be found in the wild , the resulting phenotypes mirror those of common point mutations or small insertions and deletions that cause loss of function . Using the pathway analysis tool iPAGE [24] , we found that clusters of genes whose disruption was deleterious ( clusters 1–4 ) are enriched for genes whose products participate in nucleotide and amino acid biosynthesis , functions essential in the growth media we used ( Fig . 4B , C ) . In contrast , the clusters containing beneficial insertion locations ( clusters 5–9 ) showed varied and generally weak functional enrichments , suggesting that alterations to many distinct pathways can increase fitness . As transposon insertions do not necessarily cause a null phenotype [21] , we tested in-frame deletions for a representative set of candidate genes in three of the growth conditions ( Table S4 ) . As expected , many of the null mutants grew significantly faster than the parental strain in the experimental media ( Fig . 5 and Table S5 ) . Doubling times dropped by as much as 30% for alanine media and nearly 50% for each of glutamine and asparagine media – a substantial fitness increase – showing how poorly the parental strain utilizes its existing capacities in these extreme environments . Transcriptome analyses of four fitter-than-wildtype mutants in each of alanine and glutamine media ( Dataset S2 ) revealed that each mutant had a distinct expression pattern . While overlaps among the genes up- and down-regulated in individual mutants were generally larger than would be expected by chance ( Table S6 ) , the number of genes whose expression exhibited large ( >2-fold ) changes ( Fig . S2A , B ) and the functional categories overrepresented among the differentially expressed genes varied widely among the mutants ( Fig . S2C–F ) . In particular , expression differences among chemotaxis and flagellar biosynthesis genes were especially prominent ( Fig . S3 ) . The diversity of transcriptome changes with a net beneficial effect illustrates the non-optimality of the wild-type genetic network in the experimental media and the varied possibilities for improvement . Additionally , the breadth of transcriptome changes in the Δpgi and ΔcysQ strains ( Fig . S2A , B ) demonstrates the potential for enzymatic null mutations to rewire a large portion of the cell's regulatory and metabolic network . To better understand the mechanisms by which the null mutations tested above lead to increased fitness , we used flux balance analysis ( FBA ) , which determines in a regulation-independent fashion ways a cell could use its metabolic capabilities to maximize its growth rate in a specific environment [25] . FBA simulations indicated that E . coli attains its maximum growth rate in alanine media when the glycine cleavage complex ( GCC ) is not utilized ( Fig . 6A ) , consistent with our observed benefits of deletion of GCC components ( Fig . 5A ) . The cost of synthesizing increasing amounts of serine only to degrade it to glycine likely accounts for the decreasing growth rate as GCC flux increases ( Fig . 6B ) . Simulations also indicated that phosphoglucose isomerase should be inactive during rapid growth in alanine media because flux through the enzyme creates a futile cycle ( Fig . 6C , D ) ; our results validated that prediction ( Fig . 5A ) . Both of these examples illustrate how fitness defects can be caused , not by lack of enzymatic functions , but rather their context-inappropriate utilization . Overall , flux variability analysis [26] indicated that proteins encoded by ∼860 ( numbers range from 858 for glutamine media to 869 for alanine media ) of the 1260 genes in the iAF1260 genomic reconstruction for E . coli K-12 MG1655 [27] catalyze reactions in pathways that must be ‘off’ to allow maximum growth ( see Materials and Methods ) . Likely a variety of deletions , acting both directly and indirectly , can reduce or eliminate the superfluous fluxes . Consistent with the results of our meta-analysis above , deletions of regulators also provided substantial fitness advantages; prominent examples are cpxA in alanine media ( +22% growth rate ) and lrp in both glutamine ( +82% growth rate ) and asparagine media ( +72% growth rate ) . It is also useful to note that most of the beneficial mutations studied here are neutral or deleterious in environments other than the one in which they were identified ( Fig . S4 ) , consistent with the notion that they introduce specific perturbations that increase cellular fitness in the new environment . Thus , our results indicate that even when faced with an environment that imposes severe metabolic challenges , null mutations can alter the regulatory and metabolic network of bacterial cells to greatly increase fitness without the gain of additional enzymatic functions , supporting our broad hypothesis that null mutations play a substantial role in adapting to diverse novel environments .
Our experimental results and meta-analysis of previous studies demonstrate the substantial potential of loss-of-function ( null ) mutations to aid in adaptation to novel environments through regulatory and metabolic rewiring . We find that the overarching effect of many null mutations is to improve the match between a cell's regulatory network , which is well-adapted to the organism's native habitat , and the contingencies of the new environment . This is particularly true for deletions of genes in the two functional classes in which we see the most widespread over-representation of beneficial null mutations: enzymes and regulatory proteins ( Fig . 3 ) . Regulatory mutations , especially those in cis-regulatory sequences , have long been thought to play an important role in adaptation ( reviewed in [5] ) , and our work shows that null mutations in regulators themselves also make a substantial contribution , increasing and decreasing the activity of cellular modules and facilitating the emergence of new phenotypes . The prevalence of adaptive null mutations in regulators illustrates that the phenotypic capabilities of bacterial cells – that is , the range of environments in which they possess the capacities to thrive – far exceed their regulatory capacity , the range of environments in which they can respond productively . When cells possess the biochemical capabilities for thriving under extreme conditions but fail to deploy those resources due to the constraints of the overlying regulatory network , regulatory mutations can rapidly allow the appropriate expression of those phenotypic capabilities . The GASP ( growth advantage in stationary phase ) phenotype [28] that arises in very old E . coli cultures provides a clear example: prolonged incubation in stationary phase yields cells with mutations that greatly enhance stationary phase fitness , including null mutations in the regulator lrp [29] and mutations attenuating activity of the sigma factor rpoS [30] , [31] . Enzymatic deletions also remodel cellular networks , albeit in a different way . Metabolic engineers are quite aware that well-chosen deletions can boost yields by redirecting fluxes or removing undesirable byproducts [32] , [33] , and the present work presents multiple examples of the utility of silencing enzymes . Similarly , when a cell's regulatory network erroneously expresses a metabolic pathway , knockouts of one of the component enzymes can often ameliorate the fitness deficit . For example , Bollenbach and coworkers recently found that bacterial growth in the presence of DNA synthesis inhibitors was suboptimal due to overexpression of ribosomal RNA operons under these conditions and could be improved by deletion of most copies of those genes [34] . A beneficial mutation need not cause a large fitness gain to impact the trajectory of an evolving population . Many of the beneficial null mutations discussed here have phenotypic effects , such as approximately twofold changes in antibiotic minimum inhibitory concentration ( MIC ) , smaller than the eventual level of adaptation observed in laboratory-evolved or clinical populations . Nevertheless , a population with even a small advantage under stressful conditions will be favored over time , and the increased growth rate itself will increase the odds ( per unit time ) of acquiring additional adaptive mutations . Furthermore , even the accumulation of many mutations of individually small effect can give rise to a dramatic phenotypic difference , as has been observed in the case of antibiotic resistance in both laboratory strains [18] and clinical [35] populations . The eventual evolutionary trajectory of the population may include reversion of the original adaptive null mutation , if the bacteria re-encounter conditions where the gene function is beneficial . Beneficial null mutations also enable rapid fitness increases by presenting a large mutational target size . While both null mutations and the acquisition of novel protein functions can cause widespread alterations to cellular phenotypes , the comparatively higher probability with which null mutations occur amplifies their importance in adaptive evolution . In contrast with gain of function mutations that require one of a few specific changes to a protein or regulatory element , loss of function mutations can arise from any frameshift , nonsense mutation , or insertion in a coding region if it occurs early enough in an ORF , as well as through a variety of missense mutations specific to any given protein . We very conservatively estimate that null alleles arise at a rate on the order of 10−8 per gene per cell division ( assuming the mutation rate is on the order of 10−10 per nucleotide [36] , null alleles arise only from a nonsense mutation in the first half of an ORF , the average gene length is 1 kb , and codon usage is uniformly distributed ) . Bacteria also carry genetic programs for generating additional diversity under stress through error-prone DNA repair pathways [37] , [38] , likely making it even easier for cells to acquire adaptive null mutations through the generation of frameshift or missense mutations . Genomic rearrangements mediated by insertion elements can likewise further accelerate creation of loss of function mutations . For example , beneficial loss of function of the rbs operon has been observed to arise at a frequency of 5 * 10−5 per generation in laboratory evolution experiments due to the operon's proximity to an IS150 element [39] . The combination of the rate at which null mutations arise and the breadth of circumstances under which these mutations can be beneficial may be at least partly responsible for the observation that E . coli acquire small beneficial mutations ( ∼1% change in fitness ) at a surprisingly high rate of about 10−5 per generation [40] . Consistently , beneficial null mutations have frequently been shown to make substantial contributions to fitness in laboratory evolution experiments [11] , [12] and in a wide variety of natural conditions ( reviewed in Table S1 and in the examples below ) . Most of the beneficial null mutations studied here were identified in a single culture condition ( albeit with the usual fluctuations in media composition that occur with cell growth in batch culture ) ; it is likely that adaptation to novel natural environments involves an even more complex interplay of physicochemical parameters , where antagonistic pleiotropy may reduce the adaptive potential of single null mutations . However , far from being laboratory artifacts , adaptive null mutations are being increasingly recognized in natural and clinical settings as well . For example , null mutation-mediated adaptation contributed to the divergence of Bacillus anthracis from a Bacillus cereus ancestor . In addition to two virulence-factor encoding plasmids ( pXO1 and pXO2 ) , B . anthracis is characterized by a specific and ubiquitous nonsense mutation in plcR , which encodes a pleiotropic transcriptional activator [41] , [42] . The plcR null mutation in B . anthracis leads to significant reduction in the secretion of several degradative enzymes and virulence factors [43] . Although conflicting reports exist about the evolutionary pressures underlying the selection of this null mutation [43] , [44] , the current hypothesis is that plcR inactivation is part of the co-evolution of the chromosome and the pXO1 and pXO2 plasmids that led to the emergence of B . anthracis as a separate species [42] . The evolution of pathogenic Shigella strains from their E . coli ancestors was also mediated by null mutations in several anti-virulence genes , in addition to the acquisition of pathogenicity islands and a virulence plasmid [45] , [46] . Deletion of the cadA gene and null mutations in the nadA and nadB genes in the Shigella genome prevent the formation of cadaverine and quinolinate respectively , and both these molecules inhibit multiple aspects of Shigella pathogenicity [47]–[49] . Similarly , null mutations in speG allow the accumulation of spermidine , which increases Shigella resistance to oxidative stress and survival within macrophages [50] . Beneficial null mutations not only aid in the evolution of new species of pathogens , but can also facilitate the repeated adaptation of infecting pathogens to specific host niches . For example , null mutations in key regulators mediate adaptive diversification of Pseudomonas aeruginosa during chronic lung infections in cystic fibrosis patients , leading to non-piliation , flagellum loss , lack of quorum-sensing , and mucoidity from increased alginate production [51] . The most common cause of the switch to mucoidity is loss of mucA , which encodes an anti-sigma factor that sequesters AlgT , an activator of alginate biosynthetic genes [52] . Loss-of-function mutations in lasR , which encodes a transcriptional regulator , are frequently seen in isolates from the cystic fibrosis lung and lead to quorum-sensing-negative phenotypes and reduced expression of virulence factors [53] . The phenotypes resulting from these deletions are within the physiological capabilities of the P . aeruginosa genome but are normally repressed by the regulatory network . Null mutations in important regulators alter the expression of entire modules and rewire the network to enable P . aeruginosa to adapt from its original niches as a free-living organism and acute infectious agent to long-term survival as a chronic infection in a host , although this adaptation may be important only for the specific infecting population and not for the species at large . Improved understanding of the contributions of null mutations to fitness is thus crucial for elucidating the evolutionary paths taken by evolving bacterial populations . These findings might also facilitate progress on other challenges such as understanding bacterial adaptation during chronic infections , engineering bacteria for introduction into novel environments or microbial communities , and culturing ‘unculturable’ bacteria [54] . Such ‘unculturable’ species might possess all of the biochemical capabilities necessary to grow in monoculture on common cultivation media , but simply not utilize them properly in an environment so different from their native habitat , leading to an adaptation barrier to lab conditions . Culturing such bacteria may thus require more sophisticated interventions than simple supplementation with additional nutrients .
Unless otherwise noted , media was M9 [55] lacking NaCl ( 48 mM Na2HPO4 , 22 mM KH2PO4 , 19 mM NH4Cl , 2 mM MgSO4 , 0 . 1 mM CaCl2 , and 10 µM thiamine ) , supplemented with 2 g/L of the carbon source and micronutrients [56] at the following final concentrations: 3 nM ( NH4 ) 6 ( Mo7O24 ) , 400 nM H3BO3 , 30 nM CoCl2 , 10 nM CuSO4 , 80 nM MnCl2 , and 10 nM ZnSO4 . No supplementary iron source was added . LB media was 1% Bacto Tryptone , 0 . 5% yeast extract , and 0 . 5% NaCl . Due to glutamine's limited stability in solution , we prepared glutamine media fresh for each experiment . Media used for growth curves with glucose included 0 . 01% Tween-20 to eliminate optical artifacts due to biofilm formation [23] . Unless otherwise noted , we grew cell cultures at 37°C and shook them at 250 rpm . To make clean , in-frame deletions , we transduced KanR ( kanamycin resistance cassette ) marked alleles from the Keio collection [57] into strain AH28 ( MG1655 ΔlacZ ) using P1vir phage [58] and removed the markers using a FLP recombinase system [59] . We confirmed each mutant's identity by comparing sizes of PCR products of the region containing the putative gene deletion in the mutant and parental strains . Table S4 lists all strains used in this work . Before starting the single amino acid cultures , we grew thawed aliquots from the transposon library [21] in LB for three generations and washed the cells in M9 salts lacking a carbon source . Next , we added ∼108 cells to 5 ml of M9 media with the appropriate amino acid as the sole carbon source . Using serial transfers , we maintained the cultures in exponential phase above a minimum population size of ∼108 . To reduce the impact of spontaneous mutations while allowing for the detection of subtle fitness effects , we harvested and analyzed cultures after twenty generations [60] . We carried out transposon footprinting as described previously [21] . Data ( ratios of transposon signal to genomic DNA signal ) were sum-normalized and then log-transformed ( base 2 ) to give increases and decreases similar magnitudes . Arrays were normalized to the mean of five hybridizations of the transposon library prior to selection [21] by fitting a loess [61] curve ( with the span parameter set to 0 . 3 ) to the intensities on the experimental array as a function of the mean intensities for the same genes on the reference arrays , and then subtracting from each gene the loess-predicted value . After normalization , transposon insertion locations that did not change in abundance in response to growth in single amino acid media should be distributed around zero . As a summary statistic for each gene in a given condition , we used the value closest to zero if the normalized values from all replicates had the same sign and zero if they did not . To evaluate the significance of the summary statistics , we constructed a separate null distribution of 500 , 000 “genes” for each of the four amino acids . Each gene contained either three ( for alanine , aspartic acid , or glutamine ) or two ( for asparagine ) data points . Samples for each gene came from a t-distribution with 4 degrees of freedom , with standard deviation equal to the standard deviation of the normalized experimental samples of a randomly chosen gene for the amino acid of interest and mean set to the median of the five data points for a ( possibly different ) randomly chosen gene from the normalized , unselected hybridizations . Summary statistics were calculated for the null distribution as they were for the data , and gene level p-values were set to the fraction of null genes with summary statistics exceeding the actual observed value in magnitude . We chose the significance cutoff for each amino acid separately to give an estimated 5% FDR . We excluded genes that the Profiling of E . coli Chromosome database version 4 marked as essential ( http://www . shigen . nig . ac . jp/ecoli/pec/index . jsp ) [62] . Of the 3792 genes tested for significance , 809 were significant in at least one condition . Expression profiles were subjected to k-means clustering using Euclidean distance as the distance metric . For each gene , we included the expression level in each biological replicate as well as the average across replicates for each condition . During clustering , we assigned columns of averages ten times the weight of columns of individual biological replicates . For visualization purposes , enrichment values were restricted to the range between −3 and 3 , and extreme values are shown as either −3 or 3 . FBA simulations used the iAF1260 genomic reconstruction for E . coli K-12 MG1655 [27] in MATLAB with SBML and COBRA toolboxes [26] . Simulations were done in computational minimal media [27] with the sole carbon source set to 10 mmol g DW−1h−1 with the Ec_biomass_iAF1260_core_59p81M biomass objective function . A gene was deemed non-essential for maximum growth in a medium if simulation of the full model and the model lacking that gene gave the same growth rate . We used Flux Variability Analysis [26] to identify fluxes that needed to be zero to obtain the maximum growth rate . Then , all non-essential genes that either by themselves or in combination with other genes directly catalyzed those reactions were considered to be in a pathway that needed to be zero for maximum growth . Due to numerical noise , fluxes were not required to be exactly zero; changing the thresholds did not alter the results qualitatively . All growth curves in 96-well plates used flat-bottom , untreated , polystyrene plates ( Corning #3370 ) with 150 µl of media per well . To reduce evaporation , we covered samples with 100 µl mineral oil [63] . A SynergyMx ( Biotek; Winooski , VT ) read the absorbance at 600 nm . We subtracted the absorbance of wells with media and oil but no cells from all readings as background . Unless otherwise specified , the reader shook the plates continuously on its ‘medium’ setting and maintained the temperature at 37°C . For growth curves in glucose media , we grew most strains overnight in the test media and diluted 375-fold into fresh media . Due to their slow growth rate on glucose , we grew strains ZD8 , Z18 , ZD56 , ZD59 , and ZD60 overnight in glucose media supplemented with alanine , proline , and asparagine ( 0 . 5 g/L each ) and then washed them before final dilution into glucose media . We measured absorbance every 8 minutes for 36 hr and calculated growth rates as the least squares fit to the logarithm of the part of the background-corrected growth curve between 0 . 015625 and 0 . 0625 ( before taking the logarithm ) . Most strains doubled at least three times before reaching the target absorbance range . For the remaining strains , we identified the exponential growth region by hand and adjusted the target range as necessary . The r2 value of each fit was required to be greater than 0 . 99 . To determine doubling times in alanine media , we grew cultures overnight in LB , washed them , and diluted them 300-fold into media in 96-well plates . We shook plates at 250 rpm in an incubator and measured absorbance several times a day starting at ∼20 hours after inoculation; we kept cultures in exponential phase ( background-corrected absorbance less than 0 . 15 ) using 15-fold serial dilutions . The doubling time estimates came from least-squares fits to the logarithm of the background-corrected absorbance readings multiplied by the total dilution prior to the reading . Data for each fit included at least 4 samples ( average 11 . 3 ) spanning at least 6 generations ( average 14 . 9 ) and yielded an r2 value of at least 0 . 95 . We determined growth rates in glutamine and asparagine media in two stages . As an initial filter , we attempted to determine growth rates in 96-well plates as was done for alanine media , but the wide range of doubling times resulted in lower quality data than we had obtained in alanine media . Thus , we retested those mutants that exhibited an advantage over the parental strain individually . In this second stage , which was used to generate all data reported for glutamine and asparagine media , we grew strains as 20 ml cultures in 250 ml flasks and shook them at 250 rpm . We removed culture samples several times a day and read the absorbance at 600 nm on an Ultrospec 3100 pro . We started cultures by diluting washed , LB-grown overnight cultures 100-fold into fresh test media , and after ∼2 generations of growth , we diluted cultures a second time . Sampling started after an additional ∼1 generation of growth ( ∼3 generations total in the test media ) when the absorbance reached ∼0 . 01 and continued until the absorbance exceeded 0 . 1 . We identified the linear portion of the logarithm of each growth curve manually and then subjected it to a linear least-squares fit to determine the doubling time . We washed and diluted LB-grown overnight cultures into glutamine or alanine media . After ∼5 generations of growth , we harvested samples undergoing mid-exponential phase growth and added 2 ml of culture to 4 ml of RNAprotect Bacteria Reagent ( Qiagen ) . We incubated the mixture at room temperature for 5 min and then centrifuged it at 5000 g for 10 min . We removed the supernatant and stored the pellet at −80°C . We isolated RNA using the Norgen Total RNA Purification Kit according to the manufacturer's directions except that in the last step we eluted the RNA in 35 µL of the kit's elution solution . We poly-adenylated the RNA by combining 31 µl RNA ( undiluted from the previous step ) with 4 µl 10× Poly ( A ) Polymerase Reaction Buffer ( New England Biolabs ) , 4 µl 10 mM ATP , and 1 µl ( 5 U ) E . coli Poly ( A ) polymerase ( New England Biolabs ) and incubating at 37°C for 30 minutes . Then , we cleaned samples with an RNeasy Mini Kit ( Qiagen ) and labeled them with cyanine 3-CTP or cyanine 5-CTP dye using the Low Input Quick Amp Labeling Kit ( Agilent ) starting with 200 ng of RNA per sample . We labeled strain AH28 with Cyanine 5-CTP and mutants with Cyanine 3-CTP . We then hybridized samples to an Agilent E . coli Gene Expression Microarray ( 8×15K format , Catalog # G4813A-020097 ) according to the manufacturer's instructions , scanned the resulting slides using a High-Resolution C Scanner ( Agilent ) , and extracted features using Agilent's Feature Extraction Software version 9 . 5 using protocol GE2-v5_95_Feb07 without spike-in controls . We used the ‘LogRatio’ value in subsequent analyses . We averaged all values for the same ORF and values from the two biological replicates performed for each comparison . To estimate the false positive rate , we approximated the null distribution by taking the difference of the values from the two biological replicates for the same gene and dividing by two . This produced a data set with a zero mean and the same noise distribution as that produced by averaging . We calculated a single null distribution for all 8 samples ( 4 mutants in alanine and 4 in glutamine ) . Then , the chance of a false positive was the number of samples from the null distribution greater than 1 or less than −1 ( i . e . , a two-fold change ) . The false discovery rate is the estimated number of false positives divided by the number of genes deemed significant . For each mutant , we ran iPAGE [24] in discrete mode on three sets of genes: those whose expression increased at least 2-fold between the mutant and the parental strain , those whose expression decreased at least 2-fold , and the remaining genes . We also ran iPAGE in continuous mode with various numbers of bins and identified categories similar to those in Fig . S2 . Expression data are in Dataset S2 and in the Gene Expression Omnibus ( accession GSE30345 ) . We used a total of 144 data sets showing the fitness effects of null mutations in E . coli K12 strains; we obtained 113 from the comprehensive characterization of knockout strains ( in the BW25113 background ) performed by Nichols et al . [14] , with the remainder coming from a series of experiments on transposon mutagenized libraries ( in the closely related MG1655 background ) performed by the Tavazoie laboratory [17] , [18] , [21] , [23] , [64] including this work . We excluded all genes identified as potentially essential during the construction of a gene-by-gene deletion library in BW25113 [57] or in a series of chromosomal deletions [62] from analysis , as null mutations of essential genes are clearly impossible . In combining the studies , we followed the significance calling metrics of the original authors as closely as possible . For the datasets from Girgis et al . [18] we used the published significance criteria . For data from Freddolino et al . [23] , we generated a p-value for each gene by resampling the probe level scores from the full genome-wide distribution 10 , 000 times to create a null distribution , and then applied a 1% FDR for significance calling . Otherwise , for conditions with two or more biological replicates , we determined significance at a FDR of 5% as we did for the single amino acid experiments in this work . The selections from Girgis et al . [21] were extremely stringent , making insertions resulting in average and below-average fitness effectively indistinguishable; hence , for those data sets we only included beneficial insertions in the meta-analysis . Amini et al . 's [64] data set on biofilm induction by poly-N-acetylglucosamine did not contain any significant genes at a 5% FDR , so we instead marked as significant only the three gene deletions whose phenotypes the work experimentally confirmed . Similarly , when only a single biological replicate was available for a condition ( motility in high-salt media [21] or fitness in various ethanol concentrations [17] ) , we counted as significant only those deletions whose fitness contributions the studies individually verified . We assembled a single ( non-concentration-specific ) set of deletions altering fitness in ethanol . To identify significant deletions in the Nichols et al . [14] data set , we retained the authors' normalization ( each of 324 experiments individually normalized to zero mean and IQR = 1 . 35 ) and the authors' null model ( normal distribution with zero mean and standard deviation of one ) . Then , considering all experiments collectively , we chose a cutoff corresponding to a 5% FDR . Finally , for each series of dosage titrations for a given condition , we used only the data from the highest dose ( 113 experiments total ) . We excluded data from strains carrying hypomorph alleles of presumed essential genes . To assess the significance of the numbers of beneficial or deleterious null mutations of different classes relative to that expected if the class labels were not significant , we performed the following resampling test: for each gene class/condition combination , we generated simulated distributions with the same total number of elements as the number of genes considered from that class in the corresponding condition in the real data , with the probability of each element being ‘true’ ( that is , beneficial or deleterious ) equal to the average probability of a gene being beneficial or deleterious ( as appropriate ) across all genes under that condition . For each gene class , we then took the sum of ‘true’ elements across all conditions as a summary statistic . The ( one-tailed ) p-value for enrichment of beneficial ( or deleterious ) genes in each class is obtained by comparing the observed number of beneficial ( or deleterious ) genes in that class to 10 , 000 simulated draws for the same class; the p-value is the fraction of those simulated draws which yield a summary statistic greater than or equal to the observed value . Significance of these classes was then determined by applying the Benjamini-Hochberg procedure [65] to the raw p-values , to identify classes that were significant at an FDR of 0 . 01 . The resampling procedure described here yields the distribution shown in Figure 3CD and the q-values in Table S2 . Expression Data has been uploaded to the Gene Expression Omnibus ( GEO ) ( accession GSE30345 ) . | When bacteria encounter a new challenge in their environment , such as treatment with an antibiotic or a poor nutrient source , their population faces tremendous selective pressure to evolve in order to grow better under the new conditions . We typically think of bacterial evolution in terms of what is gained: a bacterium might , for example , acquire an antibiotic resistance gene , or modify an existing enzyme to make better use of a nutrient source . By analyzing the fitness of bacterial populations under more than 100 different conditions , we show that in fact what they lose can be equally important: by rewiring the cell's metabolism , loss of function mutations can provide substantial fitness benefits under many challenging conditions , even cases such as exotic nutrient combinations where some new enzymatic function might seem to be required . Loss of function mutations occur at a much higher frequency than gains of specific functionality due to the larger mutational target area available . The combination of the rapid acquisition and broad functionality of loss-of-function mutations suggests that they play a major role in the early adaptation of bacterial populations to new challenges . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"biotechnology",
"escherichia",
"coli",
"systems",
"biology",
"prokaryotic",
"models",
"genomics",
"model",
"organisms",
"genetics",
"synthetic",
"biology",
"biology",
"computational",
"biology",
"evolutionary",
"biology",
"microbiology"
] | 2013 | Bacterial Adaptation through Loss of Function |
The function of neuronal networks relies on selective assembly of synaptic connections during development . We examined how synaptic specificity emerges in the pontocerebellar projection . Analysis of axon-target interactions with correlated light-electron microscopy revealed that developing pontine mossy fibers elaborate extensive cell-cell contacts and synaptic connections with Purkinje cells , an inappropriate target . Subsequently , mossy fiber–Purkinje cell connections are eliminated resulting in granule cell-specific mossy fiber connectivity as observed in mature cerebellar circuits . Formation of mossy fiber-Purkinje cell contacts is negatively regulated by Purkinje cell-derived BMP4 . BMP4 limits mossy fiber growth in vitro and Purkinje cell-specific ablation of BMP4 in mice results in exuberant mossy fiber–Purkinje cell interactions . These findings demonstrate that synaptic specificity in the pontocerebellar projection is achieved through a stepwise mechanism that entails transient innervation of Purkinje cells , followed by synapse elimination . Moreover , this work establishes BMP4 as a retrograde signal that regulates the axon-target interactions during development .
The specificity of synaptic connectivity in the central nervous system is a prerequisite for brain function . The neuronal circuits in the vertebrate cerebellum represent a remarkable example of wiring specificity . This was first recognized by Santiago Ramón y Cajal when he chose cerebellar circuits as revealed by the Golgi method for his early studies on brain organization ( discussed in [1] ) . In its simplest form , the cerebellar microcircuit integrates input from two afferent classes—climbing and mossy fibers . Climbing fibers selectively innervate Purkinje cells . By contrast , mossy fiber afferent activity is relayed to Purkinje cells via granule cells in the inner granular layer of the cerebellum ( IGL ) [2]–[4] . In the IGL , mossy fibers also form synapses on Golgi cells , a class of inhibitory interneurons that provide feed-forward inhibition in the cerebellar circuit . Climbing and mossy fiber information is then integrated in Purkinje cells and transduced via cerebellar efferent projection neurons in the deep cerebellar nuclei . Despite the apparent simplicity of the cerebellar circuit , it is unknown how the specificity of synapse formation emerges during development for each of the principal cerebellar afferent systems . Indeed , the molecular mechanisms regulating synapse specificity for most circuits in the mammalian brain have remained obscure . Two key steps determining the incipient pattern of synaptic connectivity during development are axon-target contact formation and synaptic differentiation . Ultrastructural reconstruction of mature neuronal circuits suggests that only a subset of contacts differentiates into bona fide synapses [5] . The fraction of actual synapses compared to cellular contacts ( potential synapses ) has been termed “filling fraction” , with a filling fraction of 1 . 0 representing a case where all contacts are synaptic structures [6] . In vertebrate and invertebrate systems several attractive and repulsive factors have been identified that contribute to synaptic specificity [7]–[13] . However , pinpointing whether these specificity factors regulate primarily selective contact formation , synaptic differentiation , or both has been challenging , given the limited resolution of light microscopy in assessing direct cellular contacts in vivo . One possibility is that some signaling pathways regulate primarily contact formation , whereas other factors drive the synaptic differentiation process after axon-target contacts are established . The ponto-cerebellar projection represents an excellent model system to explore mechanisms of synaptic specificity in the mammalian brain [14] . Mossy fiber axons emerging from the basilar pons ( PGN ) in the ventral brain stem form a major projection to the cerebellar cortex which relays information from sensory and motor cortex . Structurally , mossy fiber afferents exhibit synaptic specificity at two levels: Mossy fiber axons elaborate synapses exclusively with granule and Golgi cells but not Purkinje cells . At the subcellular level , mossy fiber synapses are restricted to the proximal regions of Golgi cells within the IGL but are excluded from the molecular layer where distal Golgi cell dendrites arborize . Cell culture studies indicated that immature granule cells provide a stop-signal for mossy fiber growth [15] . Maturing granule cells , in contrast , contribute positive signals for the differentiation of mossy fiber synapses and the elaboration of mossy fiber glomeruli [16]–[19] . However , it is unknown how contact and synapse specificity emerges for mossy fibers and their granule and Golgi cell targets . Moreover , negative signals that suppress contact of mossy fibers with inappropriate target cells in vivo have not been identified . Previous anatomical studies indicated that axons with the appearance of mossy fibers do not exhibit strict targeting specificity but form broader projection patterns during early postnatal stages [20]–[22] . Fibers with mixed mossy fiber and climbing fiber morphology ( “combination fibers” ) were observed to contact Purkinje cells during development though the extent of these interactions remained unknown . A developmental gene expression analysis in pontine nuclei revealed distinct transcriptional programs for axonal growth and synaptic differentiation of pontine mossy fibers [23] . Surprisingly , the termination of the axonal growth program did not require the presence of granule cells in the cerebellar cortex but was perturbed in mutant mice with Purkinje cell degeneration accompanied by granule cell death [23] . In combination , these studies raised the question of whether Purkinje cells may provide signals for the development of mossy fiber projections . To define the cellular nature of mossy fiber–Purkinje cell interactions and the rearrangements resulting in the specific synaptic wiring pattern , we undertook a systematic analysis of axon-target interactions in the mouse ponto-cerebellar system . Using correlated light-electron microscopy analysis we quantitatively mapped physical contacts and synaptic structures formed between identified pontine mossy fibers and Purkinje cells . Using this methodology , we observed extensive transient mossy fiber contacts and synapses on Purkinje cells that are subsequently eliminated . Patterning molecules such as WNTs , FGFs , and BMPs have been shown to exert novel neuronal signaling functions at the Drosophila neuromuscular junction and in the mammalian central nervous system [24]–[26] . Of this class of molecules , the BMPs have been extensively studied as retrograde signals at the Drosophila neuromuscular junction [27]–[32] and as trophic factors in mammalian neurons [33]–[35] . However , their signaling functions in vertebrate axon–target interactions have not been determined . We explored a role for BMP signaling in mossy fiber transient target interactions in the developing cerebellum . Using expression analysis , in vitro assays , and conditional knock-out mice we identify BMP4 as Purkinje cell-derived signal that specifically controls mossy fiber–target contact selectivity during development .
To examine the emergence of synaptic specificity of ponto-cerebellar mossy fibers we adopted an in utero electroporation approach [36] , [37] . Pontine precursor cells are selectively electroporated by injection of DNA constructs into the 4th ventricle at embryonic day 14 . 5 ( Figure S1 ) . Following differentiation and migration , these cells settle into the pontine gray nucleus ( Figure 1A , B ) [36] , [37] . Pontine axons labeled by electroporation of an EGFP expression plasmid project to the cerebellar cortex assume typical mossy fiber morphology and are restricted to the IGL at postnatal day 21 ( P21 ) , consistent with the selective elaboration of mossy fiber–granule cells synapses in the adult cerebellar circuitry ( Figure S1 ) . To examine how this target specificity emerges , we examined earlier developmental time points . At P7 , we identified a significant number of GFP-positive mossy fiber extensions projecting beyond the IGL into the Purkinje cell layer ( PCL ) ( Figure 1C–E ) . Using three-dimensional analysis of high resolution confocal stacks , mossy fiber varicosities were found in direct proximity with calbindin-positive Purkinje cell somata and axons , suggesting direct mossy fiber–Purkinje cell contacts ( Figure 1F , G ) . These contacts contained synaptic markers , as they concentrated the endogenous synaptic vesicle protein VAMP2 or a synaptophysin-fluorescent protein fusion that was introduced by electroporation into the pontine projection neurons ( Figure 1H , I ) . Mossy fibers emerge not only from the PGN but multiple pre-cerebellar nuclei . The elaboration of mossy fiber–Purkinje contacts was not unique to PGN-derived mossy fibers as it was also observed in GFP-O transgenic mice where multiple other mossy fiber populations are marked by EGFP ( Figure S1 ) [38] . Some of these contacts concentrated the postsynaptic scaffolding protein Shank1a ( Shank1a-positive: 16% of somatic contacts , n = 50 contacts; 44% of contacts with proximal PC axon segments , n = 167; Figure 1J ) . In sum , these findings suggest that mossy fiber afferents establish transient synapse-like contacts with Purkinje cells during postnatal development . In order to visualize the developmental progression of mossy fiber–Purkinje cell contacts and their differentiation into synapses we undertook a systematic light-electron microscopy analysis of PGN-derived mossy fiber axons . Pontine projections were labeled by DiI tracing followed by photoconversion of the dye ( Figure 2 ) . Mossy fiber projection patterns and Purkinje cell interactions in the cerebellar hemispheres ( Crus1 , Crus 2 , and Simplex lobules ) were quantitatively examined by light microscopy and apparent contacts were subsequently analyzed by electron microscopy . By light microscopy , mossy fiber rosettes and the thin protrusions extending from them are seen in great detail , and somata of Purkinje and granule cells of the cerebellar cortex can be clearly identified by DIC microscopy ( Figures 2B , C , S2A–D ) . To further confirm the identity of the Purkinje cell territory , some sections were additionally labeled with antibodies to calbindin ( Figure S2I , J ) . Camera lucida drawings of mossy fiber terminals from 50 µm sections at P0 , P7 , P14 , and P21 revealed that the mature mossy fiber projection pattern emerges from a series of transformations during postnatal development that includes an extensive invasion of and eventual withdrawal from the Purkinje cell territory ( Figure 2D , E ) . At P0 mossy fibers extend far into the developing cerebellar cortex where Purkinje cells are unevenly distributed and intermixed with migratory and maturing granule cells of the emerging IGL ( Figures 2D , S2 ) . At P7 , Purkinje cells form a recognizable monolayer above the IGL . However , pontine mossy fibers substantially invade this Purkinje cell territory with close to 30% of all labeled segments in the IGL penetrating into the PCL ( Figure 2E , D , see Methods and figure legends for details on quantitative analysis ) . This invasion of the PCL was significantly reduced at postnatal days 14 and 21 , yielding the mature mossy fiber projection pattern . At postnatal days 7 and 14 mossy fibers in close apposition to Purkinje cell somata often exhibited marked varicosities , resembling the presumptive mossy fiber–Purkinje cell synapses identified using the in utero electroporation approach ( Figure 2C , D ) . Fifty-five putative contacts identified at the light microscopy level in P0 , P7 , P14 , and P21 tissues were examined by electron microscopy . Tissue sections ( 50 µm ) were re-sectioned into 7 µm semithin sections and re-examined again by light microscopy . Sections encompassing the putative mossy fiber–Purkinje cell contacts were then thin-sectioned ( 70 nm ) and processed for ultrastructural analysis ( Figure S2E–H ) . Cytological characteristics defined in previous studies allowed unambiguous identification of Purkinje and granule cell somata in electron micrographs , as well as other relevant cellular components of the cerebellar cortex ( Figure S3 ) [3] , [39] . Over 90% of putative contacts between mossy fibers and Purkinje cell somata identified by light microscopy at P0 and P7 indeed represent direct cellular appositions ( Figure 3A , B , E ) . At P0 , none of the mossy fiber–Purkinje cell ( or mossy fiber–granule cell ) contacts in the developing IGL/PCL had synaptic features , representing a filling fraction ( synapses per contacts [6] ) of 0 . 0 ( Figure 3A , F ) . However , at P7 a substantial number of mossy fiber–Purkinje cell contacts exhibited ultrastructural characteristics of synapses ( Figure 3B , E , F; several consecutive sections shown in Figure S4A , filling fraction = 0 . 37 ) . At P14 , direct contacts were still observed ( 5 direct contacts verified by EM of 13 putative contacts analyzed ) but only one of them was synaptic ( Figure 3C , F , filling fraction = 0 . 2 ) . Finally , at P21 no direct mossy fiber–Purkinje cell contacts or synapses could be identified ( Figure 3D ) . During the apparent removal of contacts and the elimination of synapses between P7 and P14 we frequently observed mossy fibers separated from the Purkinje cell soma and/or ensheathed by glial processes ( Figure S4C–F ) , reminiscent of pruning processes with axosome shedding observed in peripheral axons [40]–[42] . In addition , we observed instances where mossy fiber axons were engulfed by Purkinje cells ( Figure S4C , D ) . Glial process ensheathing of Purkinje cells became even more prominent at P21 . Some mossy fibers were positioned as close as 150 nm from the Purkinje cell soma ( Figure 3D ) but glial processes separated mossy fiber endings and the Purkinje cell soma . In summary , during the first 10 postnatal days approximately 30% of all labeled IGL mossy fibers derived from the PGN establish direct contacts with Purkinje cells . The quantitative analysis uncovers remarkable developmental changes in the “filling fraction” , i . e . the differentiation of direct mossy fiber–Purkinje cell contacts into synapses , rising from 0 . 0 at birth to 0 . 37 at postnatal day 7 ( Figure 3F ) . In the second to third postnatal week , these synapses are eliminated and the contacts withdrawn , resulting in the selective innervation of granule and Golgi cells in the IGL . Notably , not all mossy fiber axons contact Purkinje cells . Therefore , specific signaling mechanisms must exist , first , to limit the invasion of pontine mossy fiber axons into the Purkinje cell territory during the first postnatal days and , second , to promote the removal of pontine mossy fiber–Purkinje cell synapses in the second postnatal week of development . In Drosophila melanogaster , growth factors of the bone morphogenetic protein ( BMP ) family regulate synaptic growth , axon arborization , and synaptic homeostasis [27] , [28] , [43]–[45] . To explore whether a comparable signaling function might be conserved in the mouse cerebellum , we surveyed the expression of BMP signaling molecules in the developing ponto-cerebellar projection system . Using in situ hybridization , we detected mRNAs for BMP receptor 1A ( BMPR1A ) , BMP receptor 1B ( BMPR1B ) , and BMP receptor type 2 ( BMPR2 ) in the PGN at P0 , the time when pontine mossy fiber axons extend into the cerebellar cortex . By P14 , detection of BMPR1B mRNA was reduced , while signals for BMPR1A and BMPR2 expression persist ( Figure 4A ) . Within the cerebellar cortex significant expression of several BMP ligands was observed consistent with previous reports ( [46]–[48] and unpublished data ) . We focused our analysis on BMP4 as it is highly expressed in Purkinje cells and dynamically regulated during the refinement of mossy fiber connectivity ( Figure 4B ) . At P0 , BMP4 mRNA is abundant in proliferating and premigratory granule cells of the EGL , and in scattered Purkinje cells ( identified by their large diameter ) . BMP4 expression in Purkinje cells was reduced at P7 , the time when mossy fiber–Purkinje cell synapses are most common , and expression was strongly up-regulated in Purkinje cells by postnatal day 14 ( Figure 4B ) . At P21 , BMP4 was highly expressed in Purkinje cells . In addition a subset of large diameter cells in the IGL ( presumably Golgi cells ) expressed BMP4 . In summary , BMP4 and its signaling receptors are appropriately positioned to regulate mossy fiber target selection during postnatal development . BMP-receptor activation results in phosphorylation of cytoplasmic SMAD proteins that translocate to the cell nucleus and activate transcription [49] , [50] . Classical morphogenetic functions of BMPs depend on SMAD phosphorylation but phospho-SMAD ( pSMAD ) -independent BMP signaling read-outs have also been described [34] , [51]–[53] . Robust SMAD phosphorylation was detected when recombinant BMP4 was added to cultured pontine explants in vitro and phosphorylation was prevented by co-application of the antagonist noggin ( Figure S5A ) . Quantitative evaluation of SMAD phosphorylation in PGN in vivo using Western blot and immunohistochemistry revealed a dynamic regulation , with moderate levels at P0 , strongly increased levels at P14 , and persistent pSMAD immune-reactivity at P21 ( Figure S5B–D ) . SMAD1 , 5 , 8 protein levels were not significantly altered during this developmental time period , suggesting that regulation of SMAD signaling occurs primarily at the level of SMAD phosphorylation ( unpublished data ) . This demonstrates a functional BMP signaling pathway in developing pontine neurons in vitro and in vivo . Given that BMP4 was dynamically expressed in Purkinje cells we asked whether Purkinje cell-derived BMP4 was required for SMAD activation in pontine neurons . We analyzed conditional BMP4fl/fl::Pcp2cre/cre knockout ( BMP4 cKO ) mice lacking BMP4 expression selectively in Purkinje cells . In the Pcp2cre knock-in line , cre-mediated recombination is detected during late embryonic stages and specifically in Purkinje cells ( [54] and Figure S6 ) . Ablation of BMP4 expression was verified by in situ hybridization ( Figure S6C ) . Interestingly , SMAD activation in the pontine gray nucleus was not dramatically altered in BMP4 cKO mice ( Figure S5D , E ) . While we cannot completely exclude that some of the pSMAD signal is due to incomplete ablation of the BMP4 expression in Purkinje cells , these results indicate that Purkinje cell-derived BMP4 might not be essential for pSMAD activation in pontine nuclei during postnatal development . Notably , Purkinje cells express significant amounts of BMP7 and other BMP growth factors during postnatal development which might be responsible for the persistent SMAD phosphorylation in the absence of BMP4 ( unpublished data ) . Importantly , signaling activities have been described for specific BMP growth factors that control cytoskeletal rearrangements through pSMAD-independent pathways [34] , [51]–[53] . In commissural spinal neurons such BMP signals represent extrinsic cues for the initial polarization of axons [55] , [56] . Therefore , we examined the possibility that target-derived BMP signaling might regulate axon development and axon-target interactions of pontine mossy fibers . Previous work demonstrated that cerebellar explants cultured in vitro release a growth inhibiting activity for mossy fibers which is thought to resemble a target-derived stop signal for afferents [57] . Explants from the PGN exhibit robust radial axon outgrowth . However , when pontine explants are co-cultured with cerebellar tissue , axon growth on the side facing the cerebellar tissue is reduced , suggesting the presence of a growth inhibiting activity derived from cerebellar tissue ( Figure 5A ) . In order to assess a possible role for BMPs in this process , we applied the soluble BMP antagonist noggin to the culture medium . Noggin addition blocked cerebellar growth retardation activity in this assay ( Figure 5B , D ) . To directly examine whether BMP4 might exert such growth-inhibiting activity , we combined pontine explants with BMP4-expressing HEK293 cells in collagen gel co-cultures . Using this assay , we observed that BMP4 was sufficient to negatively regulate mossy fiber growth in vitro ( Figure 5C ) . Importantly , this activity could be neutralized by addition of noggin to the collagen gel matrix , indicating that the growth regulation was indeed mediated by BMP signaling ( Figure 5E ) . Therefore , BMP4 negatively regulates pontine mossy fiber growth in vitro . Based on the dynamic regulation of BMP4 expression in Purkinje cells and the repulsive activity of BMP4 towards mossy fiber axons in vitro , we hypothesized that BMP4 might control either initial mossy fiber–Purkinje cell interactions , the detachment of mossy fiber–Purkinje cell contacts , or both . To explore these possibilities , we further examined the BMP4 cKO mice . Given the important patterning functions of BMP signaling in early cerebellar development [46] , [47] , [58] we first asked whether the overall anatomical organization of the cerebellar cortex or specification of cerebellar cell types was perturbed in the mutant mice . No significant changes were detected in the foliation pattern , cerebellar layering , specification of the major cell types , and expression of transcriptional markers and signaling molecules ( Figure S7 ) . In DiI labeled preparations , the number of labeled pontine mossy fiber axons or density of Purkinje cells observed in the cerebellar cortex of BMP4 cKO mice was not significantly different from control littermates or wild-type animals ( Figure S7G and unpublished data ) . Finally , the development of climbing fibers and formation of vGlut2-positive climbing fiber synapses on the Purkinje cell dendrites was not noticeably altered in the BMP4 cKO mice ( Figure S8 ) . Next , we examined whether loss of Purkinje cell-derived BMP4 resulted in defects in mossy fiber–Purkinje cell contact formation , synapse formation , and/or synapse elimination . In BMP4 cKO mice the fraction of pontine mossy fibers that penetrated into the Purkinje cell territory at postnatal day 0 was increased approximately 2-fold as compared to control ( Figure 6A , B ) . Moreover , the number of Purkinje cells receiving mossy fiber contacts was increased 7-fold at P0 and remained significantly increased over the following 2 wk . In our wild-type analysis ( Figure 2 ) we identified a peak in mossy fiber elimination from the Purkinje cell territory in the second postnatal week . Mossy fiber elimination was quantitatively compared using an elimination index for the fraction of mossy fibers removed from the PCL between P7 and P14 ( [MFsPCL P7–MFsPCL P14] / MFsPCL P7 ) . In the cKO animals elimination of mossy fibers still occurred but the elimination index for control and cKO mice was reduced to about 50% of that in control animals ( Figure 6C ) . When normalized to the length of mossy fiber segments in the Purkinje cell territory , the density of contacts per 100 µm mossy fiber length was more than 3-fold increased at postnatal day 7 ( Figure 6C ) . These observations highlight an essential function for BMP4 in the control of initial mossy fiber–Purkinje cell contact formation during the first postnatal week as well as the subsequent removal of mossy fiber processes from the Purkinje cell territory . Considering the developmental regulation of the filling fraction observed in wild-type animals ( Figure 3 ) we further examined mossy fiber–Purkinje cell contacts , synapses , and filling fractions at P7 using correlated light-electron microscopy . As in wild-type and control tissue , the majority of putative mossy fiber–Purkinje cell somatic contacts identified in BMP cKO mice indeed represented direct cellular appositions ( 38 direct contacts out of 40 potential contacts analyzed , Figure 6D , E ) . Some mutant contacts were characterized by unusual , irregular synapse-like profiles ( Figures 6D , S4B ) . However , the filling fraction was substantially reduced in the cKO as only 11% of these contacts exhibited synaptic ultrastructure ( Figure 6E ) . Based on the correlated light-EM analysis and the calculated filling fraction , the total density of mossy fiber–Purkinje cell synapses was not significantly changed , indicating that the excess contacts do not efficiently differentiate into synaptic structures . These experiments identify BMP4 as a retrograde signal that specifically controls mossy fiber-Purkinje cell contact formation and highlight that independent programs regulate contact versus synapse formation during postnatal development . If loss of BMP4 from Purkinje cells results in exuberant pontine mossy fiber–Purkinje cell contacts during early postnatal development , do these aberrant interactions perturb the placement or specificity of synapses in the mature cerebellum ? Using in utero electroporation , we marked pontine mossy fibers in control and BMP4 cKO animals and examined their projection pattern at P21 when cerebellar development is essentially complete ( Figure 7 ) . While in wild-type mice mossy fibers were restricted to the IGL and did not protrude into the molecular layer , we observed overshooting axons in the BMP4 cKO mice ( Figure 7A ) . A subset of mossy fiber axons penetrated more than 20 µm beyond the Purkinje cell somata into the molecular layer , a phenotype never observed in control cerebella ( Figure 7B ) . Within the molecular layer , most mossy fiber axons had a smooth appearance but some developed swellings comparable to simple mossy fiber rosettes . Overshooting mossy fiber axons have been observed previously in mouse mutants with perturbed granule cell migration [59] . However , we did not observe ectopic granule cells in the molecular layer of BMP4 cKO mice ( Figure 7C ) . Instead , high-resolution analysis of the overshooting mossy fiber axons revealed that some established direct contacts with the dendritic tree of Purkinje cells . Other overshooting axons formed contacts with neurogranin-positive Golgi cells ( Figure 7C ) . Notably , Golgi cells are one of the specific synaptic targets of ponto-cerebellar mossy fibers in the IGL . However , in wild-type mice mossy fiber synapses are excluded from the distal dendritic arbors in the IGL . Finally , we examined the position of pontine mossy fiber synapses in the IGL and observed a significant shift of mossy fiber rosettes towards the PCL ( within 40 µm of the Purkinje cell somata , Figure 7 ) . Therefore , loss of Purkinje cell-derived BMP4 results in persistent alterations in mossy fiber connectivity in the mature cerebellum .
BMPs are key regulators of patterning and cell fate decisions , but novel functions in neuronal wiring are emerging [49] , [60]–[62] . In the vertebrate central nervous system BMPs ( and the related TGFbeta growth factors ) control initial axon orientation and axon regeneration [55] , [56] , [63]–[65] . Moreover , retrograde , target-derived BMP signaling has been examined in the peripheral nervous system [66]–[69] . At the Drosophila neuromuscular junction a muscle-derived BMP-analogue regulates synaptic growth and homeostatic signaling [27] , [29] , [30] , [43] , [70] . Whether BMP growth factors have similar retrograde signaling activities in the central nervous system and , specifically in axon-target interactions in vertebrates , has remained unclear . In our experiments , we explored retrograde BMP signaling in the mouse pontocerebellar system and uncovered a novel function during the development of synaptic target specificity . In this system , BMP4 acts as a negative signal that limits interactions of mossy fibers with Purkinje cells , a transient target cell . The dynamic regulation of BMP4 expression in Purkinje cells mirrors the assembly of mossy fiber-Purkinje cell contacts and synapses , with a transient peak at P7 where BMP4 expression is low . Thereafter , BMP4 is strongly up-regulated and mossy fiber-Purkinje cell contacts are eliminated . The mossy fiber phenotypes in the BMP4 cKO mice highlight a critical function of Purkinje cell-derived BMP4 in mossy fiber–Purkinje cell interactions . In the cKO mice , there is a substantial increase in mossy fiber–Purkinje cell contacts at early postnatal stages ( P0–P7 ) . This supports an essential repulsive role for BMP4 in target recognition which limits the initial mossy fiber–Purkinje cell contacts and restricts the invading mossy fiber axons to their target territory in the IGL . The correlated light-electron microscopy analysis enabled us to dissociate changes in contact and synapse formation in the cerebellar system . Notably , while BMP4 cKO mice exhibit a 3-fold increase in mossy fiber–Purkinje cell contact density we did not detect a comparable increase in synapse density . Therefore , axon target contacts and synapse formation are controlled by different signaling systems . The subsequent , removal of mossy fiber–Purkinje cell contacts and elimination of mossy fiber processes from the Purkinje cell territory was significantly delayed , and after completion of cerebellar development , we observed persistent overshooting mossy fiber projections in the Purkinje cell and molecular layers . Some overshooting axons retain interactions with Purkinje cells , while others form contacts on distal Golgi cell dendrites . Notably , Golgi cells are appropriate synaptic partners of mossy fibers , but in BMP4 cKO cerebella mossy fiber–Golgi cell interactions are observed ectopically in the molecular layer . These findings support an important role for Purkinje cell-derived BMP4 in eliminating mossy fiber projections from this area , in addition to its function in regulation of the early mossy fiber–Purkinje cell contacts . Within the IGL , the placement of mossy fiber rosettes was shifted towards the Purkinje cell layer , further supporting a repulsive role for Purkinje cell-derived BMP4 . However , the fact that most mossy fiber axons did not overshoot to the molecular layer indicates that there are additional signals that restrict mossy fibers to the IGL . BMP2 and 7 transcripts are up-regulated in Purkinje cells of the cKO mice ( unpublished data ) and may partially compensate for the loss of BMP4 . Moreover , the specificity of mossy fiber connectivity is likely to emerge not only from negative , Purkinje cell-derived signals but from an interplay with positive signals derived from the appropriate target cells . Granule cells express FGF22 , Wnt7a , and neuroligins which all have been demonstrated to have positive , synaptogenic activities towards mossy fiber afferents [16]–[18] . Therefore , presentation of these synaptogenic signals by mature granule cells which strongly increase in number at later postnatal stages ( P7–P21 ) may compete with the constant number of Purkinje cells for mossy fiber contact . A prediction of this model is that direct mossy fiber-Purkinje cell synapses would persist in the absence of granule cells . This is , indeed , observed in agranular cerebella of mouse mutants or after irradiation where mossy fiber target selectivity can be examined in the absence of the appropriate synaptic targets [71] . Importantly , while most mossy fibers were appropriately restricted to the IGL , the positioning of mossy fiber rosettes within the IGL was shifted closer to the Purkinje cell layer ( Figure 7D ) , consistent with the loss of a negative regulator of synaptic connectivity in Purkinje cells of BMP4 cKO mice . The finding that the development of mossy fiber target specificity involves not only extensive contact but also synapse formation with Purkinje cells argues against a model of absolute recognition specificity for unique synaptic targets . This remodeling of transient target interactions is reminiscent of interactions in the thalamo-cortical projection and for Cajal Retzius cells in the hippocampus [72]–[74] . In both cases , afferents enter the target territory before their appropriate target cells have fully differentiated and form transient synapses on a third cell type ( subplate neurons and Cajal Retzius cells , respectively ) . This situation in the hippocampus is comparable to the transient mossy fiber–Purkinje cell synapses described in our study that are elaborated during early postnatal development when only few granule cells have descended into the forming IGL . While the initial assembly of such transient contacts is comparable , the mechanism of contact removal is fundamentally different . Elimination of transient synapses received by subplate and Cajal Retzius neurons occurs via programmed cell death of the transient target cells . By contrast , removal of mossy fiber–Purkinje cell interactions occurs independently of Purkinje cell death and requires signals for contact destabilization . The existence of widespread mossy fiber–Purkinje cell interactions during development poses the question of whether these synapses simply represent an imprecision in the initial trans-synaptic interactions or whether transient contacts serve a purpose in the development of functional cerebellar circuits . In the hippocampus , Cajal Retzius cells appear to be required for the laminar specificity of entorhinal axon projections [74] . Similarly , in the absence of subplate neurons , thalamocortical axons do not establish appropriate synaptic connectivity [75] , [76] . Therefore , transient mossy fiber–Purkinje cell interactions might similarly contribute to the assembly of cerebellar circuits . The cerebellar cortex is subdivided into longitudinal bands identified by specific molecular codes of gene expression in Purkinje cells [77] , [78] . This code develops during the first postnatal weeks , coincident with emergence of mature cellular and sub-cellular targeting specificity of both climbing and mossy fiber afferents . Recent tracing studies indicate that there is a precise somatotopic matching of pontine and climbing fibers [79] , [80] . This raises the possibility that transient mossy fiber–Purkinje cell interactions might provide a mechanism to coordinate mossy fiber and climbing fiber development and , thereby , serve a functional role in the assembly of cerebellar circuits .
All animal experiments were reviewed and approved by the institutional animal care and use committee of Columbia University and the cantonal veterinary office Basel , respectively . Mice were of the NMRI ( Figure 1 ) and C57BL/6 strains ( all other experiments ) . PCP2cre knock-in mice were previously described [54] . The conditional BMP4 floxed allele ( BMP4fl ) was generously provided by Dr . Brigid Hogan [81] . Htr5b-GFP mice are BAC transgenic mice generated by the GENSAT consortium [82] and were obtained from the MMRRC repository . Thy1 . 2-GFP ( GFP-O ) mice were generated by Drs . Sanes and Feng [38] and were obtained from the Jackson Laboratory . R26-lox-stop-lox-YFP were described in [83] . Timed-pregnant mice ( NMRI or C57BL6 background ) were used at embryonic day 14 . 5 following the protocol described in [37] . After electroporation , the mice were brought to term , pups were sacrificed by transcardial perfusion with 4% paraformaldehyde in 100 mM Na-phosphate buffer ( pH7 . 4 ) , and tissue from successfully electroporated pups ( P7 , P14 , P21 ) was processed for immunohistochemistry . The following primary antibodies were used in this study: rabbit anti-Shank1a [84] , goat anti-Car8 ( Frontiers Institute ) , rabbit and mouse anti-Calbindin D-28K ( Swant ) , rabbit anti-GFP [85] , rabbit anti Pax6 ( Covance ) , guinea pig anti-vGlut1 ( Chemicon ) , mouse anti-vGlut2 ( Chemicon ) , goat anti-Parvalbumin ( Swant ) , rabbit anti-Smad1 ( Millipore ) , rabbit anti-phospho-SMAD1/5/8 ( Millipore ) , rabbit anti-SMAD1 , 5 , 8 ( Imgenex ) , mouse anti-NeuN ( Millipore ) , rabbit anti-neurogranin ( Abcam ) , mouse anti-actin ( Sigma ) , and mouse anti-VAMP2 ( Synaptic Systems ) . Most procedures followed standard protocols; see Text S1 for details . High-resolution images of 30 to 40 µm z-stacks consisting of 0 . 45 µm thick optical sections were acquired using Zeiss LSM510 , a Zeiss LSM5 Exciter , or a LIS-spinning disk confocal system . Direct apposition of cellular markers was identified by rotating the 3D reconstruction of the stacks using Imaris Software ( Bitplane ) . Quantitative assessment of SMAD1 , 5 , 8 activation was performed using pSMAD1 , 5 , 8 and NeuN immunolabelling on 50 µm thick sagittal section using Metamorph software . The percentage of pSMAD1 , 5 , 8 positive cells among the NeuN positive cells and the pSMAD1 , 5 , 8 fluorescence intensity per NeuN area was determined through intensity thresholding and integrated morphometry analysis using MetaMorph software . DiI ( 1 , 1′-dioctadecyl-3 , 3 , 3′ , 3′-tetraindocarboyanine perchlorate , Molecular Probes ) labeling was modified from a previously published procedure [86]; see Text S1 for details . Quantitative assessment of mossy fiber invasion into Purkinje cell territory was performed on camera lucida drawings of DiI and calbindin double-labeled material ( 50 µm coronal “thick” sections , 100× objective ) of P0 , P7 , P14 , and P21 cerebellar hemispheres ( crus1 , crus2 , and simplex lobules ) . All camera lucida drawings contained all of the labeled mossy fiber segments and Purkinje cell outlines drawn from a fixed area size of 175 µm horizontal×120 µm vertical×30 µm deep ( thickness of one Purkinje cell soma ) , encompassing the upper IGL and PCL , and spanning a stretch containing on average 40 Purkinje cells at P0 ( before PC alignment occurs ) , and 9 Purkinje cells at P7–P21 . For the sake of consistency , and since Purkinje cell density and the angle of the mossy fiber segments penetrating the PCL differs at the base , versus apex , versus sides of the folia , areas for analysis were always drawn from the sides of the folia . The percentage of mossy fiber segments invading into the PCL out of all mossy fiber segments drawn per area was scored ( >20 segments per area from >25 areas obtained from 5–9 animals per time point , >3 50 µm section per animal were analyzed ) . For the quantification of Purkinje cells receiving putative somatic contacts from mossy fibers , contacts were judged as varicosities in the mossy fiber axon at the site apparently immediately adjacent to Purkinje cell soma . For quantification of contact density per mossy fiber length ( in Figure 7D ) camera lucida drawings were scanned at 600 dpi , and the length of mossy fiber segments in the PCL was measured using line tool in NeuronJ [87] . The number of putative contacts on Purkinje cell somata per mossy fiber segment was scored visually , using the criteria described above . The filling fraction was calculated as actual synapses divided by the number of contacts ( EM-verified ) . The elimination index for mossy fibers projecting into the Purkinje cell layer ( MFsPCL ) was calculated using the data points for P7 and P14 ( graph Figure 6B ) as follows: [MFsPCL P7–MFsPCL P14] / MFsPCL P7 . | Brain functions rely on highly selective neuronal networks which are assembled during development . Network assembly involves targeted neuronal growth followed by recognition of the appropriate target cells and selective synapse formation . How neuronal processes select their appropriate target cells from an array of interaction partners is poorly understood . In this study , we have addressed this question for the axons emerging from the pontine gray nucleus , a major brainstem nucleus that relays information between the cortex and the cerebellum , a brain area responsible for the control of skilled movements but also emotional processing . Using advanced microscopy techniques , we find that developing mossy fibers establish synaptic contacts rather promiscuously , and elaborate extensive synapses with Purkinje cells , an inappropriate target . These contacts are subsequently eliminated , and proper synaptic connectivity is then restricted to granule and Golgi neurons . We identify bone morphogenetic protein 4 ( BMP4 ) as a regulator of these inappropriate mossy fiber-Purkinje cell contacts . BMP growth factors are best known for their functions in cell specification during embryonic development , and our results support an additional retrograde signaling function between axons and their target cells in early postnatal stages . In summary , we show that the specificity of the synaptic connections in the ponto-cerebellar circuit emerges through extensive elimination of transient synapses . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"molecular",
"neuroscience",
"cellular",
"neuroscience",
"neuronal",
"morphology",
"neuroanatomy",
"neural",
"networks",
"signaling",
"pathways",
"biology",
"neuroscience"
] | 2011 | Development of Axon-Target Specificity of Ponto-Cerebellar Afferents |
Burkholderia pseudomallei causes the tropical infection melioidosis . Pneumonia is a common manifestation of melioidosis and is associated with high mortality . Understanding the key elements of host defense is essential to developing new therapeutics for melioidosis . As a flagellated bacterium encoding type III secretion systems , B . pseudomallei may trigger numerous host pathogen recognition receptors . TLR5 is a flagellin sensor located on the plasma membrane . NLRC4 , along with NAIP proteins , assembles a canonical caspase-1-dependent inflammasome in the cytoplasm that responds to flagellin ( in mice ) and type III secretion system components ( in mice and humans ) . In a murine model of respiratory melioidosis , Tlr5 and Nlrc4 each contributed to survival . Mice deficient in both Tlr5 and Nlrc4 were not more susceptible than single knockout animals . Deficiency of Casp1/Casp11 resulted in impaired bacterial control in the lung and spleen; in the lung much of this effect was attributable to Nlrc4 , despite relative preservation of pulmonary IL-1β production in Nlrc4−/− mice . Histologically , deficiency of Casp1/Casp11 imparted more severe pulmonary inflammation than deficiency of Nlrc4 . The human NLRC4 region polymorphism rs6757121 was associated with survival in melioidosis patients with pulmonary involvement . Co-inheritance of rs6757121 and a functional TLR5 polymorphism had an additive effect on survival . Our results show that NLRC4 and TLR5 , key components of two flagellin sensing pathways , each contribute to host defense in respiratory melioidosis .
Burkholderia pseudomallei is a tropical soil saprophyte and Tier 1 select agent that causes the infection melioidosis [1] . The bacterium may be inoculated through the skin , inhaled , or ingested . Although infection can manifest in myriad ways , pneumonia is identified in 50% of cases . Mortality from melioidosis ranges from 14–40% despite appropriate antibiotic treatment , and the risk of death is higher with pulmonary involvement [2] , [3] . This indicates an urgent need for a better understanding of host-pathogen interactions in melioidosis and adjunctive immuno-modulatory therapies . Innate immune mechanisms of recognition of invading bacteria include membrane-bound Toll-like receptors ( TLRs ) and cytosolic NOD-like receptors ( NLRs ) [4] , [5] . These pathogen recognition receptors bind conserved pathogen associated molecular patterns and drive the host response . For example , as a Gram-negative , flagellated bacterium , B . pseudomallei is predicted to activate sensors of LPS ( such as TLR4 ) and flagellin ( such as TLR5 ) . We have found that B . pseudomallei LPS is a TLR4 ligand that drives much of the innate immune response to B . pseudomallei , and that human genetic variation in TLR4 is associated with susceptibility to melioidosis [6]–[8] . We have also shown that B . pseudomallei activates TLR5 , and that polymorphisms in TLR5 are associated with survival from melioidosis [9] , [10] , however TLR5-deficient mice have not been infected with B . pseudomallei to demonstrate the role of TLR5 in an experimental setting . These findings point to an important role for flagellin in activation of immune responses in melioidosis . Whereas TLR5 detects flagellin at the cell surface , cytosolic flagellin is detected through NLRC4 , an inflammasome that activates caspase-1 [11] . NLRC4 is one of a number of NLRs that can assemble a canonical caspase-1-dependent inflammasome that in turn cleaves pro-IL-1β and pro-IL-18 to their active forms and induces pyroptosis [5] , [12] . More recent work identified murine NAIP5 and NAIP6 as direct flagellin sensors that signal through NLRC4 [13] , [14] . NLRC4 also contributes to the sensing of bacterial components other than flagellin: murine NLRC4-NAIP1 and NLRC4-NAIP2 inflammasomes recognize bacterial type three secretion system ( T3SS ) needle and rod proteins , respectively [13]–[16] . In contrast to mice , humans have only a single NAIP , and in human U937 monocytes the NLRC4-NAIP inflammasome recognizes a T3SS needle protein but not flagellin [14] . The functional interpretation of the three NLRC4 agonists is similar – flagellin , rod , and needle protein are all believed to be accidentally injected into the cytosol by bacterial T3SS . This is in contrast to TLR5 detecting extracellular flagellin , the presence of which will not be strictly linked to a particular virulence trait . In addition to the canonical caspase-1-dependent inflammasome , a noncanonical inflammasome involving another inflammatory caspase , caspase-11 , has recently been described in mice [17] . Caspase-11 protects mice from B . pseudomallei infection [18] . In this study , our primary objective was to determine the relative importance of NLRC4 in murine respiratory melioidosis in comparison to TLR5 , and with respect to canonical and noncanonical inflammasomes . Our secondary objective was to test whether genetic variation in NLRC4 was associated with outcome in human respiratory melioidosis .
All animal experiments were approved by the University of Washington Institutional Animal Care and Use Committee ( protocol number 2982-03 ) . The University of Washington complies with all applicable provisions of the federal Animal Welfare Act and with the Public Health Service ( PHS ) Policy on Humane Care and Use of Laboratory Animals . The University of Washington Human Subjects Division Institutional Review Board; the Ethical Review Committee for Research in Human Subjects , Ministry of Public Health , Thailand; and the Ethics Committee of the Faculty of Tropical Medicine , Mahidol University , Bangkok , Thailand approved the human genetic studies on subjects who had provided or whose next of kin had provided written informed consent for enrollment into clinical studies of melioidosis at the time of recruitment . B . pseudomallei 1026b was grown in LB broth shaking in air at 37°C , washed twice , resuspended in PBS containing 20% glycerol , and frozen at −80°C . Immediately before each aerosol infection experiment , the freezer stock was thawed and diluted in PBS to the desired concentration , as previously described [19] . NLRC4 SNP identification and selection was performed using the Genome Variation Server ( http://gvs . gs . washington . edu/GVS/ ) . Coding SNPs in the gene and haplotype-tagging SNPs were selected . Within the region encompassed by 50 , 000 bases upstream and downstream of NLRC4 , SNPs with a minor allele frequency ≥2% in populations identified as Japanese , Chinese and Asian were binned into groups with R2≥0 . 8 to identify haplotype-tagging SNPs . Genotyping was performed using an allele-specific primer extension method ( Sequenom Inc . , San Diego , CA , USA ) with reads by a MALDI- TOF mass spectrometer [8] . Comparisons of two and three groups of data expected to follow a normal distribution were made using Student's t test and ANOVA with a Bonferroni post-test , respectively . CFUs were log10 transformed before analysis . Survival analyses were performed with the log rank test . SNPs were tested for deviation from Hardy-Weinberg equilibrium using the exact test . The association of genotype with death was performed using a Chi square test or , for contingency tables with cell counts <10 , the exact test . For multivariate analysis of genetic associations , logistic regression was performed adjusting for age , gender , diabetes , renal disease , or liver disease . A conservative Bonferroni correction was not performed as the variants are unlikely to be independent . Effect modification was assessed by testing the incorporation of an interaction variable into the regression model , using the likelihood ratio test . Statistics were performed with GraphPad Prism 5 . 0f ( San Diego , CA ) or Stata 11 . 2 ( College Station , TX ) . A two sided p value of ≤0 . 05 was considered significant .
Given our previous identification of a strong association between a nonsense TLR5 polymorphism that renders TLR5 insensitive to flagellin and survival from melioidosis [9] , we examined whether the presence of Tlr5 in murine melioidosis altered survival . We infected mice with 361 CFU B . pseudomallei per lung , a dose that approximates the median lethal dose ( Figure 1A ) . We found that Tlr5−/− mice had significantly poorer survival than wild type mice , a phenotype that contrasts with that observed in Tlr2−/− or Tlr4−/− mice [23] . We next asked how the absence of Nlrc4 modulated this phenotype . We infected Tlr5−/− and Tlr5−/−Nlrc4−/− mice with a similar dose ( 400 CFU/lung ) of B . pseudomallei but found no difference in survival between mouse strains ( Figure 1B ) . This finding suggested that the absence of flagellin sensing at the cell surface sufficiently impaired the host response such that impaired cytosolic detection of the pathogen did not substantially impact survival further . We then tested whether lack of Nlrc4 alone altered survival in respiratory melioidosis , and how this differed from combined deficiency of Tlr5 and Nlrc4 . To increase the sensitivity of our model , we chose a lower inoculum that is non-lethal to wild type mice ( 91 CFU/lung ) . We found that Nlrc4-deficient mice were more susceptible to melioidosis than wild type mice , consistent with results from Ceballos-Olvera [24] , but that there was no difference in survival between Nlrc4−/− mice and Tlr5−/−Nlrc4−/− mice ( Figure 1C ) . Together , these experiments demonstrate that TLR5 and NLRC4 each contribute to host defense in murine respiratory melioidosis . Caspase-11 has recently been identified as a component of the noncanonical , caspase-1-independent inflammasome . We and others have found that Casp1−/−Casp11−/− mice infected with B . pseudomallei by the respiratory route failed to control infection ( unpublished data , [24] , [25] ) . To examine the effects of NLRC4 relative to other caspase-1- and caspase-11- dependent inflammasomes , we directly compared bacterial burdens in organs of wild type , Nlrc4−/− , or Casp1−/−Casp11−/− mice infected with B . pseudomallei . Twenty four hours after an inoculum of 314 CFU/lung , bacterial growth in the lungs of both Nlrc4−/− and Casp1−/−Casp11−/− mice was about 0 . 87 log10 CFU greater than in wild type mice , and there was no significant difference between CFU in Nlrc4−/− compared to Casp1−/−Casp11−/− mice ( Figure 2 ) . Forty eight hours after infection , bacterial growth in the lungs of both Nlrc4−/− and Casp1−/−Casp11−/− mice had increased significantly compared to wild type mice ( by 1 . 87 log10 CFU and 2 . 69 log10 CFU , respectively ) . Despite a trend towards greater pulmonary bacterial burdens in Casp1−/−Casp11−/− mice than in Nlrc4−/− mice , this did not reach statistical significance . Bacterial burdens in the spleens are an indication of dissemination beyond the pulmonary compartment . Although bacteria were detectable 24 hours after infection , there were no significant differences between the three mouse strains . Forty eight hours after infection , CFU were significantly greater in Casp1−/−Casp11−/− mice compared to Nlrc4−/− mice which in turn had greater bacterial burdens compared to wild type mice . These data confirm that while deficiency of both caspase-1 and caspase-11 severely impairs control of B . pseudomallei replication in the lung , NLRC4 accounts for much of the inflammasome-dependent phenotype [24] . We next evaluated selected cytokine and chemokine responses in the lungs of these mice ( Figure 3 ) . There were no differences in TNF-α or MIP-2 levels between mouse strains at 24 hours . As expected , IL-1β was very low in Casp1−/−Casp11−/− mice but was not impaired in Nlrc4−/− mice . Chemokine KC was higher in Nlrc4−/− mice compared to wild type and to Casp1−/−Casp11−/− mice . By 48 hours after infection , TNF-α levels in Casp1−/−Casp11−/− mice were significantly greater than wild type . MIP-2 and KC levels in Casp1−/−Casp11−/− and Nlrc4−/− mice were higher than in wild type mice . IL-1β was elevated in all mice compared to 24 hour levels , but was significantly elevated in Nlrc4−/− mice in comparison to wild type and to Casp1−/−Casp11−/− mice . In serum 24 hours after infection , TNF-α , MIP-2 , and Il-1β levels were low but KC was readily detectable and higher in Nlrc4−/− mice compared to Casp1−/−Casp11−/− mice . At 48 hours , despite higher bacterial burdens in the spleens of Nlrc4−/− and Casp1−/−Casp11−/− mice compared to wild type mice , serum TNF-α and IL-1β remained uniformly low . In contrast , MIP-2 and KC levels increased substantially in both Nlrc4−/− and Casp1−/−Casp11−/− mice . In line with previously published data [24] , these results point to non-NLRC4-mediated pathways of IL-1β production in the lung , but suggest that systemically , NLRC4 mediates TNF-α and IL-1β but not MIP-2 or KC release . Inhalation of B . pseudomallei results in scattered , dense cellular pulmonary infiltrates [19] . Histopathologic examination of the lungs of Nlrc4−/− and Casp1−/−Casp11−/− mice 24 hours after airborne infection with B . pseudomallei showed relatively similar sized neutrophilic infiltrates and percent of lung involved in these mice compared to wild type mice although there was minor variation in morphologic features , such as earlier evidence of nuclear fragmentation in Casp1−/−Casp11−/− mice ( Figure 4 ) . However , at 48 hours , inflammation was more severe , particularly in Casp1−/−Casp11−/− mice , which displayed larger and necrotic parenchymal lesions that lacked identifiable intact inflammatory cells . We have found that a human genetic polymorphism in TLR5 is associated with outcome from melioidosis [9] . Given the clear role for Nlrc4 in murine respiratory melioidosis , we investigated whether human genetic variation in the NLRC4 region was associated with death in human respiratory melioidosis . We genotyped five NLRC4 region single nucleotide polymorphisms ( SNPs ) ( rs455060 , rs212703 , rs410469 , rs462878 , and rs6757121 ) selected as described in the methods in 173 melioidosis patients with clinical evidence of pulmonary involvement . The call rate for four SNPs was above 97 . 5%; one ( rs212703 ) was discarded due to a low call rate . Fifty eight of the 173 subjects ( 34% ) died . In survivors , no variant deviated from Hardy-Weinberg equilibrium . rs6757121 was associated with protection against death in a general genetic model , p = 0 . 012 ( Table 1 ) . Adjusting for age , sex , and pre-existing conditions , the effect was strongest in a dominant model [odds ratio ( OR ) 0 . 35 , 95% CI:0 . 13–0 . 91 , p = 0 . 03] . rs6757121 is located about 0 . 3 kb downstream of NLRC4 and occurs with a minor allele frequency of 10% . We next tested whether our previously reported association between TLR51174C>T – a nonsense polymorphism that truncates the receptor in the extracellular domain rendering it non-responsive to flagellin – and survival in melioidosis [9] is also seen in the subset of melioidosis patients with respiratory disease . We found that the adjusted OR of death was 0 . 14 , 95% CI 0 . 03–0 . 64 , p = 0 . 01 . To determine whether co-inheritance of this TLR5 variant and the NLRC4 region variant rs6757121 alters the risk of death from respiratory melioidosis , we assessed the effect of including both together in the model . The OR of death for each variant remained unchanged , although the effect of a cross-product interaction term could not be determined due to 100% survival in carriers of both variants . The estimated OR of death for carriers of both variants was 0 . 04 , 95% CI: 0 . 006–0 . 27 , p = 0 . 001 ( Table 2 ) . Together , these data show that the NLRC4 and TLR5 variants are each associated with survival and that co-inheritance of the variants has an additive but not synergistic effect .
The results of our investigations show that NLRC4 and TLR5 , key components of two flagellin sensing pathways , each contributes to host defense in murine respiratory melioidosis . We did not detect any additional impact of deficiency of both Nlrc4 and Tlr5 on survival . Furthermore , NLRC4 is responsible for much of the failure of pulmonary bacterial containment seen in caspase-1/-11-deficient mice . In humans , we show that an NLRC4 genetic variant is associated with survival in respiratory melioidosis , and there is an additive effect of co-inheritance of risk variants in TLR5 and NLRC4 . Recent investigations have demonstrated that NLRC4 is involved in recognition of several bacterial ligands such as components of the T3SS or flagellin , and this specificity is determined by various NAIPs [13]–[16] . In contrast , the only reported ligand of TLR5 is flagellin [4] . B . pseudomallei activates TLR5 [9] and aflagellated B . pseudomallei induces impaired TLR5-dependent NF-κB activation in vitro [unpublished data] . Our present results show that Tlr5−/− mice are more susceptible to B . pseudomallei in a model of respiratory infection , in contrast to deficiency in Tlr2 , which actually confers resistance , or Tlr4 , which has no apparent effect on survival [23] . Although MyD88 is an adapter molecule for all three of these TLRs , mice deficient in Myd88 show a similar phenotype to deficiency in Tlr5 after respiratory infection with B . pseudomallei [26] . Flagellin-sensing appears to be a crucial element of host defense in murine respiratory melioidosis . However , we have not observed significant impairment in TNF-α production from Tlr5−/− alveolar macrophages stimulated ex vivo with killed B . pseudomallei [unpublished data] and it is notable that our studies of murine respiratory infection with B . thailandensis ( a related and flagellated but less virulent organism ) have not shown any Tlr5-dependent phenotype [20] . Thus , in vitro data , and infections with model organisms may not fully recapitulate the complexity of in vivo infections with fully virulent B . pseudomallei . Like TLR5 , NLRC4 appears to play a central role in host defense in respiratory murine melioidosis . Interestingly , while NLRC4 detects flagellin from many bacterial species , it appears to not detect B . thailandensis ( and presumably B . pseudomallei ) flagellin [14] , thus , the effect of NLRC4 in vivo may be attributable to T3SS sensing . B . pseudomallei expresses several T3SSs [27] and T3SS3 facilitates virulence in a number of ways [28]–[31] . The B . pseudomallei T3SS rod and needle proteins BsaK and BsaL , respectively , are detected in an NLRC4-dependent fashion in mice [15] , [32] . Recent work by Bast et al demonstrates the importance of BsaK for NLRC4-dependent caspase-1 activation in B . pseudomallei-infected macrophages and for virulence in murine melioidosis [33] . Intriguingly , despite the different sensing functions of TLR5 and NLRC4 , absence of only one sensor imparts significant clinical impairment; there is no additive effect on survival of combined Tlr5 and Nlrc4 deficiency in murine melioidosis , even at doses that are non-lethal to wild type mice . NLRC4 is just one of many pathogen recognition receptors that activate the caspase-1-dependent inflammasome . The inflammasome processes pro-IL-1β and pro-IL-18 to their active forms and also induces pyroptosis , a caspase-1-dependent lytic cell death pathway . In our studies , Nlrc4−/− mice did not show a significant difference compared to Casp1−/−Casp11−/− mice with respect to bacterial replication in the lung following respiratory infection , but did show a difference in disseminated infection to the spleen , consistent with the work of Ceballos-Olvera et al [24] . This difference in dissemination may be due to caspase-11 , which also has been implicated in defense against B . pseudomallei [18] . Relative to Casp1−/−Casp11−/− mice , Nlrc4−/− mice showed preserved pulmonary IL-1β production . These data raise the possibility that much of the early effect of inflammasome-dependent control of bacterial replication in the lung is primarily NLRC4-dependent and the function of NLRC4 may be due to pyroptosis or as-yet-undefined roles of NLRC4 rather than cytokine processing . It may be that a secondary canonical inflammasome , perhaps NLRP3 , responds to B . pseudomallei infection only once bacterial burdens become extremely high , resulting in the observed IL-1β secretion . These observations are concordant with the work by Ceballos-Olvera et al , who additionally showed that processing of pro-IL1β to the active form was not impaired in Nlrc4−/− bone marrow-derived macrophages or in the bronchoalveolar lavage fluid of Nlrc4−/− mice infected with B . pseudomallei [24] . Furthermore , despite differences in experimental methods and timing , the histology of lungs from Nlrc4−/− mice infected with B . pseudomallei in our study appeared comparable to that of wild type mice treated with IL-1β and infected with B . pseudomallei by Ceballos-Olvera et al [24] . Notably , however , we found that systemic IL-1β and TNF-α levels were almost undetectable in Nlrc4−/− mice , despite high bacterial burdens in the spleen . This contrasted with high MIP-2 and KC concentrations in the serum , suggesting that there may be distinctly different regulatory effects of NLRC4 in various compartments . Our data also demonstrate that NLCR4 inflammasome-dependent innate immune signaling is not the same for B . pseudomallei as other Gram-negative pulmonary pathogens . Respiratory infection with Legionella pneumophila , another Gram-negative , flagellated , intracellular pathogen is also restricted by NLRC4 and this effect is dependent on the presence of flagellin [21] , [34] . However , following L . pneumophila infection there was no difference in bronchoalveolar lavage fluid cell counts or in lung cytokine levels of Nlrc4−/− mice compared to wild type mice , although there was greater histologic inflammation in the lungs of Nlrc4−/− mice [21] . As in B . pseudomallei infection , Nlrc4−/− mice are more susceptible to Klebsiella pneumoniae ( a non-flagellated , extracellular pathogen ) infection by the pulmonary route , with greater bacterial replication in the lungs , dissemination to the spleen , and death [35] although this effect was not observed at higher doses [36] . In contrast to our findings , pulmonary inflammation as assessed by TNF-α , KC , IL-1β , and MIP-2 levels and histologic score is reduced in Nlrc4−/− mice infected with K . pneumoniae [35] . These differences may be due to the presence of flagellin or the intracellular nature of B . pseudomallei , or to the apparent lack of NLRC4-mediated pyroptosis induced by K . pneumoniae [35] . Our human genetic study provides adjunctive evidence for the importance of NLRC4 in respiratory melioidosis although it requires validation . Few clinically associated polymorphisms in NLRC4 have been described thus far and the function of rs6757121 is otherwise unknown . We have previously reported the association of variation in TLR5 with survival in melioidosis regardless of site of infection and here show that the association holds in respiratory disease [9] , [10] . Unlike in mice , modeling suggests that co-inheritance of variants in NLRC4 and in TLR5 increases the effect in an additive manner . Another important difference between mice and humans is that in humans , blunting of TLR5 function – as found in carriers of a nonsense polymorphism – is in fact protective against death from melioidosis [9] , [10] . This seemingly opposite phenotype from mice underscores the challenges of mimicking human sepsis in mice [37]–[39] . In conclusion , we show that NLRC4 and TLR5 are essential elements of host defense in murine respiratory melioidosis , and that genetic variation in these genes is associated with outcome from human respiratory melioidosis . | Melioidosis is an infection caused by Burkholderia pseudomallei , a bacterium that is found in tropical soil and water . Melioidosis can present in a variety of ways , but lung involvement is common and usually severe . The host response to infection governs outcome . In this study , we examined the role of two host sensors of bacterial components–TLR5 and NLRC4–to determine their necessity in respiratory melioidosis . Although both proteins are involved in detection of bacterial flagellin , in mice we defined specific and individual roles for TLR5 and NLRC4 in protecting against death from melioidosis . In humans with melioidosis involving the lung , genetic variation in these receptors also had independent associations with survival . These results underscore the importance of these elements of host defense in respiratory melioidosis and support further studies of the underlying mechanisms . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"bacterial",
"diseases",
"infectious",
"diseases",
"medicine",
"and",
"health",
"sciences",
"genetics",
"of",
"the",
"immune",
"system",
"clinical",
"immunology",
"melioidosis",
"biology",
"and",
"life",
"sciences",
"immunology",
"bacterial",
"pneumonia",
"immune",
"r... | 2014 | NLRC4 and TLR5 Each Contribute to Host Defense in Respiratory Melioidosis |
We discovered a novel interaction between phage P22 and its host Salmonella Typhimurium LT2 that is characterized by a phage mediated and targeted derepression of the host dgo operon . Upon further investigation , this interaction was found to be instigated by an ORFan gene ( designated pid for phage P22 encoded instigator of dgo expression ) located on a previously unannotated moron locus in the late region of the P22 genome , and encoding an 86 amino acid protein of 9 . 3 kDa . Surprisingly , the Pid/dgo interaction was not observed during strict lytic or lysogenic proliferation of P22 , and expression of pid was instead found to arise in cells that upon infection stably maintained an unintegrated phage chromosome that segregated asymmetrically upon subsequent cell divisions . Interestingly , among the emerging siblings , the feature of pid expression remained tightly linked to the cell inheriting this phage carrier state and became quenched in the other . As such , this study is the first to reveal molecular and genetic markers authenticating pseudolysogenic development , thereby exposing a novel mechanism , timing , and populational distribution in the realm of phage–host interactions .
Due to billions of years of co-evolution and their overpowering abundance in the biosphere , viruses of bacteria ( i . e . bacteriophages or phages ) have a profound impact on the conduct and ecology of their hosts [1] , [2] . Lytic proliferation of phages for example can affect host mutation rates [3] , structure microbial consortia [4] , and contribute significantly to the global biogeochemical carbon flux [5] . Lysogenic proliferation as stable prophages , on the other hand , increases the genetic repertoire and genome plasticity of the host , thereby often extending its adaptive potential in terms of virulence and ecological fitness [6] . While the basic molecular events and genetic circuitry behind lytic and lysogenic development have traditionally received a lot of attention and are reasonably well understood for a number of model phages [7]–[9] , the increasing wealth of novel phage genes with no known homologs and function nevertheless suggests an unforeseen intricacy in phage – host interactions [1] , [10] . Furthermore , in many ecological niches phage – host associations often appear to defy the classical bifurcation into strict lytic or lysogenic development , as a large number of reports indicate a lysogeny-independent but stable co-existence between phages and their hosts . These phenomena are often vaguely referred to as pseudolysogeny , and hypothesize the existence of stable “phage carrier” cells in which the incoming phage has temporarily refrained from lytic or lysogenic development [11] . This suspended state is believed to play an important role in the long term survival strategy of viruses , as it might ( i ) prevent poor replication or even degradation of the phage chromosome in a host that is too starved to support further steps in lytic or lysogenic development , and/or ( ii ) provide a transient intracellular refuge for the phage chromosome in environments characterized by low host densities and short capsid half-lives [12] , [13] . Despite its ecological importance [11] , [14] , however , no formal molecular evidence currently exists for the presence of such a state , let alone its possible impact on the physiology of the cell . In this study , we extend on the intricacy of phage – host interactions and provide both genetic and direct cell biological evidence for the existence of a dedicated pseudolysogenic state in the Salmonella Typhimurium – phage P22 model system .
During routine screening of a MudK based lacZ promoter-trap library in Salmonella Typhimurium LT2 on LB X-Gal agar plates , our attention was drawn to a colony displaying an inhomogeneous distribution of LacZ activity ( i . e . blue coloration; Figure 1A ) that was neither symmetrical , nor sectorial . Moreover , after streaking out on new LB X-Gal plates , this particular clone segregated both into plain white colonies and colonies with an irregular blue coloration similar to that of the parent colony ( Figure 1B ) . Interestingly , however , when the latter colonies were replica-plated on green indicator agar , the blue patches on LB X-Gal agar overlapped perfectly with the dark green sites of cell lysis that were revealed by the green indicator agar ( compare Figure 1B and 1C ) . As we reasonably assumed this cell lysis to stem from infection by residual P22 HT105/1 int201 transducing phage that was initially used to deliver the MudK element during construction of the library , we hypothesized LacZ activity of the isolated clone to be triggered by exposure to phage P22 . In order to further examine this phenotype , the MudK insertion of the corresponding clone was transduced into a fresh LT2 strain and a phage-free transductant ( designated LT2K7 ) was streaked across wild-type P22 ( P22 wt ) on LB X-Gal agar ( Figure 1D ) . As a result , we found LT2K7 to turn from white to blue upon encountering P22 , suggesting that phage P22 is causally involved in triggering lacZ expression in LT2K7 . The MudK insertion site of LT2K7 was mapped to the dgoRKAT operon . DNA sequence analysis revealed that the MudK insertion resulted in a translational fusion of the lacZ reporter gene to dgoT ( Figure 2A ) . The dgoR gene located at the beginning of the operon is predicted to function as an autorepressor [15] and indeed LT2K7 ΔdgoR constitutively expressed the dgoT::MudK fusion ( Figure 2B ) , regardless of infection by P22 wt . Furthermore , increasing the level of DgoR by providing the corresponding gene on a multicopy plasmid ( pFPV-dgoR ) was able to abolish induction of dgoT::MudK by P22 wt , but had no obvious effect on phage infection per se ( Figure 2C ) . These data suggest that infection by P22 interferes with autorepression of the dgo operon in LT2 . Subsequently , we noticed that derepression of dgoT::MudK upon phage infection was a feature supported by P22 , but not by another S . enterica specific temperate phage such as ES18 ( Figure S1 ) . This raised the possibility that induction of the dgo operon stemmed from a genetic circuit in P22 , rather than from a generic host response to phage infection . To examine this , a plasmid library of random P22 genomic fragments was screened for loci able to render the LT2K7 indicator strain blue on LB X-Gal . As such , a 521 bp P22 fragment could eventually be obtained that triggered dgoT::MudK upon conditional expression ( using the arabinose inducible PBAD promoter ) in LT2K7 . More specifically , this fragment was found to correspond to a small and unannotated locus situated between orf25 and orf80 in the late region of the P22 genome [16] , [17] ( Figure 3A , boxed region ) , and was subsequently designated as pid ( for phage P22 encoded instigator of dgo expression ) . Interestingly , close inspection of the pid region revealed it to be a genuine moron locus [6] , as it is integrated at a site where related phages have either no ( cfr . PS34 in Figure 3B ) or another insert ( Figure 3B ) . In addition , the pid locus is further characterized ( i ) by the fact that it is divergently transcribed relative to its surrounding genes , indicating that its regulatory control might deviate from that of the late region , and ( ii ) by a 3′ Rho-independent transcriptional termination site . During our efforts to discriminate whether the pid locus encoded a small regulatory RNA or a small protein , we discovered the appearance of a distinct low molecular weight protein band on SDS-PAGE upon triggering transcription of the locus from a plasmid ( pFPV-PBAD-pid ) ( Figure 4A ) . Moreover , sequencing of this protein indeed revealed peptide signatures encoded by one of the possible reading frames of the moron locus ( Figure 4B ) . While the stop codon of this open reading frame could be inferred , the start codon was predicted by the presence of an upstream canonical Shine-Dalgarno sequence ( AAGGAG ) [18] ( Figure 4C ) . Importantly , introduction of a −1 frame shift in the start codon ( Figure 4C ) simultaneously abolished both expression of the characteristic protein band and induction of dgoT::MudK in LT2K7 ( Figure 4D ) , establishing this 86 amino acid and 9 . 23 kDa protein ( termed Pid; Figure 4B ) ( and not a small RNA species putatively originating from the same locus ) as the actual trigger of the interaction . It should be noted that subsequent deletion of the pid open reading frame in P22 correspondingly abolished induction of the dgo operon upon infection ( Figure 4E ) , but had no noticeable impact on the ability to develop lytically or lysogenically . Since upon infection the propagation of P22 wt can either proceed lytically or lysogenically , we wondered which of these two distinct developmental routes would actually mount the Pid/dgo interaction ( Figure 5A ) in the cell . Surprisingly , however , dgoT::MudK expression was completely absent both when LT2K7 was subjected to obligate lytic infection with P22 c2 ( Figure 5C ) or when the reporter strain carried P22 wt as a prophage ( Figure 5D ) . The latter finding is in fact consistent with our initial observation of the Pid/dgo interaction being fully supported by the P22 HT105/1 int-201 transducing phage ( Figure 1A , 1B and Figure 5B ) despite its inability to integrate in the host chromosome as a prophage . To further corroborate this finding , we extended the P22 pid open reading frame with a strep-tag encoding sequence ( leading to P22 pid-strep ) to facilitate Pid detection by western blot , and checked whether the observed absence of dgoT::MudK expression also correlated with attenuated levels of Pid . In agreement with the results above ( Figure 5 ) , Pid production was abundant in LT2 infected with P22 pid-strep ( Figure 6A ) , while it was severely attenuated in LT2 infected with the obligate lytic P22 c2 pid-strep derivative ( Figure 6B ) and completely absent in LT2 carrying P22 pid-strep as a prophage ( Figure 6C ) . To determine whether or not compromised Pid production stemmed from attenuated pid transcription , the pid open reading frame of P22 was replaced with the yfp fluorescent reporter gene , and the resulting phage ( i . e . P22 Δpid::yfp , carrying yfp under the control of the native pid promoter ) was used to interact with LT2 . In agreement with our previous findings ( Figure 5 and Figure 6 ) , cells infected with an obligate lytic derivative of P22 Δpid::yfp ( i . e . P22 c2 Δpid::yfp ) only displayed very faint fluorescence in the few minutes before cell lysis ( Figure 7C ) , while cells carrying P22 Δpid::yfp as a prophage displayed no detectable fluorescence ( Figure 7D ) . On the contrary , cells infected with P22 Δpid::yfp ( Figure 7A ) or its int derivative ( i . e . P22 Δint Δpid::yfp ) ( Figure 7B ) clearly showed a plethora of cells exhibiting YFP expression to different extents . Interestingly , the finding that expression of pid and subsequent derepression of the dgo operon are not supported during lytic or lysogenic propagation of P22 strongly suggests that the Pid/dgo interaction might be dedicated to a different state of P22 development . Spurred by the above observations , time-lapse fluorescence microscopy was used to more closely examine the timing and dynamics of pid expression during infection of LT2 with P22 Δpid::yfp at single cell resolution . While this approach demonstrated that the pid locus indeed became expressed in lineages emerging from non-lytic infection with the reporter phage , it also revealed that this expression was a feature that subsequently segregated asymmetrically between siblings ( Figure 8 ) . Surprisingly , in fact , only one individual within the growing lineage consistently displayed the ability to express pid , thereby revealing an unprecedented timing and populational distribution of this phage – host interaction . It should also be noted that disruption of the int gene in P22 Δpid::yfp did not affect the timing nor the asymmetric distribution of pid expression ( Figure S2 ) , corroborating that the actual chromosomal integration event leading to the establishment of a prophage was not required for this phenomenon . In order to more closely examine the possible role of the P22 chromosome in this peculiar asymmetric segregation phenotype , P22 Δpid::yfp was equipped with a parS site ( resulting in P22 Δpid::yfp parS ) , allowing its whereabouts during infection to become fluorescently tractable in an LT2 strain expressing the ParB protein fused to mCherry ( i . e . LT2 pCW-mCherry-parB ) . Interestingly , soon after infection of LT2 pCW-mCherry-parB with P22 Δpid::yfp parS , a single and coherent mCherry cloud appeared in cells destined for non-lytic infection ( Figure 9A and 9B ) , indicative for the presence of one ( or possibly more ) P22 chromosome ( s ) . Furthermore , upon subsequent cell divisions , this cloud became asymmetrically segregated between siblings , with pid expression remaining tightly linked to the cell inheriting and carrying the unintegrated P22 chromosome ( s ) ( Figure 9 ) . The gradual dilution of YFP molecules in siblings not inheriting this phage carrier state is consistent with the heterogeneity in YFP fluorescence in liquid cultures of LT2 infected with P22 Δpid::yfp observed earlier ( cfr . Figure 7 ) .
Given the penetration and importance of bacteriophages in global ecology , understanding their possible associations with a host is of tremendous importance . In this report , the S . Typhimurium – phage P22 model system yielded both molecular and genetic evidence authenticating the existence of a dedicated phage carrier state in which an unintegrated phage chromosome is stably maintained in the cell and asymmetrically inherited by only one of the siblings upon further divisions . This behavior differs fundamentally from cells undergoing lytic or lysogenic phage development , which are forced either to lyse after the production of new virions or to symmetrically segregate the prophage chromosome ( integrated in the host chromosome or existing as a stable episome ) among siblings [19] , [20] , respectively . The phage carrier ( or pseudolysogenic ) state is believed to have a tremendous impact on phage ecology , as the ability to postpone the commitment to lytic or lysogenic development might improve phage survival in inhospitable environments [11]–[14] . Specifically with regard to the biology of phage P22 , our findings at the single cell level are in remarkable agreement with very early observations made by Zinder , who anticipated that upon infection P22 could be maintained in a pseudolysogenic form during several generations before integrating itself as a prophage [21] . Despite the long-standing assumption of its alleged existence and its ecological importance , however , the phage carrier state has so far hardly been documented from a molecular or genetic point of view . In fact , although it has been proposed that the phage remains idle or inert while being in this state [12] , [13] , our results on the contrary provide the first evidence that a dedicated phage – host interaction ( as exemplified by Pid/dgo ) can be mounted in phage carrier cells . Clearly , the existence of dedicated genetic programs that are executed solely in phage carrier cells substantiates their biological significance and allows them to differentiate from uninfected cells or cells destined for lytic or lysogenic development . On itself , the induction of the LT2 dgo operon by the P22 Pid ORFan protein is also peculiar , since only a very limited number of phage – host interactions have so far been discovered in which the phage deliberately and specifically interferes with host gene expression . Indeed , in currently recognized interactions , phage encoded functions either ( i ) hijack cellular machinery and generally shut down host gene expression to support phage reproduction during lytic proliferation [22] , or ( ii ) contribute virulence factors that support the pathogenicity of the host during lysogenic development [6] , [23] . A notable exception was only recently described for λ lysogens of E . coli , in which the λ CI repressor was shown to compromise cellular gluconeogenesis by physically obstructing the host pckA promoter [24] . Interestingly , the dgo operon encodes proteins involved in the uptake and metabolism of D-galactonate , which is considered to be an important source of carbon and energy during intracellular survival and proliferation of Salmonella spp . [25] . Moreover , a dgoT knock-out was correspondingly found to attenuate the virulence of S . enterica serovar Choleraesuis in pigs [26] . It remains to be established , however , how exactly the Pid/dgo interaction is mounted within the carrier state , and whether it would endow carrier cells with increased virulence or rather constitutes a way for the phage to decide on how long to maintain this state . In summary , our results authenticate the existence of the phage carrier state as a distinct developmental route in phage biology that differs from strict lytic or lysogenic propagation . The phenotypic consequences of the interactions taking place in phage carrier cells are likely to provide the missing link in the proper and accurate interpretation of phage – host dynamics occurring throughout microbial ecosystems .
Bacterial strains , phages and plasmids used throughout this study are listed in Table 1 . For culturing bacteria , Lysogeny Broth ( LB; [27] ) medium was used either as a broth or as agar plates after the addition of 15% ( for spreading plates ) or 7% ( for soft-agar plates ) agar . Cultures were grown in LB broth for 15–20 h at 37°C under well-aerated conditions ( 200 rpm on a rotary shaker ) to reach stationary phase . Exponential phase cultures were in turn prepared by diluting stationary phase cultures 1/100 or 1/1000 in fresh pre-warmed broth , and allowing further incubation at 37°C . When appropriate , the following chemicals ( Applichem , Darmstadt , Germany ) were added to the growth medium at the indicated final concentrations: ampicillin ( 100 µg/ml; Ap100 ) , chloramphenicol ( 30 µg/ml; Cm30 ) , kanamycin ( 50 µg/ml; Km50 ) , tetracycline ( 20 µg/ml; Tc20 ) , glucose ( 0 . 02% ) , L-arabinose ( 0 . 02% ) , and 5-bromo-4-chloro-3-indolyl-β-D-galactopyranoside ( X-Gal; 40 µg/ml ) . Phages were propagated on S . Typhimurium LT2 as plaques in LB soft-agar or as lysates in LB broth as described previously [28] . Phage stocks were filter sterilized with 0 . 2 µm filters ( Fisher Scientific , Aalst , Belgium ) and chloroform was added to maintain sterility . Generalized transduction was performed with phage P22 HT105/1 int-201 as described previously [28] , [29] . This mutant is unable to integrate into the host chromosome as a prophage due to the lack of integrase ( Int ) activity . To discriminate phage infected from uninfected colonies , plates containing green indicator ( GI; [28] ) agar were used to indicate cell lysis . The latter medium contains glucose as a carbon source , and a pH indicator dye that turns dark green at sites where phage infection causes cell lysis and the concomitant release of organic acids . Please note that for clear visualization of spotted or ( cross- ) streaked bacterial and/or phage populations , agar plates were printed to Whatman filter papers ( GE Healthcare , Diegem , Belgium ) before photographing . Expression of β-galactosidase ( LacZ ) was inferred from the hydrolysis of either 5-bromo-4-chloro-3-indolyl-β-galactopyranoside ( X-Gal ) or o-nitrophenyl-β-D-galactoside ( ONPG ) . X-Gal was typically added to agar plates ( 40 µg/ml ) , where its hydrolysis by β-galactosidase yielded an insoluble blue precipitate . For quantitative measurements of lacZ expression , Miller units were determined as described previously [30] using the CHCl3-sodium dodecyl sulfate permeabilization procedure . Particles of phage P22 were purified by passing a lysate through a 0 . 45 µm pore-size filter , after which particles were concentrated by centrifugation ( 4 , 000×g , 20 min ) in the presence of polyethylene glycol ( PEG ) 8 , 000 ( 8% , w/v ) and 1 M NaCl . Subsequently , further purification was attained by ultracentrifugation ( 140 , 000×g , 3 hours ) using a layered CsCl step gradient of 1 . 33 , 1 . 45 , 1 . 50 and 1 . 70 g/ml . This resulted in a distinct blue band containing the concentrated P22 particles . This band was subsequently collected and dialysed against phage buffer ( 10 mM Tris-HCl pH 7 , 10 mM MgSO4 , 150 mM NaCl ) three times using a Slide-A-Lyzer dialysis cassette ( Pierce , Rockford , IL , USA ) . For DNA extraction , the purified and dialysed phage particles were incubated at 56°C for 1 h in the presence of 0 . 5% SDS ( w/v ) , 20 mM EDTA and 2 µg/ml proteinase K . Subsequently , DNA was extracted and purified from this mixture by phenol/chloroform [27] and precipitated with Na-acetate/ethanol . Finally , the sample was treated with RNase A ( 0 . 1 mg/ml ) ( Fermentas , St . Leon-Rot , Germany ) for 1 h at room temperature to remove any residual RNA . Next , the resulting purified P22 genomic DNA was partially digested with the blunt 4 bp-cutter BsuRI restriction enzyme ( Fermentas ) and separated by agarose gel electophoresis ( 1% agarose ) , after which fragments between 1–2 kb were isolated from the gel using the GeneJET Gel Extraction Kit ( Fermentas ) . Parallel to this , pFPV-PBAD–gfp ( pAA100; [31] ) was digested with XbaI and HindIII ( Fermentas ) to remove gfp , and treated with calf intestinal alkaline phosphatase ( Fermentas ) to prevent self-ligation . The genomic P22 DNA fragments and the cut pFPV-PBAD vector were subsequently ligated after blunting with T4 ligase and Klenow polymerase ( Fermentas ) , and transformed by electroporation into LT2K7 . After plating on LB Ap100 , this random P22 shotgun library was replica-plated on LB Ap100 X-Gal with and without 0 . 02% arabinose to screen for plasmids able to trigger LacZ expression in LT2K7 . Phages containing homologous regions to the region surrounding pid were selected with nucleotide Blast [32] . Whole genome alignment was performed manually and was based on the Blast-search results . The resulting conclusions were later confirmed by a progressive Mauve alignment [33] on the full genomes using default settings . Strain LT2K7 stems from a random MudK library , generated as described previously [34] , and harbors a translational lacZ fusion to the LT2 dgoT gene ( i . e . dgoT::MudK ) . In strain LT2 ΔdgoR , the dgoR gene was deleted via recombineering [35] , using an amplicon ( Phusion DNA polymerase; Fermentas ) prepared on pKD3 [35] with the primers dgoR_pkd3_Fw and dgoR_pkd3_Rev ( Table 2 ) . The cat cassette replacing dgoR was flipped out using pCP20-borne Flp to recombine the two frt-sites [36] , resulting in a small frt-scar followed by a new ribosome binding site [35] . Strain LT2K7 ΔdgoR was subsequently constructed by transducing dgoT::MudK to LT2 ΔdgoR . For the construction of P22 Δpid::yfp , the yfp-frt-cat-frt cassette was PCR amplified ( Phusion DNA polumerase; Fermentas ) from plasmid pAc [37] with primers pid_YFP_Cm_Fw and pid_YFP_Cm_Rev ( Table 2 ) , and used to replace the pid gene in LT2 lysogenized with wild-type P22 via recombineering [35] . Subsequently , the cat cassette was flipped out using pCP20-borne Flp to recombine the two frt-sites [36] , and the resulting P22 Δpid::yfp phage was isolated and purified from the corresponding lysogen . For the construction of P22 Δint Δpid::yfp , the integrase gene ( int ) in LT2 lysogenized with P22 Δpid::yfp was deleted by recombineering , using a PCR amplicon prepared on pKD3 with primers P22_Int_Fw and P22_Int_Rev ( Table 2 ) [35] . Please note that the frt-flanked cat cassette was not removed by site specific Flp recombination , since this would interfere with the frt-scar already present in the pid locus . The resulting P22 Δint Δpid::yfp phage could be released by amplifying rare excision events through growth on wild-type LT2 in order to allow detection and purification of plaques . Please note that these phages produced normal turbid plaques and were unable to from true lysogens on LT2 . For the construction of P22 Δpid::yfp parS , the parS-frt-cat-frt cassette was PCR amplified from pGBKD3-parS [38] with primers P22_parS_Fw and P22_parS_Rev ( Table 2 ) , and inserted between the gtrC and 9 genes in LT2 lysogenized with P22 Δpid::yfp via recombineering [35] . Please note that the frt-flanked cat cassette was not removed by site specific Flp recombination , since this would interfere with the frt-scar already present in the pid locus . For the construction of P22 pid-strep , a strep-tag encoding sequence ( strep ) was added to the 3′ end of the pid open reading frame by recombineering a strep-frt-cat-frt amplicon prepared on pKD3 [35] with primers pid_Strep_Fw and pid_Strep_Rev ( Table 2 ) in LT2 lysogenized with wild-type P22 . Subsequently , the cat cassette was flipped out using pCP20-borne Flp to recombine the two frt-sites [36] , and the resulting P22 pid-strep phage was isolated and purified from the corresponding lysogen . Please note that the C-terminal addition of the strep-tag to Pid was shown to have no effect on the ability of Pid to trigger the dgo-operon . Finally , clear mutants P22 c2 Δpid::yfp and P22 c2 pid-strep were constructed by oligo-mediated mutagenesis [39] of the corresponding P22 Δpid::yfp and P22 pid-strep lysogens in LT2 , using olignucleotide Oligo_C2_Stop ( Table 2 ) . This oligo introduced two flanking stop-codons after the first 11 amino acids of the P22 C2 repressor . After recombination , transformants were inoculated in LB with wild type LT2 and grown overnight at 37°C to amplify the corresponding clear mutants . Afterwards , P22 c2 Δpid::yfp and P22 c2 pid-strep were isolated by plaquing on LT2 , and the c2 mutation was verified by sequencing . Plasmid pFPV-PBAD-pid was constructed by digesting pFPV-PBAD–gfp with XbaI and HindIII ( Fermentas ) , and subsequently replacing gfp with pid . The latter amplicon ( Phusion DNA polymerase; Fermentas ) was obtained using primers pid_Fw and pid_Rev ( Table 2 ) , digested with XbaI and HindIII prior to ligation . Plasmid pFPV-dgoR expresses the LT2 dgoR gene under the control of its own promoter , and was constructed by ligating an XbaI digested PCR amplicon of the LT2 dgoR locus , obtained with primers dgoR_Fw and dgoR_Rev , into the XbaI site of pFPV25 [40] . Plasmid pCW-mCherry-parB was constructed by first making an amplicon of the pALA2705 vector [41] with primers pALA_Out_Left and pALA_Out_Right . These primers amplify the entire plasmid except its gfp gene , and added an EcoRI and a SacI restriction site at the end of the amplicon . Subsequently , the mCherry gene was amplified from pRSet-B-mCherry ( kind gift from Roger Tsien , University of California , USA ) with primers Mcherry_Fw and Mcherry_Rev ( Table 2 ) , and both amplicons were digested with EcoRI and SacI prior to being ligated to each other . The resulting plasmid , pCW-mCherry-parB , expresses an N-terminal fusion of mCherry to ParB under control of an IPTG inducible promoter . Please note , however , that leaky expression of the latter promoter in the absence of IPTG was already sufficient , as mentioned previously [42] . For site-directed mutagenesis , the “Phusion Site-Directed Mutagenesis Kit” protocol ( Thermo Scientific , Epsom , United Kingdom ) was followed . As such , plasmid pFPV-PBAD-pid was used as a template for amplification with primers pid_FS_Fw and pid_FS_Rev ( Table 2 ) for constructing a frame shift mutation in the actual pid start codon . After phosphorylating the 5′ ends of the primers according to the manufacturer's instruction , the primer pair was used to PCR amplify pFPV-PBAD-pid ( Phusion polymerase; Fermentas ) . The resulting linear fragment was purified from an agarose gel using the GeneJET Gel Extraction Kit ( Fermentas ) , subsequently self ligated , and finally transformed by electroporation to E . coli DH5α . The resulting pFPV-PBAD-pidFS plasmid from transformants selected on LB Ap100 was further confirmed by sequencing , prior to transformation to LT2 and LT2K7 . Samples were lysed in standard lysis buffer containing 50 µl/ml Bugbuster ( Novagen , Darmstadt , Germany ) . Total protein concentration was assessed by the BCA protein assay kit ( Novagen ) and SDS-PAGE was performed as described previously by Sambrook and Russel [27] ) . Finally , gels were stained with coomassie [27] and when necessary , silver staining was employed as previously described [43] . For protein identification , the corresponding protein band was excised and trypsin-digested according to the method described earlier [44] . Subsequently , the digested peptides were identified by LC–ESI MS/MS ( Thermo Electron , San Jose , CA ) and further analyzed using Mascot ( Matrix Sciences , London , UK ) against the NCBI database ( http://www . ncbi . nlm . nih . gov/ ) . For western-blotting , equal amounts of proteins were separated with PAGE and transferred to a nitrocellulose membrane ( Hybond-C Extra; GE Healthcare ) by semi-dry electroblotting for 1 hour at 0 . 15 A using a Trans-Blot SD Semi-Dry Electrophoretic Transfer Cell ( Bio-Rad Laboratories ) and transfer buffer ( 50 mM Tris; 40 mM glycine; 0 . 075% SDS; 20% Methanol ) . Strep-tagged Pid was subsequently detected by StrepMAB-Classic , an anti-strep monoclonal antibody conjugated with Horse radish peroxidase ( IBA , Göttingen , Germany ) . Horse radish peroxidase activity was assessed with Pierce ECL Western Blotting Substrate ( Thermo Scientific ) , and detected on photo-sensitive film ( Hyperfilm ECL; GE Healthcare ) . The strep-tagged protein ladder ( IBA ) was used as a molecular ruler and positive control of the blotting process . Fluorescence microscopy and time-lapse fluorescence microscopy were performed with a temperature controlled ( Okolab Ottaviano , Italy ) Ti-Eclipse inverted microscope ( Nikon , Champigny-sur-Marne , France ) equiped with a TI-CT-E motorised condensor , a YFP filter ( Ex 500/24 , DM 520 , Em 542/27 ) , an mCherry filter ( Ex 562/40 , Dm 593 , Em 641/75 ) , and a CoolSnap HQ2 FireWire CCD-camera . For imaging , cells were placed between LB agar pads and a cover glass , essentially as described previously [45] , and incubated at 37°C . Please note that for experiments involving LT2 pCW-mCherry-parB , cells were grown on agar pads of AB-minimal media supplemented with 0 . 02% D-glucose , 100 µg/ml Uracil and 100 µg/ml Thiamine , Ap100 , and incubated at 30°C , as described previously [42] . Images were acquired using NIS-Elements ( Nikon ) and resulting pictures were further handled with open source software ImageJ ( Downloaded from http://rsbweb . nih . gov/ij/ ) . | Viruses of bacteria , also referred to as ( bacterio ) phages , are the most abundant biological entity on earth and have a tremendous impact on the ecology of their hosts . It has traditionally been recognized that upon infection by a temperate phage the host cell is forced either to produce and release new virions during lytic development or to replicate and segregate the phage chromosome together with its own genetic material during lysogenic development . These developmental paths are orchestrated by a dedicated set of phage–host interactions that are able to sense and redirect host cell physiology . In addition to this classical bifurcation of temperate phage development , many studies on phage biology in natural ecosystems hypothesize the existence and significance of stable phage carrier cells that are not engaged in either lytic or lysogenic proliferation . Using Salmonella Typhimurium and phage P22 as a model system , we provide substantial evidence authenticating the existence of the phage carrier state and demonstrate that this state ( i ) is asymmetrically inherited among carrier cell siblings and ( ii ) enables the execution of a novel phage–host interaction that is not encountered during lytic or lysogenic proliferation . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"bacteriology",
"microbial",
"metabolism",
"gene",
"regulation",
"microbiology",
"gene",
"function",
"molecular",
"genetics",
"bacterial",
"pathogens",
"microbial",
"physiology",
"gene",
"expression",
"microbial",
"pathogens",
"biology",
"salmonella",
"microbial",
"ecology"... | 2013 | Expression of a Novel P22 ORFan Gene Reveals the Phage Carrier State in Salmonella Typhimurium |
A lack of access to sanitation is an important risk factor child health , facilitating fecal-oral transmission of pathogens including soil-transmitted helminthes and various causes of diarrheal disease . We conducted a meta-analysis of cross-sectional surveys to determine the impact that community-level sanitation access has on child health for children with and without household sanitation access . Using 301 two-stage demographic health surveys and multiple indicator cluster surveys conducted between 1990 and 2015 we calculated the sanitation access in the community as the proportion of households in the sampled cluster that had household access to any type of sanitation facility . We then conducted exact matching of children based on various predictors of living in a community with high access to sanitation . Using logistic regression with the matched group as a random intercept we examined the association between the child health outcomes of stunted growth , any anemia , moderate or severe anemia , and diarrhea in the previous two weeks and the exposure of living in a community with varying degrees of community-level sanitation access . For children with household-level sanitation access , living in a community with 100% sanitation access was associated with lowered odds of stunting ( adjusted odds ratio [AOR] = 0 . 97 , 95%; confidence interval ( CI ) = 0 . 94–1 . 00; n = 14 , 153 matched groups , 1 , 175 , 167 children ) , any anemia ( AOR = 0 . 73; 95% CI = 0 . 67–0 . 78; n = 5 , 319 matched groups , 299 , 033 children ) , moderate or severe anemia ( AOR = 0 . 72 , 95% CI = 0 . 68–0 . 77; n = 5 , 319 matched groups , 299 , 033 children ) and diarrhea ( AOR = 0 . 94; 95% CI = 0 . 91–0 . 97 ) ; n = 16 , 379 matched groups , 1 , 603 , 731 children ) compared to living in a community with < 30% sanitation access . For children without household-level sanitation access , living in communities with 0% sanitation access was associated with higher odds of stunting ( AOR = 1 . 04 , 95% CI = 1 . 02–1 . 06; n = 14 , 153 matched groups , 1 , 175 , 167 children ) , any anemia ( AOR = 1 . 05 , 95% CI = 1 . 00–1 . 09; n = 5 , 319 matched groups , 299 , 033 children ) , moderate or severe anemia ( AOR = 1 . 04 , 95% CI = 1 . 00–1 . 09; n = 5 , 319 matched groups , 299 , 033 children ) but not diarrhea ( AOR = 1 . 00 , 95% CI = 0 . 98–1 . 02; n = 16 , 379 matched groups , 1 , 603 , 731 children ) compared to children without household-level sanitation access living in communities with 1–30% sanitation access . Community-level sanitation access is associated with improved child health outcomes independent of household-level sanitation access . The proportion of children living in communities with 100% sanitation access throughout the world is appallingly low . Ensuring sanitation access to all by 2030 will greatly improve child health .
An estimated 1 billion people live without access to any type of sanitation facility , i . e . a toilet or latrine [1] . This lack of sanitation access fails to contain human feces , which are responsible for transmission of various diarrheal diseases as well as soil-transmitted helminthes ( STH ) primarily through the fecal-oral route where fecal matter is ingested via water , dirt or food [2] . Diarrheal diseases kill millions of children each year [3] , and for those who survive present the problem of malnutrition and developmental delays [4] . STH cause malnutrition and stunting in addition to developmental delays [5] . Furthermore hookworm ( Necator americanus or Ancylostoma duodenale ) are known risk factors for anemia [6] . Infections with Ascaris lumbricoides ( roundworm ) and Trichuris trichiura ( whipworm ) may also be risk factors for anemia although the evidence is inconclusive [7] . The prevalence of anemia is high in lower-income countries , estimated at 47% of children in 2005 [8] , though recent reports suggest the prevalence is decreasing [9] . Due to the importance of iron to various cellular functions including immune system functionality [10 , 11] , iron deficiency anemia is implicated as a cause of mortality for millions of children under five years of age each year [12 , 13] . Beyond a cause of mortality , anemia also decreases cognitive function [14–16] , and energy levels which leads to decreased productivity and economic well-being [17 , 18] . For subsistence farmers in lower-income countries decreased productivity can in turn lead to low crop yields and food insecurity , perpetuating a vicious cycle of malnutrition . Through containment and disposal of human feces , individual-level access to sanitation is known to decrease both diarrheal disease and STH infection [19–23] . A previous examination of survey data 1986–2007 found decreased risk of child mortality , diarrhea and stunting for children living in households with access to improved sanitation [24] . However , limiting sanitation to a household-level risk factor while ignoring the community-effect may greatly underestimate the impact that sanitation has on human health [25] . Poor sanitation in the community leads to increased exposure to fecal matter for all in that community , a significant risk factor for environmental enteropathy and subsequent child malnutrition [26] . Indeed , in India the behavior of open defecation was associated with reductions in child growth in an ecological analysis [27] , and in Cambodia community-sanitation behavior was associated with increased child growth more prominently than household-sanitation behavior [28] . Numerous community-randomized controlled trials of total sanitation campaigns have suggested that increasing access to sanitation can improve child health [28–31] , while others have found little to no effect of these interventions on child health [32–34] . Herein we present a study estimating the impact of community-level access to sanitation on child health as measured through child growth , anemia , and diarrhea symptoms using survey data compiled into an individual-level meta-analysis .
We sought to measure the impact that living in a community with 100% sanitation access has on the outcomes of child growth stunting among children aged 12–59 months , anemia among children under 5 years of age , and diarrhea in the previous two weeks from nationally-representative surveys . To do so we pooled surveys to create an individual-level meta-analysis [35] . We included multiple indicator cluster surveys ( MICS ) , demographic and health surveys ( DHS ) , malaria indicator surveys ( MIS ) , and AIDS indicator surveys ( AIS ) that were nationally-representative and publicly available as of July 2016 . As part of original survey protocol all data were anonymized prior to download from repositories to protect participant privacy . Anthropomorphic data are regularly collected in nationally-representative surveys . In these surveys height for age z-scores are computed for children under 5 years of age based upon World Health Organization growth reference standards . We classified children as stunted or not based upon the child’s height for age z-score being less than 2 standard deviations of the WHO growth reference standard . The outcome of stunting was available for 267 of 301 datasets . Nationally-representative surveys typically use the HemoCue system to measure hemoglobin levels for children age 5 and under and adjust these values for altitude . Depending upon the level of hemoglobin in the blood anemia is classified as none ( ≥12 . 0 g/dl ) , mild ( 10 . 0–11 . 9 g/dl ) , moderate ( 7 . 0–9 . 9 g/dl ) , and severe ( < 7 . 0 g/dl ) . We conducted analyses with two separate anemia outcomes , children with any anemia ( mild , moderate , or severe ) and children with moderate to severe anemia . The anemia outcomes were available for 104 of 301 datasets . Caregivers of children under five are also asked whether their child has had any commonly occurring illnesses such as fever , diarrhea , or cough . We classified children with diarrhea as those whose caregivers reported them having diarrhea in the previous 2 weeks , and children without diarrhea as those whose caregivers reported them not having diarrhea in the previous 2 weeks . The outcome of diarrhea was available for 281 of 301 datasets . In order to estimate the incremental effect of increasing community-level sanitation access on the outcomes of child growth stunting and anemia among children we classified children as living in households with any type of sanitation facility ( unimproved or improved ) , or not having any access to a sanitation facility . If households reported sharing a sanitation facility with others they were classified as having any type of sanitation facility . We defined community as the survey sampling area or cluster , and calculated the proportion of households having any sanitation facility ( unimproved or improved ) to serve as a measure of community-level sanitation access . We excluded datasets where > 95% of children live in communities with 100% sanitation access from any further analyses . Children in households with sanitation facilities or in communities with high sanitation access are likely to be predisposed to less risk of stunting and anemia , independent of sanitation access . To account for this selection bias and potential confounding we used two separate methodologies . First , we stratified our analyses by children in households with any sanitation access and children in households without any sanitation access . Second we used exact matching on community-level measures to circumvent the inherent selection bias of living in communities with more access to sanitation . Using the MatchIt package [36] in R version 3 . 2 . 3 [37] we matched children on numerous community-level and other covariates . To do so , we first took the cluster mean of child-level immunization coverage ( 3 doses of diphtheria , pertussis and tetanus ) . We then took the cluster mean of household wealth quintile and household access to a water source that was not considered surface water ( rivers , dams , ponds , lakes or unprotected springs ) . Once these cluster-level estimates were estimated we categorized estimates of cluster-level immunizations into tertiles , community-level wealth above and below the median , and community-level access to a non-surface water source above and below the median . In addition to the community-level measures we matched on household-level wealth ( dichotomized into rich or poor ) and mother’s education ( dichotomized into completed primary or not ) . The exact matching was conducted in accordance with the following equation: mijkl = β0 + βCi + χHj + δPk + ϕSl where mijk is a matched group for child i in household j in cluster k in survey l , Ci is an estimate of the mother’s education , Hj is an estimate of household wealth , Pk is a vector of cluster characteristics and Sl is a survey dummy . The matching procedures and all covariates were selected a priori . Pooling all datasets to create an individual-level meta-analysis we first examined the relationship between the outcomes and community-level sanitation access through a Lowess smoothing figure . To account for observable non-linearity in the exposure of interest we attempted to fit a cubic spline , however the spline was unable to account for the large decrease in the odds of the outcomes when going from 99% sanitation access to 100% sanitation access . We therefore categorized community-level sanitation access at 0% , 1–30% , 31–60% , 61–99% , and 100% to both align with the knots of the cubic spline ( 0 . 6 and 0 . 99 ) and to provide an appropriate comparison group ( 1–30% ) . Second , we calculated the unadjusted association between the exposures and outcomes of interest . For the unadjusted analysis we included the dataset and household sanitation access as covariates and adjusted the standard errors for correlated data at the survey cluster level . Finally , we used a generalized linear model with the matched group as a random intercept and a logit link to assess an adjusted association between the exposures and outcomes of interest . We included the following covariates to decrease the potential for confounding , with variable selection determined a priori: household sanitation access , urban or rural , child’s age in years , mother’s education ( quantified as none , some , and completed primary or higher ) , household wealth quintile , insecticide treated mosquito net ( ITN ) coverage ( no ITN in household , household owns ITN but child did not use previous night , and child used ITN previous night ) , child’s weight for height ( no wasting , 0–2 standard deviations below reference , >2 standard deviations below reference ) , child has a health or immunization card ( no , yes ) , child immunizations ( none , some , or all according to WHO standards ) , previous birth interval ( < 24 months or not ) , birth order ( firstborn , second born , third born , or later ) , mother’s age of the child in 5 year increments ( i . e . , 15–19 , 20–24 , etc . ) , household size in terms of number of people ( <6 , 6–15 , >15 ) , whether the household uses an open water source ( defined as a river , stream , pond , or unprotected spring ) , national gross domestic product retrieved from the World Bank database for the year of the survey as a continuous variable , and dataset as a dummy variable . The general model we use to assess the relationship between the outcomes and the exposure of interest is given by the following equations: yijklm|πijklm~Binomial ( 1 , πijklm ) logit ( πijklm ) =β1Sanj×β2Sank+χCijk+δHj+κSl+ζmζm~N ( 0 , ψ ) where πijklm is a dichotomous outcome for child i in household j in cluster k in survey l in matched group m , Sanj is whether the household has access to any sanitation or not , Sank is the level of sanitation access in the community , Cijk is a vector of child characteristics , Hj is a vector of household characteristics , Sl is a vector of survey characteristics and ζm is a random intercept for matched group m that is assumed to be normally distributed with a mean of zero . All analyses were conducted in Stata version 13 . 1 .
Among children living in households with sanitation access , living in a community with 100% sanitation access is associated with lower odds of stunting ( Table 2 ) . The lower odds of being stunted is only observed at 100% sanitation access; there was no effect of increasing community-level sanitation access for children in households with a sanitation facility located in clusters with < 100% sanitation access ( Fig 2 ) . Among children living in households without sanitation access , living in communities with zero sanitation access was associated with higher odds of stunting compared to children living in communities with 1–60% sanitation access ( Table 2 ) . Among children living in communities with high access to sanitation ( 60% or more of households with sanitation access ) not having household-level sanitation access was associated with higher odds of stunting compared to children living in communities with 1–60% sanitation access ( Fig 2; Table 2 ) . For the outcomes of any anemia as well as moderate or severe anemia , increasing community-level access to sanitation is associated with lower odds of anemia for children in households with sanitation access as well as children in households without sanitation access ( Fig 2; any anemia Table 3; moderate or severe anemia Table 4 ) . Increasing protection for all children occurred with increasing community-level sanitation access . For the outcome of diarrhea symptoms in the previous two weeks , increasing community-level access to sanitation is not associated with lower odds of diarrhea for children in households without access to sanitation ( Fig 2 , Table 5 ) . For children in households with access to sanitation living in a community with 100% sanitation access was associated with a lower odds of diarrhea ( Fig 2 , Table 5 ) . Children living in houses with access to a sanitation facility was associated with lower odds of stunting and any anemia at all levels of community-access to sanitation compared to children in houses with no access to a sanitation facility ( Table 6 ) . Living in a house with access to a sanitation facility was associated with lower odds of the outcomes of moderate or severe anemia and diarrhea compared to living in a house with no access to a sanitation facility only when community-sanitation access was higher .
We found that community-level access to sanitation is associated with lower odds of stunting and anemia for children independent of household-level sanitation access , and lower odds of diarrhea for children in houses with a sanitation facility . For children with sanitation access our analyses suggest that further gains in reducing the risk of stunting , anemia and diarrhea can be made as their communities move toward universal sanitation access . For children without household-level sanitation access our analyses suggest that community-level sanitation in addition to household-level sanitation is an important factor in child health . Unexpectedly for children without individual-level access to sanitation , living in a community with higher access to sanitation ( 60–99% ) was not beneficial compared to living in a community with no access to sanitation in terms of both stunting and diarrhea . ( It was beneficial for the outcome of anemia ) . We suspect that lacking a sanitation facility when the majority of neighbors have one is an indicator of vulnerability and for an outcome such as stunting with a multi-factorial causal etiology the vulnerability may represent a risk factor . In contrast for the outcome of anemia a significant benefit was observed for this particular population . Diarrhea in the previous two weeks was not associated with community-level access to sanitation , except for those children living in communities with 100% sanitation access . Also household-level access to sanitation was only associated with lower odds of diarrhea when community-level sanitation exceeded 60% . These findings that found improved sanitation at the household level to be associated with lowered risk of diarrhea [24] . Unmeasured confounding is a primary threat to these types of analyses , and the lack of impact may be due to unmeasured risk factors . Furthermore , there is great uncertainty around the validity of self-reported diarrhea in surveys [38] , and the subsequent misclassification error may lead to an underestimation of the impact of sanitation on diarrhea . Decreased fecal matter in the environment is likely to decrease circulation of diarrhea-causing agents , however there was no way to account for handwashing behavior in this analysis which is suggested to drive the relationship between diarrheal disease and sanitation [39 , 40] . The association between higher community-level sanitation access and the outcomes of anemia and stunting ( at lower levels of community-level access ) are consistent with the theory that environmental enteropathy is a significant risk factor for child malnutrition and health [26] . A recent modeling analysis and literature review suggests that community-level sanitation acts through a type of “herd-immunity” mechanism [41] , and an observational study demonstrated the protective nature of herd-immunity from sanitation in rural Ecuador [42] . These analyses confirm that a lack of sanitation at the community level poses a risk to members of that community , independent of household sanitation access and that the greatest gains occur as communities achieve universal access to sanitation . Our findings are in line with the scientific understanding of how fecal-oral transmission of various pathogens impact child health [26 , 41] . The measurement of sanitation access at the level of primary sampling unit of nationally-representative surveys is an innovation that improves upon previous analyses of survey data that only measure sanitation access at the district level [27] , or consider sanitation as a household-level risk factor [24] . Still , these findings should be treated cautiously for a number of reasons . First we greatly simplified sanitation access as having a sanitation facility or not . The sanitation ladder is much more nuanced [43] , with the greatest benefits to health coming from improving sanitation beyond a simple pit latrine . The simplification of sanitation access to having a facility or not allowed for its measurement at community level . Second , survey data are subject to error including recall and information error . Both outcomes included in this analysis were measured by survey personnel , and are not likely to be associated with the exposure of interest . However responses in survey questions about sanitation access may have suffered from social-desirability bias . Finally the use of the primary sampling unit as the community is not a perfect measure of community , given that primary sampling units may comprise various villages . Given the comprehensive nature of the datasets used and the random sampling of children selected we do not anticipate any publication or reporting bias to threaten the validity of these results . These results suggest that the greatest gains in health from sanitation are made when communities achieve universal access to sanitation . Until access to sanitation is universal within a population , even those with access carry risk derived from those without access to sanitation . Access to sanitation was included in the Millennium Development Goals as target 7 . C , with the goal of reducing by half the population without access to safe drinking water and basic sanitation . Progress was minimal; the target was missed by nearly 1 billion people [1] . These data show that poor community-level sanitation access is a significant risk factor for child growth stunting and anemia , both for children living in households with access to sanitation and for children living in households without . The number of children living in communities where any households lack sanitation access is alarmingly high throughout the world , and efforts must be made to achieve the Sustainable Development Goal of eliminating open defecation by 2030 . | A lack of access to a sanitation facility , i . e . a toilet and/or latrine , leads to numerous health challenges such as parasitic worms and environmental enteropathy . Parasitic worms are transmitted through human feces and cause multiple health complications in children including anemia and child growth stunting . Environmental enteropathy occurs with repeated and long-term inflammation of the small intestine which then reduces nutrient uptake and can cause child growth stunting , anemia and diarrhea . One-sixth of the world population has no access to any type of sanitation facility , and are therefore at higher risk of these challenges . Scientific literature on the impacts of sanitation typically examines household access to sanitation rather than community-level access to sanitation . We used national survey data to assess the impact that community-level access to sanitation has on child health , both for children with access to a sanitation facility and children without access to a sanitation facility . We found that a lack of sanitation access in the community is a significant risk factor for anemia and child growth stunting , but not for incidence of diarrhea . This risk decreases if a child has access to a sanitation facility , but even among those children with a sanitation facility poor sanitation access in the community is still a risk factor for anemia , child growth stunting and diarrhea . In addition to improving household access to adequate sanitation , community-level sanitation access needs to be addressed to improve child health . These results will add impetus to the Sustainable Development Goal to ensure sanitation access for all by 2030 . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"children",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"anemia",
"pediatrics",
"health",
"care",
"diarrhea",
"age",
"groups",
"research",
"design",
"child",
"growth",
"mathematics",
"signs",
"and",
"symptoms",
"statistics",
... | 2017 | An individual-level meta-analysis assessing the impact of community-level sanitation access on child stunting, anemia, and diarrhea: Evidence from DHS and MICS surveys |
The hydrolytic deamination of adenosine to inosine ( A-to-I editing ) in precursor mRNA induces variable gene products at the post-transcription level . How and to what extent A-to-I RNA editing diversifies transcriptome is not fully characterized in the evolution , and very little is known about the selective constraints that drive the evolution of RNA editing events . Here we present a study on A-to-I RNA editing , by generating a global profile of A-to-I editing for a phylogeny of seven Drosophila species , a model system spanning an evolutionary timeframe of approximately 45 million years . Of totally 9281 editing events identified , 5150 ( 55 . 5% ) are located in the coding sequences ( CDS ) of 2734 genes . Phylogenetic analysis places these genes into 1 , 526 homologous families , about 5% of total gene families in the fly lineages . Based on conservation of the editing sites , the editing events in CDS are categorized into three distinct types , representing events on singleton genes ( type I ) , and events not conserved ( type II ) or conserved ( type III ) within multi-gene families . While both type I and II events are subject to purifying selection , notably type III events are positively selected , and highly enriched in the components and functions of the nervous system . The tissue profiles are documented for three editing types , and their critical roles are further implicated by their shifting patterns during holometabolous development and in post-mating response . In conclusion , three A-to-I RNA editing types are found to have distinct evolutionary dynamics . It appears that nervous system functions are mainly tested to determine if an A-to-I editing is beneficial for an organism . The coding plasticity enabled by A-to-I editing creates a new class of binary variations , which is a superior alternative to maintain heterozygosity of expressed genes in a diploid mating system .
Since it was first discovered over 20 years ago [1] RNA editing has emerged as an important source of genetic coding variations in diverse life forms . One prominent mechanism for RNA editing is the deamination of adenosines in the precursor mRNA molecules , pertaining to most organisms in the metazoan lineage , including insects and mammals [2–4] . The deamination event , namely A-to-I editing , converts specific adenosines ( A ) to inosines ( I ) . Inosines are decoded as guanosines ( G ) in translation , thus resulting in codon changes that often lead to amino acid substitutions in the protein products . In addition to genetic recoding , A-to-I editing is also known to affect alternative splicing [5 , 6] , modify microRNAs , and alter microRNA target sites [5 , 7 , 8] . The major component of the A-to-I RNA editing machinery is the so called adenosine deaminases acting on RNA ( ADAR ) family of enzymes , which act on double stranded RNA structures ( dsRNAs ) within the substrate molecules [3 , 4 , 9] . Details about substrate targeting and regulation of editing activities are sparse; however , evidence indicates that A-to-I editing was cotranscriptional [10] , and the ADAR targeting sites were delineated to prefer certain non-random sequence patterns [11 , 12] , and depended in large part on the tertiary structure of RNA duplexes [4 , 13 , 14] . Genetic variability generated by A-to-I RNA editing expands the diversity and complexity of transcriptome , which serves as an important mechanism helping support critical biological functions . Lacking A-to-I RNA editing due to ADAR mutation in animal models resulted in embryonic or postnatal lethality in mice [15 , 16] , or displaying neurological defects in flies [17 , 18] . Many A-to-I editing targeted genes were documented in previous studies in human , mice , rhesus , and fly [19–22] . Reported cases of editing targets include the neuronal receptors [23 , 24] , ion transporters [25] , and immune response receptors [26] . While examples of A-to-I RNA editing on critical genes have been known for years , from the evolutionary perspective how and to what extent that A-to-I editing diversifies and shapes the transcriptome and proteome is not fully characterized in the evolution . And very little is known about how RNA editing itself is constrained by selective forces through evolution . There are variable views on the adaptive potentials provided by A-to-I RNA editing . While it was suggested that A-to-I editing on coding genes was non-adaptive from the studies on rhesus and human [22 , 27] , the ‘continuous probing’ hypothesis presented some likely scenario for ‘functional significant editing sites’ [28] . This hypothesis proposed that novel RNA editing sites that emerged on transient double-strand RNA structures , were continuously probed during evolution and became the basis for adaptive selection . And more recently , the non-synonymous high-level A-to-I editing events were proposed to be beneficial in human [29] . The next-generation sequencing technology and the Model Organism ENCyclopedia Of DNA Elements ( modENCODE ) Project [30] enabled an unprecedented resource on the model organisms , like Drosophila and Caenorhabditis , that made it possible for the multi-genome large scale analysis to compare RNA editing patterns in the evolution . To explore the landscape of RNA editing and characterize the selective constraints imposed on A-to-I editing through evolution , we assembled a study based on the modENCODE resource , involving seven Drosophila species for which there were both reference genome and corresponding transcriptome sequencing data available . The study was also complemented with data from other sources , including NCBI Sequence Read Archive ( SRA ) [31] , NCBI Gene Expression Omnibus ( GEO ) [32] , FlyBase [33] , and FlySNPdb database [34] . Using the Drosophila genus as a model system that represents an evolutionary timeframe of approximately 45 million years , we identified a total of 9281 A-to-I RNA editing events . Validations of the events were performed by comparing with results of previous studies and with data from fly tissue/development samples or ADAR mutants , and by carrying out mass array-based validation experiments . Through phylogenetic analysis , the A-to-I RNA editing events were categorized into three distinct types based on the conservation of the editing sites . The profiles and physiological significance of each editing type were analyzed in association with selective constraints through evolution and with functional enrichment in the context of gene ontology ( GO ) . Further evidence revealed the changing patterns of different editing types during holometabolous development and in post-mating response , thus implying the active involvement of RNA editing in short-term response and in normal physiological processes . This work represents a comprehensive study on A-to-I RNA editing in flies at an unprecedented scale , which offers new insight into the evolutionary dynamics of A-to-I editing events , and the critical roles of RNA editing events in fly nervous system .
To explore the A-to-I RNA editome and characterize the evolutionary dynamics of the RNA editing events , we first sought to compile a reference set of events from evolutionarily related model organisms . The Drosophila genus offers some unique advantage for our purpose , as the flies originate from a common ancestor from approximately 45 million years ago ( mya ) ( Fig 1A ) , and many have well annotated quality genome and corresponding transcriptome sequencing data . Upon careful searching the modENCODE data collections , the seven fly species , D . ananassae , D . melanogaster , D . mojavensis , D . pseudoobscura , D . simulans , D . virilis , D . yakuba , were found to meet our needs . Our study utilized the genome and transcriptome sequencing data from samples of whole fly , different tissue types and developmental stages ( S1–S3 Tables; see Methods for details ) . Additional data were acquired to complement the modENCODE data , including D . melanogaster pharate adult dataset , D . pseudoobscura and D . simulans tissue datasets , D . melanogaster genome re-sequencing data , and head RNA-Seq data of the Adar5G1 mutant and paired wild type strain w1118 ( S6 and S12 Tables; see Methods for details ) . To identify A-to-I RNA editing events for the seven species , their whole fly deep-sequencing transcriptome data ( S1 Table ) were initially analyzed . To call A-to-I RNA editing events , we used a modified pipeline ( see Methods for details ) similar to what was described by Ramaswami [36] . We identified totally 9281 A-to-I editing candidate events to generate a reference set , ranging from 826 in D . ananassae to 2052 in D . simulans ( Fig 1A and S4 Table ) . When compared to non A-to-G mismatches from our pipeline , percentage wise the A-to-G editing change was 16- and 14-fold higher than the average of other base change types in all sites and in CDS sites , respectively ( S1 Fig ) . Assuming that all non-canonical mismatches were background noise , and the error rates for all 12 base change were equal , the false positive rate for A-to-G change type was estimated to be 5 . 59% for all sites , and 6 . 32% for CDS sites [36] . ( Annotation of A-to-I editing events is described in the next section . ) These values were in line with those of previous studies , which suggested that almost all non-canonical base changes were due to sequencing errors or alignment artifacts [20 , 36] . To validate the A-to-I editing events and estimate the error rate from our process , we sampled and scrutinized the subset from D . melanogaster , which included 1299 events . First , we compared the D . melanogaster subset with those from previous studies on the same species . 37 , 345 , 361 , 96 , and 564 A-to-I editing events from the studies of Hoopengardner and Stapleton [37 , 38] , Graveley [39] , Rodriguez [10] , Ramaswami [36] and St Laurent [21] overlapped with ours , respectively ( Fig 1B and S5 Table ) . Notably , 37 of the 44 events collected and manually validated by Hoopengardner and by Stapleton were included in our D . melanogaster subset . Collectively , the combined data from those previous studies covered 664 ( 51 . 1% ) of the editing events in our D . melanogaster subset . Second , to further examine the rest 635 events not overlapping with previous studies , we obtained additional transcriptome sequencing data sets generated from pharate adults ( S6 Table ) [40] , from nine tissue types ( S2 Table ) [41] , and from four developmental stages ( S3 Table ) . Within the 635 events , 194 , 294 , and 293 were found with the above datasets , respectively ( Fig 1B and S5 Table ) . Merging together they supported 492 bona fide A-to-I editing events in the group of 635 , which account for another 37 . 9% of the D . melanogaster subset . Taken together , 1156 of 1299 events ( 89 . 0% of the D . melanogaster subset ) either overlapped with the previous studies or were reproduced with new tissue/development samples . When counting editing events in gene coding regions ( CDS ) separately , 675 of 748 CDS events ( 90 . 2% ) were supported by previous data ( Fig 1B ) , which is slightly higher than that for all events . Third , to validate the identified editing events catalyzed by the ADAR enzyme , we obtained and analyzed the RNA-Seq datasets from paired D . melanogaster samples of wild-type strain ( w1118 ) and Adar5G1 mutant [36] . The Adar5G1 mutant flies were found previously to be defective in A-to-I RNA editing [36] . Out of 1299 events in the D . melanogaster subset , 523 were present in the head of the wild type . However , in the head of the Adar5G1 mutant , 485 of the 523 ( 92 . 7% ) were found to have adenosine residues only ( S7 Table ) , confirming the vast majority of identified events are associated with ADAR activity in D . melanogaster . The false positive rate estimated with the Adar5G1 mutant data is 7 . 3% ( 38/523 ) for all events and 8 . 7% ( 27/312 ) for CDS events , in line with other studies using similar scheme [10 , 36] . Forth , we also estimated the false positive rate in the D . melanogaster subset that is due to possible genomic variation , e . g . single nucleotide polymorphism ( SNP ) . We first created a genomic variant database for D . melanogaster , combining the SNP data from FLYSNPdb [34] with variants identified from genome sequencing data ( see Methods for details ) . We then crosschecked our D . melanogaster subset with the genomic variant database ( S1 Text ) . We reasoned that if an A-to-I editing site was found to match an A/G genomic variant , the editing event might be a suspect , possibly resulted from a genomic variant . 110 of the 1299 ( 8 . 95% ) editing events in the D . melanogaster subset and 74 of 748 ( 9 . 89% ) CDS events found A/G correspondents in our genomic variant database . So the estimated false positive rate due to genomic variation is 8 . 95% for all editing events ( 9 . 89% for CDS events ) by our pipeline . We attempted similar analysis to estimate the success rate of A-to-I editing events in other fly species . We were able to recover 74 . 24% and 75 . 91% of all events ( 72 . 77% and 70 . 36% of CDS events ) ( S14 Table ) only for two species , D . pseudoobscura and D . mojavensis , respectively , with RNA-seq data from separate sources ( S12 Table ) . Due to limited tissue types and smaller datasets from these species , the recovery rates for D . pseudoobscura and D . mojavensis are lower than that ( 86 . 0% ) for D . melanogaster . Finally , we carried out mass array-based validation experiments using the Sequenom's MassARRAY platform as described [20 , 22] . On randomly selected A-to-I editing events form all seven fly species , the overall success rates were 86 . 7% for all events , and 89 . 9% for CDS events . So using mass array-based validation approach , the non-confirming rates for all seven species were 13 . 3% for all events and 10 . 1% for CDS events , respectively . They are likely to represent the upper limit of the false positive rate in our work , as many events in the non-confirming category may be missed due to the lower sensitivity of mass array genotyping compared to RNA-seq [20] . Looking more closely into species , the success rates estimated for D . melanogaster , D . mojavensis , D . simulans , D . pseudoobscura , D . yakuba , D . ananassae , and D . virilis were 84 . 6% , 88 . 5% , 100 . 0% , 71 . 0% , 92 . 6% , 90 . 5% , and 83 . 3% , respectively , for all events , and 82 . 4% , 94 . 4% , 100 . 0% , 91 . 3% , 91 . 3% , 92 . 9% , and 76 . 5% , respectively , for CDS events ( S16 Table ) . In summary , analyses of the sampled data suggest our process is effective and reliable for the identification of A-to-I editing events in Drosophila . The seven fly species were found to have comparable success/false positive rates when estimated using mass array-base validation approach . These results are in line with those of the previous studies [21 , 36] , re-enforcing confidence in our analysis pipeline . To characterize the genome distribution of A-to-I RNA editing events in Drosophila , the editing sites were to be annotated with the gene structure information from FlyBase . However , in the current genome releases , the gene models for D . yakuba , D . ananassae , D . simulans , D . mojavensis , and D . virili lacked the definition for 5’- and 3’-UTRs ( untranslated regions ) . So we first redefined the UTR boundaries for gene models in these five species with the help of trancriptome sequencing data ( see Methods for details ) . The UTRs for a total of 62 , 193 gene models were completed ( S2 Text ) . The A-to-I editing sites were then annotated with the newly updated gene structures ( Table 1 and S4 Table ) . Between 16 . 8% and 32 . 1% of events were found in the intronic or intergenic regions in various species ( Table 1 ) . Some events in intergenic regions coincided with non-coding RNAs . For example , in D . malenogaster 30 events were located within its non-coding RNA sequences ( S8 Table ) . The exonic events ( in UTRs or CDS ) accounted for 74 . 5% of all events , for which the majorities ( 74 . 5% ) were found in the CDS that could lead to amino acid coding changes . Indeed , with the exceptions of D . virilis and D . ananassae , A-to-I editing events in CDS regions occupied more than 50% of all events . The RNA editing events were significantly biased toward CDS regions ( S17 Table , Fisher's Exact Test , p-value < 5 . 24E-60 ) , strongly implying function of RNA editing on gene coding sequences in flies . To reveal the tissue profile of A-to-I RNA editing events , we performed hierarchical clustering analysis on the D . melanogaster subset cross nine tissue types ( Fig 2A ) . The A-to-I editing events grouped tissue samples into two apparent clusters , namely nervous tissues ( central nervous system , and head ) versus the rest ( accessory gland , fat body , ovary , salivary gland , digestive system , imaginal disc , and testis ) . We next analyzed the profiles of genes targeted by RNA editing in the D . melanogaster tissues . Considerable variances were displayed in both gene expression abundance and editing level across the tissue types ( Fig 2B ) . To determine the effect of ADAR gene [3 , 42] expression on RNA editing level in flies , we plotted ADAR expression level in all the tissues ( Fig 2B , bottom panel ) . While ADAR exhibited a large variation cross tissue types , to our surprise a poor correlation between ADAR expression and median A-to-I editing levels in D . melanogaster tissues was observed ( Kendall’s tau-b coefficient = -0 . 315 ) . Other confounding factors apart from ADAR expression are suspected to be involved in the regulation of A-to-I editing activity in tissues . Representing the first documented profile for A-to-I editing in flies , the large variances in editing levels in tissues resemble those found in mice [20] , rhesus [22] , or human [43 , 44] . Secondary structure forming around the RNA editing sites plays important role in the substrate-enzyme recognition , thus affecting the efficiency of A-to-I RNA editing . Structural RNAs have lower folding energy [46–49] . We calculated the minimum free energy for secondary structures [50] for the identified editing sites in D . melanogaster and compared them with those for randomly picked sites . Significant difference was observed between sequences flanking editing sites and those random ones ( S4 Fig , Wilcoxon-Mann-Whitney rank sum test , p-value = 5 . 094E-06 ) . The lower median minimum free energy from the editing sites indicates a tendency to form more stable secondary structure around them . In comparison , early studies [11–14] suggested that both the secondary structure and the sequencing context of editing sites were important factors affecting the editing activities . However , apart from the lower median minimum free energy , no strict sequence feature concerning the RNA editing sites was identified in our work . The large fraction of A-to-I editing events concentrating in the CDS regions in Drosophila has a strong functional implication of RNA editing on coding genes . It is imperative to ask what adaptive advantage in evolution , if any , is gained from A-to-I RNA editing . To understand what biological processes and functions are involved in by different A-to-I editing types , we performed Gene Ontology ( GO ) enrichment analysis on the genes of three editing types in D . melanogaster ( see Methods for details ) . There was no GO term reaching the significance threshold ( p-value <0 . 001 ) for the type I events . For the types II events , the top enriched GO categories were potassium ion transport ( p = 1 . 6E-5 ) , extracellular matrix structural constituent ( p = 2 . 6E-5 ) , axon ( p = 1 . 4E-4 ) , learning or memory ( p = 2 . 2E-4 ) , sleep ( p = 3 . 7E-4 ) , ARF guanyl-nucleotide exchange factor activity ( p = 3 . 8E-4 ) , and lysosomal membrane ( p = 3 . 8E-4 ) ( Fig 4 and S13 Table ) . For the type IIIs , the top GO categories were voltage-gated calcium channel complex ( p = 2 . 2E-11 ) , voltage-gated calcium channel activity ( p = 1 . 3E-10 ) , synaptic transmission ( p = 2 . 3E-9 ) , neurotransmitter secretion ( p = 1 . 3E-8 ) , synaptic vesicle ( p = 2 . 3E-8 ) , calcium ion transport ( p = 2 . 3E-8 ) , synaptic vesicle transport ( p = 4 . 1E-8 ) , synapse ( p = 1 . 25E-7 ) , and so on ( Fig 4 and S13 Table ) . Notably , the top 13 GO categories for type IIIs had significant p-value ranging from 10−11 to 10−5 , whereas the top 6 GO terms for type IIs had p-value between 10−5 and 10−3 . The type III events have far more significant GO categories than type IIs , and are almost exclusively concentrated in the functions , components and processes of the nervous system . Similar analyses were also performed with other fly species ( S13 Table ) , and the results resembled that of D . melanogaster . To further strengthen the functional relevance of A-to-I RNA editing , we further investigated the protein domains where A-to-I editing events are located . Our results indicated that type III events were significantly concentrated in functional domains ( Hypergeometric test with p-value adjusted by FDR; p-value = 1 . 74E-38 ) , whereas type I ( FDR adjusted p-value = 1 . 0 ) and II ( FDR adjusted p-value = 0 . 049 ) events were not significant . Looking more closely , type III event-enriched domains/families were heavily related to ion-channel function , including Ion_trans ( FDR adjusted p-value = 4 . 39E-30 ) , Neur_chan_LBD ( FDR adjusted p-value = 9 . 56E-12 ) , Neur_chan_memb ( FDR adjusted p-value = 1 . 75E-08 ) , and Myosin_head ( FDR adjusted p-value = 8 . 87E-08 ) ( S15 Table ) . In light of type III events being the only type subjected to positive selection , the functions of the nervous system may play a unique role in the selection and evolution of type III editing events . Given the functional bias of different A-to-I editing types and the differential selection imposed during evolution , we further looked into their tissue distribution patterns for coordinated evidence about specialization of editing types . We analyzed the transcriptome data sets from modENCODE of nine tissue types for D . melanogaster , and of three tissue types for both D . pseudoobscura and D . simulans ( S12 Table; and see Methods for details ) . The editing events of each type were plotted in the D . melanogaster tissues ( Fig 5A ) . For the type III events , a large majority was detected in the head and the central nervous system , and a small fraction in the other tissues . The occurrence of type I and II events was also elevated slightly in the brain tissues in D . melanogaster . Similar pattern was also supported by the tissue transcriptome data available from D . pseudoobscura and D . simulans ( Fig 5B ) . It is likely that such pattern is held true in other fly species , whose data are limited so far . In agreement with the GO enrichment analysis , these results point to the importance of type III events in brain functions . The gene expression abundance and the editing level for each editing type were further analyzed in D . melanogaster . The median expression abundances in the head and the central nervous system for type III genes were higher than for either type I or II events . Such trend was reversed in all the other tissue types ( Fig 5C ) . The median editing levels in the head and the central nervous system were also higher for type III events than for either type I or II events , with the exception of the central nervous system , where a small number ( only 4 ) of type I events were counted ( Fig 5D ) . However , the median editing levels for type III events were mostly lower in the rest tissue types . Taken together , the type III genes were preferentially expressed and edited in the head and central nerve system . Although biased distribution of A-to-I editing events toward brain tissues was previously reported in rhesus [22] , mice [20] , and human [43] , we showed for the first time that preference was established toward a fraction of the editing events ( type III ) , which were subjected to positive selection associated with nervous/synaptic activities in Drosophila . It is likely that other event types occurring in brain tissues are the by-products of A-to-I RNA editing machinery . On the other hand , although positive selective constraint on type III editing events is overwhelmingly concentrated in the components and functions of the nervous system , we cannot rule out that other functions and processes drive adaptive selection on A-to-I editing events . The high expression abundance and high editing level for some events in the non-brain tissues hint on such possibility ( Fig 5C and 5D ) . To understand the physiological significance of different editing types , we investigated their patterns in two important aspects of fly life cycle: holometabolous development and mating response . First , the occurrences of A-to-I editing events at the four developmental stages in D . melanogaster were analyzed . Embryo , larvae , pupae , and adult shared 133 common editing events ( in 96 genes ) , with 37 and 93 being type II and III , respectively ( S10 Table ) . Considerable changes in A-to-I editing happened between embryo and larvae , between larvae and pupae , and between pupae and adult ( Fig 6A ) . For example , 2 , 105 , and 50 disappeared , and 3 , 27 , and 24 emerged for type I , II , and III events , respectively , in transition from embryo to larvae . They included a type II event on Npc1a ( Niemann-Pick C1 protein ) gene that was lost , and a type III event on Rdl ( glycine receptor alpha-3 ) that emerged . Also note shift in gene expression levels accompanied some of the changes in editing events ( Fig 6A ) . The shifting patterns of different editing types during holometabolous development illustrated the dynamic and active nature of A-to-I RNA editing , which are exemplified by eag and stj genes in Drosophila . eag encodes a voltage-gated potassium channel , for which A-to-I editing could alter amino acid in the critical S6 segment and the cytoplasmic C-terminal domain for binding cyclic nucleotide . We observed a striking pattern in changes of RNA editing level on seven sites throughout fly life cycles ( Fig 6B , top panel ) . Similar patterns on four of these sites were previously reported [58] . The RNA editing-induced changes on eag potassium channel were found to modulate its activation kinetics in D . melanogaster [58] . In contrast , the stj ( straightjacket ) gene , which encodes the alpha ( 2 ) delta subunit of the voltage-gated calcium channel in neurons , exhibited a different editing pattern ( Fig 6B , bottom panel ) . As a critical component involved in the neuromuscular junction development , synaptic transmission , and synaptic vesicle endocytosis [59–61] , this represents the first reported finding on the editing pattern of stj transcripts . We postulate that eag and stj proteins acquire a host of fine-tuned channel property through A-to-I editing with the combination of multiple sites at variable editing levels . The resulting diversity of eag and stj proteins enables a wide range of excitability and complex regulation in fly nervous system . Second , to investigate whether and to what extent the different types of A-to-I editing events are involved in post-mating response in flies , we analyzed the published RNA-Seq data from paired virgin and mated female flies ( S2 Table ) . Mating is known to induce profound physiological and behavioral changes in the female flies . The so-called long-term post-mating changes usually last about a week , involving changes in the expression of hundreds of genes in brain tissues [62 , 63] . Comparing the A-to-I editing events in the head tissues , significant changes in different editing types were observed between day 1 virgin and mated females , and between day 4 virgin and mated females ( Fig 6C and S11 Table ) . Notably , the changes in RNA editing in mated females concentrated in synaptic receptors and ion channels , e . g . synaptotagmin-1 , endophilin-A , glycine receptor alpha-3 , ryanodine receptor-2 , voltage-dependent calcium channel ( beta ) , etc . To our knowledge this is the first reported observation that implies that A-to-I RNA editing is actively involved in the post-mating response in Drosophila .
A-to-I RNA editing adds a critical layer of functional modulation on genes and has been recognized as an important mechanism to expand the genetic repertoire through coding plasticity . The extent of impact of A-to-I editing on the diversity of transcriptome and proteome , and the selective constraint imposed on RNA editing events through evolution are some of today’s key issues in evolutionary biology . Our study was designed to take advantage of the large collection of genome and transcriptome sequencing data that were only available recently . The analysis was performed using the combination of two dimensional data sources: fly species across a defined evolutionary timeframe , and tissue samples across a range of tissue types and developmental stages . The evolution of the A-to-I editing events in Drosophila was revealed with some important observations . First , A-to-I RNA editing on coding genes is confined to a relatively small group of transcripts in the Drosophila phylogeny . Conservatively , about 5% of coding gene families in Drosophila are targeted by A-to-I editing . The majorities of A-to-I editing events are not conserved between homologous genes . Second , based on the conservation of A-to-I RNA editing sites , there appears to be three distinct types of editing events on genes’ coding regions , corresponding to the editing events of different ages . While the type I and IIs are presumably young non-conserved editing events in singleton genes or in multi-member gene families , respectively , type IIIs are conserved events in multi-member gene families . For the majority of editing events , i . e . type IIs , non-synonymous substitutions are deleterious and purged by purifying selection . In contrast , the type III events are driven by positive selection , where non-synonymous changes are preserved . Third , the type III events were found to be concentrated in the head tissues , and highly enriched in a narrow range of components and functions of the nervous system ( Figs 4 and 5 ) . The results from enrichment analysis of type IIIs and its biased distribution suggest that the positive selection on type IIIs is associated with their involvement in the nervous/synaptic activities . While many A-to-I editing cases were reported by others to occur in the nervous system [36 , 37 , 64] , there has not been evidence like ours to show that a clear portion of editing events ( type III ) being positively selected during evolution , are overwhelmingly associated with the nervous system/brain functions . And equally importantly , a larger portion of editing events ( type I and II ) being under purifying selection , do not have such association . Forth , the patterns of different event types were found to shift between developmental stages and in post-mating response in female flies . The data suggest that the A-to-I RNA editing is actively involved in these processes , underlain by a complex regulation of A-to-I RNA editing in flies . The rapid shifts in A-to-I editing can modulate gene function dynamically , having a profound implication for fast acclimatization and rapid response to changing environmental conditions . The adaptive potentials of A-to-I RNA editing are the subject of intense debate over the past years . On one hand , un-controlled editing events can disturb or disrupt the normal gene function networks , hence reducing the fitness of living organisms . On the other hand , RNA editing offers genes coding plasticity that can be advantageous in evolution . The competing probabilities are summarized by the ‘continuous probing’ model [28] . Under this model , new low-level editing events emerge at many sites continuously , which forms the molecular basis for adaptability through continuous selection . Such pool of varying editing sites may confer acclimatizing and adaptive advantage for organisms in changing environments , representing an enhanced evolvability with a low cost in fitness as the un-edited bases are also present to function under normal conditions [28] . Our analysis of A-to-I RNA editing events in flies adds new details to the subsequent process of natural selection . It appears that the non-synonymous A-to-I editing , in general , is rather deleterious . The majority of editing events , i . e . type I and IIs , are driven by purifying selection , in which non-synonymous events are purged ( Fig 3B and 3C ) . The selection mechanism mostly likely operates at the organism level where individuals with detrimental non-synonymous editing events are counter-selected . It is also possible that such counter-selection happens within the cell at the molecular level , but it is a less likely mechanism , as no clear case has been found in support of it . In addition , the neutral non-synonymous editing events , if ever exist , would account for a very small fraction . A-to-I RNA editing observed in our study appeared in general to impose some burden on fitness . On the other hand , a minority of editing events , i . e . type IIIs , are driven by positive selection , which are conserved in homologous genes and preserved across multiple species . These beneficial events are concentrated mainly in functions and components of the nervous system . Although a few cases of beneficial A-to-I editing outside of neuronal receptors and brain-specific ion channels were documented by different researchers [7 , 24 , 65–67] , there was little indication that editing events outside of the nervous system are adaptive , which is contrasting and surprising ( Fig 4 ) . It appears that nervous system functions are mainly tested to determine if an A-to-I editing is beneficial for an organism . Underlying our conclusion , it was suggested that in the brain the broadened diversity of the transcriptome created through A-to-I RNA editing may be part of the process in memory-formation [28] . Coincidentally or not , the oldest ADAR enzymes arising at the beginning of metazoan lineage , accompanied the occurrence of the most primitive nervous system in animals [68] . Our analysis provided a thorough account about the type III events being highly involved in the nervous functions and processes . Previously , the consequences of RNA editing deficiency were revealed by the ADAR mutant flies , which displayed a phenotype of severe behavior dysfunction and neurological defects in the central nervous system [17] . The severe alterations in synaptic ultrastructure and the impaired synaptic release at larval neuromuscular junctions was identified as the cause for defects in synaptic development and for dysfunctions from motility to courtship in ADAR mutant flies [69] . In addition , our work found changes of different editing types occurred throughout the developmental cycles and in post-mating response in Drosophila ( Fig 6 ) , implying the active involvement of A-to-I editing in development and in physiological activities . Supporting our finding at the transcriptome level , individual editing sites were found by previous studies to be developmentally regulated in flies[3 , 70] and in mammals [71 , 72] . Why is the beneficial effect of A-to-I editing observed with the type III events largely limited to the central nervous system in flies , but not in a broader spectrum of biological processes or functions ? While answer to this intriguing but difficult question remains elusive to us , we may speculate that the coding plasticity enabled by A-to-I RNA editing generates a new class of binary variations that uniquely fit the property required for functioning by the animals’ central nervous system . It is possible that ion channels of heterogeneous composition created by RNA editing have become intrinsic components of the functional nervous system . It is also apparent that the ability to fine-tune ion channels and receptors by A-to-I editing cannot be supported by the ‘A/G’ heterozygote , as it is almost impossible to sustain such heterozygosity in all offspring through the diploid mating system . So the A-to-I RNA editing scheme is an effective alternative to maintain heterogeneous components of the nervous system . While we could not rule out the cases of adaptive A-to-I editing that are driven by positive selection from activities outside the nervous system , their restriction mostly to the nervous system is somewhat puzzling . One possible explanation could be that outside the nervous system the benefit of amino acid substitutions from A-to-I recoding is limited , which cannot offset their deleterious effect through evolution . In summary , with the extensive data collections from seven fly species spanning a defined phylogenetic distance , we systematically characterized their A-to-I RNA editome , establishing the prevalence of A-to-I editing and the extent of impact on transcriptome . We further unraveled the evolutionary dynamics of RNA editing events by deriving their time-course of events from closely related species . Importantly , we have shown that A-to-I editing events in CDS regions are grouped into three distinct types based on the conservation of the editing sites . Although A-to-I editing events in general are deleterious , a minority of events ( type III ) that are subjected to positive selection , are mostly associated with the components and function of the nervous system . Tissue specific profiles of the RNA editing types and their changes during holometabolous development and in post-mating response reveal the dynamic nature of A-to-I editing , which points to an underlying mechanism for complex regulation . In essence , the potential of genetic diversity and complexity created by A-to-I RNA editing , and their impact on various bio-physiological processes are shaped and realized by the balance between positive selection on beneficial editing events and the purifying of detrimental ones .
The modENCODE projects are the main source for the Drosophila data used in this study . It is complemented by additional data from NCBI Sequence Read Archive ( SRA; http://www . ncbi . nlm . nih . gov/sra ) and from NCBI Gene Expression Omnibus ( GEO; http://www . ncbi . nlm . nih . gov/geo/ ) . More details on the sequencing data are found in the S1–S3 , S6 and S12 Tables . The whole-fly transcriptome sequencing data for the Drosophila species , D . ananassae , D . melanogaster , D . mojavensis , D . pseudoobscura , D . simulans , D . virilis , D . yakuba , were obtained from modENCODE project: Transcriptional Profiling of additional Drosophila species with RNA-Seq ( Lab: Brian Oliver ) ( S1 Table ) . The tissue transcriptome sequencing data for D . melanogaster were obtained from modENCODE project: Tissue-specific Poly ( A ) Site Profiling of D . melanogaster using Illumina poly ( A ) + RNA-Seq ( Lab: Brenton Graveley ) ( S2 Table ) . The developmental-stage transcriptome sequencing data for D . melanogaster were obtained from modENCODE project: Developmental Time Course Transcriptional Profiling of D . melanogaster Using Illumina poly ( A ) + RNA-Seq ( Lab: Brenton Graveley ) ( S3 Table ) . The transcriptome sequencing data for D . melanogaster pharate adult dataset [40] used for validation was obtained from NCBI GEO under accession number GSE50711 . The head transcriptome sequencing data for the Adar5G1 mutant and paired wild type D . melanogaster strains w1118 were obtained from NCBI SRA under accession numbers: SRR629969 and SRR629970 [36] . The tissue transcriptome sequencing data for D . pseudoobscura and D . simulans were obtained from modENCODE project: Transcriptional Profiling of additional Drosophila species with RNA-Seq ( Lab: Brian Oliver ) ( S12 Table ) , and from NCBI GEO under accession numbers: GSM1258036 , GSM1258037 , GSM1258038 , GSM1258039 , GSM1258040 , GSM775506 , GSM775507 , GSM775508 , GSM775509 , GSM775510 , GSM1306668 , GSM1306669 , GSM1306670 , and GSM1306671 . The genome re-sequencing data for D . melanogaster were obtained from NCBI SRA under accession numbers: SRR485845 , SRR485846 , SRR485847 [10] , SRR1516226 ( BioProject PRJNA244953 ) , and from modENCODE project: Genome assembly and alignment of D . melanogaster OreR virgin female from Bloomington stock to reference r5 ( Lab: Brenton Graveley; DDC id:modENCODE_5518 ) . For analysis , the reference genomes and gene annotation data for Drosophila species , D . ananassae ( r1 . 3 ) , D . melanogaster ( r5 . 53 ) , D . mojavensis ( r1 . 3 ) , D . pseudoobscura ( r2 . 29 ) , D . simulans ( r1 . 4 ) , D . virilis ( r1 . 2 ) , D . yakuba ( r1 . 3 ) were downloaded from the FlyBase ( ftp://ftp . flybase . net/genomes/ ) . Those for A . aegypti ( AaegL1 . 3 , April 2012 ) were obtained from Vectorbase ( https://www . vectorbase . org ) . The raw sequencing data were first processed to remove low quality reads . The sequencing reads were trimmed from both the 5’ and 3’ ends , with a quality score threshold of 20 , using program Sickle ( version 1 . 33 ) [73] . Any reads containing N were also removed . The consequential clean datasets were evaluated with FastQC ( http://www . bioinformatics . babraham . ac . uk/projects/fastqc/ ) . The pipeline for identification of A-to-I RNA editing was modified from what was used in Ramaswami’s work [36] . First , quality RNA-Seq reads from each species were mapped to their genomes using Burrows-Wheeler algorithm [74] , employed by Tophat program ( version 2 . 0 . 8b ) [75] with the parameters ‘-G reference . gtf’ and ‘-N/—read-mismatches’ set to 3 . The reference genomes and related gene models for the Drosophila species were retrieved from FlyBase as described in section: Collection of genome and transcriptome sequencing data . Second , the RNA variances were called using Samtools ( Version: 0 . 1 . 13 ) [76] pileup program with options”-Q 15” . The resulting variant bases were reported with the numbers of reads supporting either the reference genotype or the variance genotypes . Third , the RNA variances were filtered using the following criteria to identify A-to-I editing events: 1 ) variant sites with coverage depth > = 5; 2 ) variant sites located over 10 bp away from either end of a sequence read; 3 ) variant sites with > = 2 non-identical supporting reads; 4 ) variance rate between 1% and 90%; 5 ) occurring in at least 50% of all samples for a species; 6 ) retaining only A-to-G base changing events . A-to-I RNA editing is catalyzed by the enzyme ADAR , and A-to-I editing events were found to be abolished in ADAR-mutant flies . To validate the identified A-to-I editing events and estimate the rate of false positives , we sampled the events occurring in the heads of day 5 wild type ( w1118 ) fly , and compared with those from the heads of day 5 Adar5G1 mutant [36] . The transcriptome sequencing data from day 5 wildtype fly and day 5 Adar5G1 mutant fly were processed , mapped and filtered as described in the section: pipeline for identification of A-to-I RNA editing . For those A-to-I editing events found to occur in the heads of day 5 wild type flies , their corresponding nucleotide resides in the heads of day 5 Adar5G1 mutant flies were examined . Those that were found to be adenosine residues only in Adar5G1 mutant flies are considered genuine A-to-I RNA editing events . We first created a D . melanogaster genomic variant database ( S1 Text ) by combining SNP data from FLYSNPdb [34] with the genomic variant data we identified from the D . melanogaster genome re-sequencing data . Excluding INDELs and other types of polymorphisms , the FLYSNPdb comprised more than 21307 SNP that were imported into our database . In addition , we isolated SNPs using three sets of genome re-sequencing data ( described in the section: Collection of genome and transcriptome sequencing data ) with our SNP pipeline . Briefly , the sequencing reads were mapped to the D . melanogaster genome ( r5 . 53 ) using bowtie2 ( version 2 . 1 . 0 ) [74] with options “-N 1” . The base variances were called using Samtools ( Version: 0 . 1 . 19 ) [76] mpileup program with options”-Q 20” . The resulting base variants were further filtered with following parameters: 1 ) variant sites with coverage depth > = 5; 2 ) variant sites located over 10bp away from either end of a sequence read; 3 ) variant sites with > = 2 non-identical supporting reads; 4 ) variance rate >1% . To identify genomic variants that match an A-to-I RNA editing event , we first filtered the D . melanogaster genomic variant database and retained only A-to-G base changing sites . The resulting A-to-G SNPs were compared with D . melanogaster A-to-I RNA editing sites . Any A-to-I editing site matching a genomic A-to-G SNP was suspected to be resulted from a genomic variant . For experimental validation , the samples of six fly species , D . ananassae , D . mojavensis , D . pseudoobscura , D . simulans , D . virilis , and D . yakuba , were ordered from the Drosophila Species Stock Center at the University of California , San Diego , whereas the samples of D . melanogaster were obtained from Core Facility of Drosophila Resource and Technique , Institute of Biochemistry and Cell Biology , Chinese Academy of Sciences , Shanghai . For each species 20–30 fly individuals were pooled before gDNA and total RNA were extracted in parallel . The gDNA was isolated according to the protocol of VDRC stock center ( http://stockcenter . vdrc . at/control/protocols ) . The total RNA was extracted using RNeasy kit ( Qiagen , Germantown , MD , USA ) and cDNA was synthesized using RevertAid First Strand cDNA Synthesis Kit ( Thermo Scientific , Waltham , MA , USA ) , according to the manufacturers’ instructions . Thirty to thirty-five A-to-I editing sites were randomly chose for each species , with twenty to twenty-five from CDS regions and ten from non-CDS regions . Genotyping was performed on reverse-transcripted cDNA and matching gDNA using the iPLEX Gold Assay ( Sequenom , San Diego , CA , USA ) . Assay primers were designed with the MassARRAY Assay Design software ( version 3 . 1; Sequenom ) . Allele specific extension was performed with iPLEX Gold reagent kit ( Sequenom ) . Extension products were subjected to MALDI-TOF mass spectrometry ( MassARRAY Analyzer Compact; Sequenom ) , according to the manufacturer’s instructions . Genotypes were automatically called using the MassARRAY Typer software ( Sequenom ) , and checked manually . Genotyping results from cDNA and matching gDNA were compared and positive events were confirmed with ‘G’ allele found in cDNA ( G/Total > = 0 . 10 ) and ‘A’ allele found in gDNA ( A/Total >0 . 90 ) , as described by Chen et al [22] . A-to-I RNA editing sites were annotated with ANNOVAR [77] using gene models from FlyBase for the Drosophila species , D . ananassae , D . melanogaster , D . mojavensis , D . pseudoobscura , D . simulans , D . virilis , and D . yakuba . A-to-I RNA editing sites were annotated with gene definitions , including CDS , intronic , 5’UTR , 3’UTR , and intergenic . Those within coding regions ( CDS ) were further defined as “synonymous” or “non-synonymous” based on whether they change the amino acid in protein products . Because the gene models for D . yakuba , D . ananassae , D . simulans , D . mojavensis and D . virili lack the untranslated regions ( UTR ) structure definition for genes , we had to first define their UTR structures as described in the section: Refining UTR regions in Drosophila . We then combined the refined UTR structures with the FlyBase gene models of the five species , which was used in annotation by ANNOVAR . The UTR structures for Drosophila species , D . yakuba , D . ananassae , D . simulans , D . mojavensis and D . virili , were defined with the help of available trancriptome sequencing data . The sequencing reads from whole-fly transcriptome data ( S1 Table ) were first mapped to the reference genomes of D . ananassae ( r1 . 3 ) , D . mojavensis ( r1 . 3 ) , D . pseudoobscura ( r2 . 29 ) , D . virilis ( r1 . 2 ) , and D . yakuba ( r1 . 3 ) with Tophat ( version 2 . 0 . 8b ) . The coverage depth for mapping sequences was reported with Samtools ( Version: 0 . 1 . 13 ) and BEDTools ( Version: 2 . 12 . 0 ) . Then their corresponding gene models ( CDS ) acquired from FlyBase were superimposed to their genome , before the CDS regions were extended upstream and downstream based on mapped reads . The maximum lengths for 5’UTR and 3’UTR were set at 600 bp and 1400 bp , respectively . The parameters were chosen because 95% of 5’UTRs were within 600 bp upstream of translation initiation codons , and 95% of 3’UTRs were within 1400 bp downstream of stop codons in the D . melanogaster gene models . The refined UTRs for gene models in the five species , D . yakuba , D . ananassae , D . pseudoobscura , D . mojavensis and D . virili , are available in S2 Text . The transcriptome sequencing data were processed and mapped to reference genomes as described in the section “Sequence mapping and pipeline for identification of A-to-I RNA editing” . The mapping files were processed with Cufflinks ( v2 . 1 . 1 ) [45] with options “-g * . gff” to estimate the gene expression for nine D . melanogaster tissue types . FPKM ( fragments per kilobase of transcript per million mapped reads ) was used to measure the gene expression abundance . The editing levels for A-to-I editing sites were estimated using the Samtools ( Version: 0 . 1 . 13 ) pileup program , which reported the numbers of reads supporting either the reference genotype or the edited genotype . The editing level for each site was calculated as percentage of reads in edited genotype out of total reads mapped to the site . The tissue and development stage transcriptome sequencing data ( S2 and S3 Tables ) , including the brain tissue RNA-Seq data from virgin and mated female individuals , were processed and mapped to reference genomes as described in the section “Sequence mapping and pipeline for identification of A-to-I RNA editing” . The D . melanogaster RNA editing sites from the reference list were scanned , and the numbers of reads supporting either the reference genotype or the edited genotype were reported and analyzed . Only the events meeting the following criteria were designed to be present in a tissue sample: 1 ) variant sites having coverage depth > = 5; 2 ) variant sites having at least 10 bp from either end of a sequence read; 3 ) variant sites with at least two non-identical reads supporting edited genotype . The hierarchical clustering was performed by first building a matrix based on the presence/absence of A-to-I editing events in the nine different tissue types for all D . melanogaster editing sites from the reference list . The matrix was processed with heatmap function from R ( http://www . r-project . org/ ) using “complete hierarchical cluster” algorithm and option “distfun = dist ( method = ‘euclidean’ ) ” . To calculate the secondary structure minimum free energy for A-to-I RNA editing sites , we first extracted 60 bp sequences flanking the editing sites ( 30 bp upstream and 30 bp downstream ) . The secondary structures for the 61 bp sequences for all sites were built using RNAFold ( 2 . 0 . 7 ) from ViennaRNA Package 2 . 0 [50] with options “—temp = DOUBLE;—dangles = 2;—noGU” , and the minimum free energy for the folding structures was calculated . As a control , random 61 bp CDS regions from 2000 arbitrarily picked Drosophila genes were isolated , and their secondary structures were predicted using the same protocol with minimum free energy computed as described above . To study the conservation of coding genes targeted by A-to-I RNA editing in Drosophila , the homologous gene families were constructed . The entire gene sets from the seven species , D . ananassae ( r1 . 3 ) , D . melanogaster ( r5 . 53 ) , D . mojavensis ( r1 . 3 ) , D . pseudoobscura ( r2 . 29 ) , D . simulans ( r1 . 4 ) , D . virilis ( r1 . 2 ) , D . yakuba ( r1 . 3 ) were downloaded from the FlyBase . The OrthoMCL pipeline [78] was used to cluster encoded gene products into homologous families , as previously described [79] . Briefly , poor quality coding sequences were filtered using the orthomclFilterFasta module with options “min_length = 10; max_percent_stop = 20” . Then BLAST search with blastp was conducted with the option “–e 1E-5” ( E value threshold ) . Clustering with MCL module was performed with options “-abc” and “-i 5 . 0” . The proteins from the seven Drosophila species formed 30 , 434 families , and among them 10 , 820 contained more than one member . Using the clustered homologous gene families from the Drosophila species as reference , the identified A-to-I edited genes were mapped into families . A total of 1 , 526 gene families comprised genes with A-to-I editing events; of which , 133 were singleton genes ( 8 . 72% ) and 1393 were multi-member gene family ( 91 . 28% ) . Based on the conservation of RNA editing sites , the CDS events were categorized into three types . The type I events occurred in singleton genes that did not have detectable homologous gene in other fly species . The type II events were non-conserved editing events in multi-member gene families , but each occurred in one member and had no conserved event in other members of the same family . The type III events referred to conserved editing events occurred in at least two members of a multi-member gene family . We investigated the event gains and losses of type III events along the phylogeny using the Gain Loss Mapping Engine ( GLOOME ) [52] ( http://gloome . tau . ac . il/ ) . ( Since each of type I or II events is only present in one terminal leaf of the phylogenetic tree , it is not necessary to include them in the analysis ) . The type III events were grouped into 402 clusters based on conservation of editing sites . Then the presence and absence profile ( phyletic pattern ) was generated [52] based on the clustering of type III events . With uploaded phyletic pattern matrix of type III events , GLOOME server inferred branch specific gain and loss events along the phylogeny using stochastic mapping [53] . The selective pressure on the coding genes targeted by A-to-I editing was analyzed using the Ka/Ks value ( the ratio of non-synonymous nucleotide substitution rate to the synonymous substitution rate ) [55] . The orthologous genes from A . aegypti were used as outgroup in computing Ka and Ks values . The genes harboring A-to-I editing events were paired with its orthologs from A . aegypti , which were identified using bidirectional best hits ( BBH ) algorithm [80] . The Ka and Ks values for each pair were computed with codeml program from PAML package , using maximum-likelihood method [81] . The Ka and Ks values were then corrected with Colbourne’s protocol [82] . To investigate the details of purifying selection on genes with type III events , the Ka/Ks values for the local neighbor sequences near A-to-I editing sites were calculated using shifting windows with a size of 11 codons . For each shifting window , the Ka/Ks value of a local sequence was computed with codeml program using the 11-codon aligned block between the local sequence and orthologous one from A . aegypti . The genes with different types of RNA editing events in D . melanogaster were compiled , and the lists of type I , II , and III genes were created , respectively ( S13 Table ) . Gene ontology ( GO ) enrichment analyses were performed on genes of each editing type with GOseq package [56] from R using the Hypergeometric test with p-values adjusted by false discovery rate ( FDR ) control procedure [57] . A significant GO term required at least two enrichment genes and five background genes . The GO terms at the top of the tree hierarchy , namely cellular component ( CC ) , biological process ( CC ) , and molecular function ( MF ) , were excluded from the significant list . Protein domain enrichment analyses were performed on the protein domains where A-to-I editing events fall in . Genes with A-to-I editing events were annotated with domain information using Pfam webserver ( v29 . 0 ) [83] ( http://pfam . xfam . org/ ) with default parameters . The proportion of number of editing events within domains over total event number was tested against the proportion of all domain size over all gene size . Domain enrichment analyses were performed using the Hypergeometric test similar to that described in the GO enrichment analyses . | One prevalent form of RNA editing is the deamination of adenosines ( A-to-I editing ) in the precursor mRNA molecules , pertaining to most organisms in the metazoan lineage . While examples of A-to-I editing on critical genes have been known for years , it has not been fully characterized how A-to-I editing shapes the transcriptome and proteome in the evolution . To understand how A-to-I editing affects genes’ evolution and how itself is constrained by selection , we generated a global profile of A-to-I editing for a phylogeny of seven fly species , a model system representing an evolutionary timeframe of about 45 million years . We are focused on 5150 editing sites ( of totally 9281 identified ) located in the coding region of 2734 genes . Our analysis revealed the evolution dynamics of A-to-I editing sites and functional specificity of targeted genes . The shifting patterns of A-to-I editing are documented during holometabolous development and in post-mating response in flies . This work points to the important roles of regulated RNA editing in animal development and offers new insight into the evolution of A-to-I editing events and their harboring genes . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"invertebrates",
"computational",
"biology",
"animals",
"invertebrate",
"genomics",
"animal",
"models",
"drosophila",
"melanogaster",
"model",
"organisms",
"genome",
"analysis",
"drosophila",
"research",
"and",
"analysis",
"methods",
"biological",
"databases",
"molecular",
... | 2016 | The Landscape of A-to-I RNA Editome Is Shaped by Both Positive and Purifying Selection |
Japanese encephalitis ( JE ) is a major cause of mortality and morbidity for which there is no treatment . In addition to direct viral cytopathology , the inflammatory response is postulated to contribute to the pathogenesis . Our goal was to determine the contribution of bystander effects and inflammatory mediators to neuronal cell death . Material from a macaque model was used to characterize the inflammatory response and cytopathic effects of JE virus ( JEV ) . Intranasal JEV infection induced a non-suppurative encephalitis , dominated by perivascular , infiltrates of mostly T cells , alongside endothelial cell activation , vascular damage and blood brain barrier ( BBB ) leakage; in the adjacent parenchyma there was macrophage infiltration , astrocyte and microglia activation . JEV antigen was mostly in neurons , but there was no correlation between intensity of viral infection and degree of inflammatory response . Apoptotic cell death occurred in both infected and non-infected neurons . Interferon-α , which is a microglial activator , was also expressed by both . Tumour Necrosis Factor-α , inducible nitric oxide synthase and nitrotyrosine were expressed by microglial cells , astrocytes and macrophages . The same cells expressed matrix metalloproteinase ( MMP ) -2 whilst MMP-9 was expressed by neurons . The results are consistent with JEV inducing neuronal apoptotic death and release of cytokines that initiate microglial activation and release of pro-inflammatory and apoptotic mediators with subsequent apoptotic death of both infected and uninfected neurons . Activation of astrocytes , microglial and endothelial cells likely contributes to inflammatory cell recruitment and BBB breakdown . It appears that neuronal apoptotic death and activation of microglial cells and astrocytes play a crucial role in the pathogenesis of JE .
Japanese encephalitis virus ( JEV ) continues to be the leading cause of viral encephalitis in Asia and the Western Pacific , where it is a significant cause of mortality and disability . Annually there are estimated to be up to 70 , 000 cases , with 10 , 000–15 , 000 deaths [1] . Although vaccination is the most viable option to prevent the disease , affordable vaccines are still not widely available , and there is no established treatment for JE . Despite the disease's importance , little is known about the pathogenesis . During in vitro studies neuronal apoptosis was described [2] , but its mechanisms and relevance for the disease are still unclear , in particular in relation to the inflammatory response that develops alongside direct viral cytopathology . Opportunities for in depth neuropathogenic studies on JE in humans are very limited , mainly because autopsy tissue from fatal human cases is rarely available due to cultural constraints in many areas where JE occurs . Mouse models of pathogenesis have some similarities to human disease , but there are also differences [3] , [4] . The macaque model , developed in the 1990s to test JE vaccines is a useful model for studying human disease , particularly since the macaque immune system closely resembles that of humans [5] . We therefore conducted a retrospective study on the brains of experimentally JEV-infected macaques , to dissect the inflammatory response and the cascade of events that leads to neuronal damage . We were especially interested in apoptotic pathways and inflammatory mediators including cytokines , inducible nitric oxide synthase ( iNOS ) and matrix metalloproteinases ( MMPs ) , because these may point towards new targeted treatments to control the inflammatory damage , even in the absence of antiviral therapy .
The study does not involve animal use as it was conducted on archived paraffin embedded brain tissue of rhesus macaques ( Macaca mulatta ) . The original research on challenge study was conducted in compliance with the Animal Welfare Act and other federal statutes and regulations relating to animals and experiments involving animals and adheres to principles stated in the Guide for the Care and Use of Laboratory Animals , NRC Publication , 1996 edition . The original study was approved by the Institutional Animal Care and Use Committee ( United States Army Medical Component , Armed Forces Research Institute of Medical Sciences ) and by the Animal Use Review Office , United States Army Medical Research and Materiel Command ( Permit Number: 93-11 ) . The study was performed on archived paraffin embedded brain tissue of twelve rhesus macaques challenged intranasally with a well characterized wild-type JEV strain ( KE93; Genotype Ia , GenBank accession number KF192510 . 1 ) as part of an effort to evaluate second-generation JEV vaccines [5] ( Table 1 ) . All archived specimens used in this study are from unvaccinated monkeys . The challenge study had been undertaken in several phases and with different doses , ranging from 7 . 5×105 to 2×1010 plaque forming units [6] . Monkeys originating from India and screened negative for both JEV and Dengue virus neutralizing antibodies ( aged 3–7years , of both sexes , weighing 4 . 0–9 . 9 kg ) had been intranasally inoculated either with the virus isolate passaged twice in culture ( animals 1 and 2 ) or with an isolate prepared from the brain of animal 2 that was subsequently passaged twice in suckling mice to increase both virus titer and virulence [6] . The monkeys were euthanized at the onset of stupor or coma ( 10–13 days post inoculation ) and JEV infection was confirmed by virus isolation from the brain . Five age-matched uninfected control monkeys from an unrelated study served as negative controls . Immediately after death , brains were exenterated and sections of frontal lobe , thalamus , brainstem and cerebellum fixed in 10% neutral buffered formalin for at least 72 hours . Following routine paraffin wax embedding , 3–5 µm sections were prepared and stained with haematoxylin-eosin ( HE ) or used for immunohistology . For immunohistological studies , sections of thalamus and brainstem ( exhibiting the most consistent histological changes ) and , for comparison , the cortex ( absence of inflammatory infiltrates ) were chosen . These were stained for the presence of JEV antigen , apoptosis and apoptotic pathway markers , glial and inflammatory cell markers , von Willebrand Factor ( to confirm blood brain barrier [BBB] breakdown , through the demonstration of plasma protein leakage ) , and proinflammatory markers . Commercial antibodies to human proteins were selected for this study , especially those known to cross react with Macaca mulatta . Details on the panel of antibodies and the detection methods used are provided in Table S1 . Briefly , sections were dewaxed in xylene and hydrated through graded alcohols . To inhibit endogenous peroxidase activity , they were treated with freshly prepared 3% H2O2 for 15 min . Sections underwent heat-induced antigen/epitope retrieval with a laboratory pressure cooker ( Decloaking Chamber , Biocare Medical , Concord , USA ) using citrate buffer pH 6 or pH 9 [7] . This was followed by incubation with normal serum to block non-specific binding sites in tissues , and the primary antibodies ( 15–18 hrs at 4°C ) ( see Table S1-A ) . Apoptotic cells were also identified by the terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate nick end in situ labelling ( TUNEL ) method using the Apoptag In Situ Apoptosis Detection kit ( Chemicon Inc . , Millipore , Billerica , USA ) to demonstrate the characteristic DNA changes . Appropriate controls were included for each marker: uninfected control monkey brains as negative controls for JEV and to establish constitutive expression of other markers , sections with known positivity for specific markers as positive controls , and sections incubated with normal mouse/rabbit IgG as isotype controls . Double immunolabeling was performed on selected sections of some monkeys ( animals 2 , 9 , 11 ) to characterize the populations of cells expressing apoptosis markers ( TUNEL and caspase-3 , -8 , and -9 ) and proinflammatory mediators ( cytokines , iNOS and MMPs ) and to relate them to the expression of JEV antigen . For this purpose , primary antibodies raised in different species were sequentially localized using non-overlapping secondary reagents and different chromogens ( see Table S1-B ) . Sequential staining was performed on consecutive sections , mainly to detect tumor necrosis factor alpha ( TNF-α ) expression in inflammatory cells and glial cells and to further characterize JEV-infected cells when primary antibody were used that had been generated in the same species or when the double immunolabeling was difficult to interpret . A confocal laser scanning microscope LSM 700 ( Carl Zeiss Micro Imaging , Germany ) with solid state laser excitation wavelength 488 nm ( for FITC ) and 555 nm ( for Texas Red ) and ZEN 2009 software was used to detect immunofluorescent staining . All other light microscopic assessments were undertaken with conventional microscopes .
All JEV-infected animals exhibited mild to moderate , multifocal to diffuse , non-suppurative meningoencephalomyelitis with evidence of neuronal degeneration and death . The inflammatory response was similar in its extent and composition regardless of the dose of inoculum and the day of euthanasia , and was dominated by mononuclear perivascular cuffs ( Figure 1A ) and meningeal infiltrates . These were accompanied by morphological evidence of endothelial cell activation ( represented by a tomb-stone like luminal protrusion of endothelial cells; Figure 1B ) and/or vascular damage . The latter was indicated by perivascular haemorrhage and substantial leakage of serum into the parenchyma , as demonstrated by staining for von Willebrand factor ( Figure 1C ) . Neuronal cell death was indicated by morphological neuronal changes suggestive of apoptosis , in association with satellitosis or microglial nodules ( Figure 1D , E ) . Reactive astrogliosis , represented by a multifocal increase in astrocyte numbers ( Figure 1F ) and evidence of astrocyte activation ( presence of gemistocytes ) in areas with inflammatory infiltrates was also identified in JEV infective brains . T cells ( CD3+ ) were the predominant leukocytes in both perivascular and meningeal infiltrates . They were also present in small numbers in the adjacent parenchyma ( Figure 2A ) . B cells ( CD20+ ) were sparse and primarily seen in the perivascular infiltrates ( Figure 2B ) , while moderate numbers of macrophage/microglial cells ( CD68+ ) identified in perivascular and meningeal infiltrates and the adjacent brain parenchyma ( Figure 2C ) . Staining for myeloid/histiocyte antigen , reported to stain macrophages [8] and microglial cells [9] , identified a substantial number of cells with a morphological appearance of macrophages ( Figure 2D ) , suggesting their recruitment into the tissue . Staining for CD68 , which is also expressed by microglial cells , and major histocompatibility complex ( MHC ) class II antigen ( expressed mainly by activated microglial cells ) confirmed the presence of microglial nodules but also demonstrated diffuse microgliosis and activation of microglial cells ( presence of both reactive and amoeboid microglial cells; Figure 2 C , E ) . Furthermore , endothelial cells were shown to express MHC II , confirming their activation ( Figure 2E ) . The cells surrounding neurons in satellitosis were also CD68-positive microglial cells ( Figure 2F ) . For comparison , in brain areas without evidence of viral antigen and inflammation ( cerebral cortex ) , only scattered MHCII-positive microglial cells without morphological features of activation were seen . There was no evidence of microglial MHC II expression in control brains . JEV antigen expression , seen as finely granular cytoplasmic staining , was observed in numerous neuronal cell bodies and processes disseminated in the thalamic and brain stem nuclei of all animals and in neuronal cell processes throughout the affected parenchyma ( Figure 3A ) . Most infected neurons appeared morphologically unaltered ( Fig . 3A inset ) , but some were surrounded by microglial cells ( satellitosis ) and exhibited degenerative changes ( Figure 3B ) . JEV-positive microglial cells were found in some glial nodules , but occasionally as individual cells in affected areas like brainstem and thalamus , as confirmed by sequential staining for CD68 and JEV antigen ( Figure 3C ) . In contrast , there was no evidence of JEV infection of astrocytes ( Figure 3D ) . In one animal with a particularly strong inflammatory response ( animal 2 ) , a small percentage of slender perivascular cells ( perivascular macrophages ) also expressed viral antigen ( Figure 3E ) . There was no evidence of JEV antigen in endothelial cells in any animal . Nor was there any correlation between intensity of viral infection as indicated by immunostaining and degree of inflammatory response . Negative control brain sections did not show any positive reaction . Morphological features of apoptosis were observed in degenerating neurons within glial nodules and in satellitosis , among leukocytes in the perivascular infiltrates and in individual cells with microglial features in the adjacent parenchyma . Cell death by apoptosis was confirmed by the TUNEL method which identified apoptotic JEV-infected neurons in glial nodules and satellitosis as well as apoptotic microglial cells disseminated in the parenchyma , in satellitosis and in microglial nodules ( Figure 3F ) . Occasional lymphocytes in the perivascular infiltrates were also apoptotic ( Figure 3G ) and the JEV-infected perivascular macrophages were apoptotic in animal 2 ( Figure 3E ) . Key apoptosis molecules , including caspases-8 , -9 ( both initiator caspases ) and cleaved caspase-3 ( an executor caspase ) were identified by staining to detect cells undergoing early apoptosis and not exhibiting representative morphological features . Small numbers of neurons with normal morphology expressing cleaved caspase-3 and more cells expressing caspase-8 were seen in JEV infected brains . Both caspases were also expressed by some leukocytes in the perivascular infiltrates ( Figure 4A , B ) . Caspase-9 , however , was only detected in astrocytes and microglial cells ( Figure 4C ) . Double staining for JEV and the various apoptosis markers confirmed that some JEV-infected neurons were undergoing apoptosis ( data not shown ) . In order to better understand the regulation of apoptotic processes in response to JEV infection , the expression of representative pro- and anti-apoptotic proteins was assessed . While numerous microglial cells and occasional neurons stained positive for the pro-apoptotic protein Bax ( Figure 4D ) , the anti-apoptotic protein Bcl-2 was mainly expressed by lymphocytes in the perivascular infiltrates ( Figure 4E ) . Dual staining showed JEV antigen in some Bax-positive neurons and occasional Bax-positive microglial cells ( data not shown ) . In uninfected control brains TUNEL positive cells were not identified . Caspase and Bcl-2 staining was negligible; weak and infrequent Bax expression was seen in neurons . Having characterized the inflammatory response and the patterns of cell death in the brains for monkeys infected with JEV , we aimed to identify relevant mediators of these processes frequently identified in viral mediated infections . To assess local nitric oxide ( NO ) production , we investigated the expression of iNOS and nitrotyrosine ( NT ) . We stained for MMP-2 and -9 , which are known to cause BBB disruption by degrading collagen IV , its main component [10] , interferon ( IFN ) -α , a potent antiviral cytokine and microglial activator [11] , and TNF-α which has been shown to directly activate microglia [12] and induce neuronal apoptosis [13] . Both iNOS and NT were expressed by microglial cells and astrocytes . iNOS expression was also seen in some macrophages in the perivascular infiltrates and the adjacent parenchyma ( Figure 5A , B ) where staining for NT was only very weak . MMP-2 was expressed in cells with the morphology of reactive astrocytes ( Figure 5C ) and , to a lesser extent , in microglial cells and in infiltrating macrophages , whereas MMP-9 , known to be constitutively expressed in human neurons , was intensely expressed by neurons and relatively weakly by microglial cells ( Figure 5D ) . TNF-α expression was seen in microglial cells , infiltrating macrophages and astrocytes , as confirmed by dual staining with CD68 and sequential staining with GFAP ( Figure 5E ) . It was also occasionally seen in endothelial cells ( data not shown ) . IFN-α expression , however , was seen both in uninfected and infected neurons , as confirmed by dual staining with JEV antigen ( data not shown ) , and in astrocytes and microglial cells ( Figure 5F ) . In control brains , only minimal expression of inflammatory mediators was seen , represented by staining in occasional vascular endothelial cells ( iNOS , TNF-α ) , neurons ( MMP-9 , iNOS ) and vascular smooth muscle cells ( TNF-α ) .
The present study used macaques , which have previously been established as a good model for neuropathological studies on JE in humans [5] , [6] , to evaluate the cytopathic effects of and inflammatory response to JEV in the brain . The apoptosis pathways and the full spectrum of proinflammatory factors have not been fully studied in any previous animal models of JE , or autopsy tissues . This study utilized monkeys challenged with JEV intranasally rather than a route more consistent to natural infections to increase the likelihood of encephalitis . Peripherally challenged monkeys generally do not typically develop encephalitis [14] and with direct intracerebral challenge the encephalitis develops early [15] . The intranasal route was therefore the most useful route in our model and has been reported to provide a useful model for the study of anti-viral compounds and vaccine candidates [5] , [15] albeit this unnatural infection route may be a limitation in our study . As in humans , JEV induces a non-suppurative meningoencephalitis with neuronal cell death , microgliosis and astrogliosis in macaques [16] , [17]; these classic findings are also common in other viral encephalitides [18] . However , the ‘punched-out’ areas of focal necrosis , often seen in fatal human JE cases [16] , [19] were not observed in our experimentally infected monkeys . It is possible that this pathology had not yet developed in the macaques that were euthanized at the onset of stupor or coma in contrast to human infections where histological observations are always made on post mortem material at the end of the disease process [16] , [19] . The inflammatory response in macaques even with the chosen challenge route was consistent with the changes seen in humans , characterised by perivascular mononuclear cuffs , with less intense infiltrates in the adjacent parenchyma [16] . While T cells dominated in the perivascular infiltrates and recently recruited macrophages were the largest population in the parenchymal infiltrates , B cells represented a minority and were restricted to the perivascular cuffs . Cytotoxic T cells ( CTLs ) have been reported to play a key role in mouse models of JE [20] , but it remains unclear if these cells are beneficial or deleterious , or both . In the present study , it was not possible to assess the role of CTLs , due to the non-availability of antibodies suitable for macaques . In viral encephalitis , macrophages are known to migrate from the perivascular space into the surrounding parenchyma where they become activated [21] . In addition to microglia , known to cause neuronal death in JE [3] , [19] , the relative contribution of peripheral macrophages that migrate into the CNS should be elucidated . Our study confirmed neurons as the main targets of JEV , as previously shown in fatal human cases [16] , [19] , [22] . We also demonstrated viral antigen in microglial cells , mainly within microglial nodules surrounding infected neurons , suggesting virus uptake by phagocytosis . However , productively infected microglial cells cannot be excluded , since they do support viral replication in vitro [23] , [24] . Viral antigen was not detected in other glial cell types , despite evidence that astrocytes can become infected in culture systems [23] . There was also no evidence of endothelial cell infection . A similar viral target cell pattern has been reported in human cases , with the exception that some studies found evidence also for endothelial cell infection [16] , [19] . Interestingly , we detected JEV antigen in perivascular macrophages in one animal . These cells found at the interface between blood and brain parenchyma are resident macrophages with high phagocytic activity and MHC-II expression [25] , which suggests that they had phagocytosed virus that entered the brain via the blood . Viral infection and inflammatory responses were associated with cytopathic changes , and , although not excessive , neuronal death via apoptosis was clearly observed . Apoptosis was shown by the TUNEL assay which has been used in the past to demonstrate apoptosis , although interpretation of the findings can be difficult in the presence of necrosis and autolytic changes [26]; we therefore also confirmed apoptosis by staining for cleaved caspase-3 . Apoptotic neurons were often surrounded by microglial cells ( satellitosis and formation of microglial nodules ) which indicated their impending phagocytosis . Some apoptotic neurons were JEV infected . In addition , several morphologically unaltered , infected neurons were shown to express the pro-apoptotic protein Bax , the initiator caspase-8 or the active effector caspase-3 , which indicates that these cells were destined to become apoptotic . These results confirm the in vivo relevance of previous in vitro studies which demonstrated that JEV replication can lead to neuronal apoptotic death [27] and support findings from the mouse model that JEV replication contributes to Bax activation [28] . Taken together , these findings provide clear evidence of a direct , although possibly not rapid , cytopathic effect of JEV on neurons . The demonstration of caspase-8 in affected neurons also indicates that neuronal apoptosis is initiated by the fas-mediated or extrinsic pathway , a mechanism that is central to the process of immune-mediated viral clearance [29] and seen in a number of CNS viral infections including West Nile virus [30] . Importantly , apoptotic cell death or pre-apoptotic caspase-8 expression was also seen in a proportion of JEV antigen-negative neurons , which suggests some degree of bystander neuronal death . In addition , a proportion of microglial cells , often in close proximity to infected neurons but generally not JEV-infected , were apoptotic . Furthermore , the observation of morphologically unaltered microglial cells expressing caspase-9 suggest that microglial apoptosis is initiated by the mitochondria or the intrinsic pathway . A recent in vitro study showed that JEV infection can lead to apoptosis of microglial cells [24] . Our results indicate that in vivo this direct mechanism is probably less relevant and that pro-inflammatory factors are more important; this is also seen in other CNS conditions , such as experimental autoimmune encephalomyelitis ( EAE ) where microglial apoptosis is considered an important homeostatic mechanism to control microglial activation and proliferation [31] . Apoptotic cell death was also observed in a proportion of infiltrating inflammatory cells in our JEV infected monkeys . Considering that these cells were not JEV-infected , this most likely represents a normal mechanism to eliminate activated leukocytes and thereby limit the inflammatory response in the CNS . On the other hand , infiltrating leukocytes ( predominantly T cells ) were found to express the anti-apoptotic protein Bcl-2 . This supports a murine in vivo study that provides evidence of a critical role of Bcl-2 in the survival of virus-specific CTLs [32] . The occurrence of apoptosis in apparently uninfected neurons suggests that indirect mechanisms ( bystander cell death ) contribute to neuronal damage in JE , and indeed recent in vitro and in vivo murine studies demonstrated that microglial cells can induce neuronal apoptosis via the release of pro-inflammatory mediators [3] , [4] . Also , TNF-α , via its receptor on neurons , has been shown to induce caspase-8 activation in mouse neurons [33] . Indeed , we observed TNF-α upregulation in astrocytes , microglial cells , endothelial cells and infiltrating macrophages in infected macaques . It is likely that these cells were also responsible for the TNF-α upregulation observed in JEV-infected mice [3] , [34] . TNF-α related neuronal death is also reported in a recent in vitro study with WNV [35] . The results of our study suggest that JEV might simultaneously trigger , both directly and indirectly , the caspase dependent extrinsic apoptotic pathway in neurons and the intrinsic apoptotic pathway in microglial cells . Further definition of the underlying mechanisms will allow us to understand the processes involved in disease progression and to assess the potential of anti-apoptotic treatment strategies . Alongside the inflammatory infiltration and the cytopathic effects , we found distinct evidence of activation of a range of cells , namely microglial cells , astrocytes and vascular endothelial cells . Microglial activation was confirmed by the demonstration of MHC II antigen , iNOS , NT , TNF-α and MMP expression by microglial cells and has been reported previously in JEV-infected mice [3] . To shed light on the potential mechanism of microglial activation , we assessed the expression of IFN-α ( type I IFN ) ; this potent antiviral cytokine is an activator of microglia in response to CNS viral infection [11] , and is elevated in the cerebrospinal fluid of patients with JE , where it is associated with a poor outcome [36] . We demonstrated IFN-α expression in neurons which suggests that they might be responsible for microglial activation early after infection; expression by microglia and astrocytes suggests they might be responsible for sustained microglial activation in JE . As described in earlier reports [22] , reactive astrogliosis and astrocyte activation was also observed in the present study . Astrocyte activation is considered as a non-specific response to degenerative changes including virus-induced damage in the CNS . However , a recent study provided evidence that this activation might be an effect of TNF-α release from microglial cells [23] . So far , little is known about the role of astrocytes in neuroinflammation caused by JEV , whether they are protective or pathogenic . Nevertheless , the demonstration of TNF-α , IFN-α , iNOS , NT and MMP-2 expression by astrocytes in our study provides the first in vivo evidence that astrocytes may play an important role in the pathogenesis . The same is true for microglial cells and macrophages in the inflammatory infiltrates , through release of the inflammatory mediators , all these cells might actively contribute to the damage of other cells in the brain and in particular induce bystander apoptotic death of neurons [3] , [4] . iNOS and NT expression indicate NO production , which is in accordance with results from a mouse study [37] . There , a gradual increase in iNOS activity was observed after intracranial JEV infection , and was considered a consequence of release of cytokines , such as TNF-α or IL-8 which might be beneficial through the inhibition of viral replication and release [37] . However , NO has also been discussed as a potential mediator of pathogenesis in tick-borne encephalitis virus infection [38] . MMP levels have been shown to correlate with the severity of some CNS infections [39] . MMP-9 is known to be constitutively expressed in human neurons . However , it was intensely upregulated in neurons of the JEV-infected macaques and weakly expressed by microglial cells , while glial cells and infiltrating macrophages were sources of MMP-2 . MMP release is stimulated by proinflammatory cytokines including TNF-α [40] . In JE , MMPs might play a detrimental role and not only be responsible for BBB disruption through collagen IV degradation , but also contribute to neuronal destruction via stimulation of TNF-α release . We observed endothelial cell expression of MHC II antigen and TNF-α , which confirms that they are activated and suggests they have a role in inflammatory cell recruitment and potential contribution to immune reactions , glial cell activation and neuronal apoptosis . Endothelial cells might also be a source of the increase in serum TNF-α seen in JE patients [36] . Based on our findings we postulate that infection of neurons by JEV triggers a network of inflammatory mediators [41] . Through release of IFN-α , neurons activate microglial cells which , via release of cytokines such as TNF-α , activate astrocytes and endothelial cells . Together , these mediators contribute to BBB breakdown , leukocyte recruitment into the parenchyma and further neuronal apoptosis . Glial cell apoptosis should limit the extent of inflammation . However , the release of further mediators by infiltrating leukocytes , in particular macrophages , results in sustained glial and endothelial cell activation and further leukocyte recruitment , ultimately augmenting the inflammatory response and neuronal cell loss . Although the inflammatory response is intended to be protective , and presumably is so in cases which improve and recover , if uncontrolled it can contribute to disease progression in JE . Our study is mostly descriptive as we used archived materials from a previous challenge study . However it might shed some light on some novel processes mediating pathogenesis which could aid in the experimental design for future studies investigating inflammatory responses to JE . Viral encephalitis is a major cause of morbidity and mortality worldwide . The pathogenesis of flavivirus encephalitis remains incompletely understood but it appears that the immune response is crucial in limiting viral spread to the brain [42] . The cascade of events that we have outlined for JE may also apply to other viral encephalitides . Currently there is no proven efficacious therapy for most viral infections of the CNS including JE . Novel strategies for treating viral CNS infections are urgently needed . Our results from a macaque model indicate that neuronal apoptosis and glial activation are crucial steps in the pathogenesis of JE . They imply that adjunctive therapy with inhibitors of caspases or targeted anti-inflammatory treatments might be a promising therapeutic approach for JE in the future . | Japanese encephalitis ( JE ) is one of the most important causes of viral encephalitis worldwide , with no specific antiviral treatment available . Despite some recent successes with widespread vaccination , JE will likely remain an important public health problem; because the virus is mosquito-borne and has natural animal hosts , it will never be eradicated . We have little understanding of what determines the severity and outcome of infection . Data from human post mortem studies is very limited because of cultural constraints on autopsies in areas where JE occurs . Circumstantial evidence suggests that in addition to cytopathology caused directly by infection of neurons , there may be bystander cell death of non-infected neurons , caused by an excessive inflammatory response . Our study used archived brain samples from a prior challenge study in a validated macaque model of JE . We stained for the presence of JEV antigen , apoptosis , and pro-inflammatory markers in affected areas , such as the thalamus and brainstem . We show that bystander neuronal cell death is important , and elucidate the inflammatory and apoptotic mechanisms underlying it . Currently there is no proven efficacious therapy for most viral infections of the central nervous system , including JE . Novel strategies for treating such infections are urgently needed . Our findings suggest new anti-inflammatory and anti-apoptotic therapeutic approaches may be useful in treating this debilitating disease . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"veterinary",
"pathology",
"histology",
"infectious",
"diseases",
"medicine",
"and",
"health",
"sciences",
"anatomy",
"veterinary",
"microbiology",
"biology",
"and",
"life",
"sciences",
"immunology",
"microbiology",
"neuroscience",
"veterinary",
"science"
] | 2014 | Neuropathogenesis of Japanese Encephalitis in a Primate Model |
RNA viruses exhibit substantial structural , ecological and genomic diversity . However , genome size in RNA viruses is likely limited by a high mutation rate , resulting in the evolution of various mechanisms to increase complexity while minimising genome expansion . Here we conduct a large-scale analysis of the genome sequences of 99 animal rhabdoviruses , including 45 genomes which we determined de novo , to identify patterns of genome expansion and the evolution of genome complexity . All but seven of the rhabdoviruses clustered into 17 well-supported monophyletic groups , of which eight corresponded to established genera , seven were assigned as new genera , and two were taxonomically ambiguous . We show that the acquisition and loss of new genes appears to have been a central theme of rhabdovirus evolution , and has been associated with the appearance of alternative , overlapping and consecutive ORFs within the major structural protein genes , and the insertion and loss of additional ORFs in each gene junction in a clade-specific manner . Changes in the lengths of gene junctions accounted for as much as 48 . 5% of the variation in genome size from the smallest to the largest genome , and the frequency with which new ORFs were observed increased in the 3’ to 5’ direction along the genome . We also identify several new families of accessory genes encoded in these regions , and show that non-canonical expression strategies involving TURBS-like termination-reinitiation , ribosomal frame-shifts and leaky ribosomal scanning appear to be common . We conclude that rhabdoviruses have an unusual capacity for genomic plasticity that may be linked to their discontinuous transcription strategy from the negative-sense single-stranded RNA genome , and propose a model that accounts for the regular occurrence of genome expansion and contraction throughout the evolution of the Rhabdoviridae .
RNA viruses are among the most structurally and ecologically diverse of all life forms [1] . Their genomes may consist of positive ( + ) sense , negative ( - ) sense or ambi-sense single-stranded ( ss ) RNA , or double-stranded ( ds ) RNA , and may take the form of a single or multiple segments that are packaged in single or multiple particles . RNA viruses also employ a plethora of strategies for replication and gene expression , and encode a vast array of structural and non-structural proteins , many of which are unique and have multiple , highly specialized functions [2] . Despite their diversity , RNA virus genomes are ubiquitously small , averaging only 10 kb , and with a maximum size of ~32 kb for some members of the order Nidovirales [3 , 4] . This size limitation has been linked to high mutation rates ( a mean rate of ~1 mutation /genome /replication ) due to replication with an error-prone RNA-dependent RNA polymerase that lacks proofreading capability [5 , 6] . High error rates are thought to limit genome sizes because , as size increases , the number of deleterious mutations also increases to levels beyond which reproduction of the fittest variant cannot be guaranteed [7 , 8] . Due to this fundamental evolutionary constraint , RNA viruses have employed various mechanisms of genome compression , such as the use of alternative or overlapping open reading frames ( ORFs ) and the evolution of multiple functions for individual proteins [4 , 7 , 9] . For some RNA viruses , increases in genome size have been associated with increases in the size of replicative proteins [10] and the presence of helicase and proof-reading exonuclease domains [3 , 11–13] . However , the mechanisms and evolutionary context that would favour increased genome size and complexity , given constraints on replication efficiency , are currently unknown [3 , 4] . The Rhabdoviridae is one of the most ecologically diverse families of RNA viruses . Rhabdoviruses have been identified in a very wide range of plants and animals , including mammals , birds , reptiles , and fish with many transmitted by arthropod vectors [14 , 15] . The family includes rabies virus ( RABV ) , which causes over 25 , 000 human deaths annually [16] , vesicular stomatitis Indiana virus ( VSIV ) , which has served as an important model for the study of many aspects of mammalian virus replication and virus-host interactions , and many other important pathogens of humans , livestock , farmed aquatic animals and food crops . The non-segmented [–] ssRNA rhabdovirus genome is packaged within a characteristic bullet- or rod-shaped particle comprising five structural proteins—the nucleoprotein ( N ) , polymerase-associated phosphoprotein ( P ) , matrix protein ( M ) , glycoprotein ( G ) and RNA-dependent RNA polymerase ( L ) [17] . The genome features partially complementary , untranslated leader ( l ) and trailer ( t ) sequences and five ORFs arranged in the order 3’-N-P-M-G-L-5’ . Each ORF is flanked by relatively conserved transcription initiation ( TI ) and transcription termination/polyadenylation ( TTP ) sequences which orchestrate expression of the five corresponding capped and polyadenylated mRNAs [17] . Rhabdovirus genomes may also contain additional ORFs encoding putative proteins , which are mostly of unknown function . These may occur as alternative or overlapping ORFs within the major structural protein genes or as independent ORFs flanked by TI or TTP sequences in the regions between the structural protein genes [15] , some of which appear to have arisen by gene duplication [15 , 18–22] . Here we undertake the first large-scale analysis of the evolution of genome size and complexity in a family of [–] ssRNA viruses . We demonstrate that remarkable changes in genome size and complexity have occurred in rhabdoviruses in a clade-specific manner , primarily by extension and insertion of additional transcriptional units in the structural protein gene junctions , followed by occasional losses . We also show that rhabdoviruses have evolved a large number of accessory proteins and that the use of non-canonical gene expression strategies appears to be common , particularly amongst vector-borne rhabdoviruses .
Our data set comprised the complete or near-complete genome sequences of 99 animal rhabdoviruses , including 45 viruses isolated from various vertebrates and arthropods for which we determined the sequences de novo ( S1 Table ) . Incomplete genomes lacked only the extreme terminal sequences . All rhabdovirus genomes contained the five canonical structural protein genes ( N , P , M , G and L ) ; however , there was remarkable diversity in the number and location of other long ORFs . Across the data set , we identified 179 additional ORFs ≥180 nt in length of which 142 shared no detectable protein sequence similarity with any other protein in our data set or with those in public databases ( S2 Table ) . These additional ORFs were located either within the structural protein genes or in additional transcriptional units located in regions between these genes ( Fig . 1 ) . The additional transcriptional units were annotated by using relatively conserved TI and TTP motifs . The core TI sequence ( UUGU ) was conserved with some minor variations ( CUGU , UUGC , UUGA , UCGU , UGAU ) employed in some viruses . The TTP motif G[U]7 was also conserved , with the variation A[U]7 occurring only in several genes of one virus ( CHOV ) . Due to the large number and diversity of additional ORFs , we adopted a standard nomenclature that does not necessarily reflect structural homology . Unless previously assigned a distinctive name ( e . g . , BEFV GNS , α1 , α2 , β and γ proteins ) , all ORFs ≥180 nt were assigned names according to the following rules: i ) each additional transcriptional unit was designated U ( unknown ) followed by a number as they appeared in order in the genome presented in positive polarity ( i . e . , U1 , U2 , U3 , etc ) ; ii ) the first ORF within each transcriptional unit was assigned the same designation as the transcriptional unit; and iii ) each subsequent ORF within any transcriptional unit ( alternative , overlapping or consecutive ) was designated by letter ( i . e . , U1x , U1y , U1z ) ( S2 Table ) . Alternative ORFs are defined here as those which occur in a different frame within another longer ORF; overlapping ORFs are alternative ORFs which extend beyond the end of the primary ORF; and consecutive ORFs are those which do not overlap but follow consecutively within the same transcriptional unit . The arbitrary cut-off of ≥180 nt ( ≥60 aa ) was selected on the basis that two small basic proteins of 55 and 65 amino acids ( C and C’ ) have been shown to be expressed from an alternative ORF within the VSIV P gene [23 , 24] . These are the smallest known rhabdovirus proteins . To determine the evolutionary history of the rhabdoviruses studied here , we inferred a phylogenetic tree using conserved regions of the L protein of all 99 viruses in our data set as well as the recently described North Creek virus ( NORCV ) [25 , 26] ( Fig . 2 ) . All but two of these 100 rhabdoviruses ( NORCV and MOUV ) clustered into 17 well-supported monophyletic groups ( bootstrap proportion [BSP] ≥ 85 ) ; however , many of the deeper nodes were unresolved throughout the phylogeny . Eight of the well-supported clades corresponded to the eight established genera ( Lyssavirus , Vesiculovirus , Perhabdovirus , Sigmavirus , Ephemerovirus , Tibrovirus , Tupavirus and Sprivivirus ) and we assigned a further seven clades as proposed new genera ( Almendravirus , Bahiavirus , Curiovirus , Hapavirus , Ledantevirus , Sawgravirus and Sripuvirus ) . The taxonomic assignment of the two remaining clades was considered to be ambiguous ( S1 Table ) . For simplicity of expression we refer here to all as ‘genera’ , whether existing or proposed , but we recognise that taxonomic proposals require consideration and ratification by the International Committee on Taxonomy of Viruses ( ICTV ) . Although the analysis was limited by the availability of single isolates of most viruses , apparent structure by geographic location or reservoir host was not observed in the phylogeny . However , multiple genera appeared to be primarily associated with bats ( i . e . , ledanteviruses , lyssaviruses ) , fish ( i . e . , perhabdoviruses , spriviviruses ) or ungulates ( i . e . , ephemeroviruses , tibroviruses , vesiculoviruses ) . Vector-borne rhabdoviruses were present in 12 of the 17 groups , dominating the dimarhabdovirus supergroup , but were largely absent from clades associated with bats ( Lyssavirus ) , flies ( Sigmavirus ) and fish ( Perhabdovirus , Sprivivirus ) ( Fig . 2 ) . The exception to this trend was the Tupavirus clade , which comprised viruses that have not yet been associated with a vector species , and for which little is known about their ecology or distribution . Each of the seven newly proposed rhabdovirus genera formed an independent , well-supported monophyletic group in the L protein phylogeny ( BSP ≥ 85 ) , and comprised viruses with similar genome organization ( Fig . 1; Fig . 2 ) . In several instances , viruses clustered closely with other members of a genus , yet we considered them to be unassigned species due to major differences in genomic architecture ( see below ) . For example , the newly proposed genus Curiovirus comprises a monophyletic group of four viruses isolated from biting midges ( Culicoides sp . ) , sandflies ( Lutzomyia spp . ) and mosquitoes ( Coqillettidia and Trichoprosopon spp . ) from the forests of South America and the Caribbean ( S1 Table ) . The genomes of CURV , IRIRV , RBUV and ITAV all have one or more ORFs located between the M and G genes , and the G and L genes . In contrast , the closely related ARUV and INHV lack additional genes between the M and G and for this reason we have excluded them from the genus Curiovirus at this time . We also recognize the previous suggestion that CURV and ITAV should be assigned to a new genus for which the name Bracorhabdovirus ( Brazilian Amazonian Culicoides rhabdoviruses ) was proposed [27] . However , our analysis clearly indicates that this monophyletic group has a broader host range and geographic distribution than this regionally-derived name suggests . Five of the novel viruses ( comprising four putative new species ) identified in this study were assigned to established genera . Two of these , KOOLV and YATV , clustered within the existing Ephemerovirus clade , ( BSP ≥ 85 ) and possessed the characteristic genome organization of ephemeroviruses , including a non-structural glycoprotein gene ( GNS ) followed by a viroporin ( α1 ) and several other small proteins ( Fig . 1; Fig . 2 ) . Similarly , two novel viruses isolated from biting midges ( Culicoides insignis ) , SWBV and BAV , clustered within the genus Tibrovirus ( BSP ≥ 85 ) and exhibited the conserved N-P-M-U1-U2-G-U3-L genome organisation ( Fig . 1; Fig . 2; S1 Table ) . SWBV was assigned as a new species ( Sweetwater Branch virus ) , but BAV is closely related to TIBV and may be regarded as the same species ( Tibrogargan virus ) . Finally , a novel tupavirus ( KLAV ) identified from two species of vole ( Microtus and Clethrionomys spp . ) , clustered with the TUPV and DURV clade in the L protein phylogeny ( Fig . 2; S1 Table ) . A more detailed rationale for the assignment of viruses to existing and proposed new genera is provided as supplementary text . We identified a 48 . 5% variation in genome size from the smallest genome ( FUKV , Ledantevirus; 10 , 863 nt ) to the largest in our data set ( KOOLV , Ephemerovirus; 16 , 133 nt ) . All genomes , including those for which extreme terminal sequences were unresolved , appeared to fall within this range . Variations in genome size were associated with: i ) variation in the length of intergenic regions ( IGRs ) between transcriptional units; ii ) variation in the length of 3’ and 5’ untranslated regions ( UTRs ) within individual transcriptional units; iii ) the presence of additional transcriptional units containing long ORFs; and iv ) the presence of overlapping or consecutive long ORFs within individual transcriptional units . An examination of genome size across the phylogeny revealed a general trend towards larger genomes in the lower third of the tree , which is comprised of the hapaviruses , curioviruses , tibroviruses and ephemeroviruses , as well as several unassigned viruses ( S1 Fig . ) . Although this may indicate that an enhanced capacity for genome expansion is a property specific to this group , variation in genome size can also be observed between viruses in the majority of genera in the data set . Several clade-specific patterns were evident when the lengths of the transcriptional units and IGRs were compared within and between rhabdovirus genera ( Table 1 ) . Ledantevirus genomes were smallest on average ( 1 . 75 × the length of the L ) whereas ephemeroviruses genomes were the largest ( 2 . 37 × the length of the L , Table 1 ) . Interestingly , although substantial variation in the length of gene junctions was observed in several genera ( including ephemeroviruses and lyssaviruses ) , most variation in genome size occurred as the result of the presence of new , non-canonical ORFs in the regions between the structural protein genes ( Table 1 ) . Although new ORFs were observed in each IGR across the phylogeny ( N-P , P-M , M-G and G-L ) their location was primarily restricted to a single IGR within each genus . For example , while hapavirus genome expansion occurred primarily in the P-M junction , genome expansion in the ephemeroviruses occurred at the G-L junction and tibrovirus and curiovirus genomes contained additional ORFs primarily in the M-G junction ( Table 1 ) . This suggests that once a new ORF arises at a particular gene junction within a lineage , further expansion is more likely to continue at the same gene junction , rather than begin anew elsewhere in the genome . Whilst the genome architecture in some viruses was highly compact , others featured long stretches of sequence with non-ascribed function that occurred primarily as 5’UTRs and 3’UTRs within transcriptional units ( Fig . 3 ) . The proportion of untranslated sequences within or between transcriptional units ranged from 0 . 5% ( FUKV; 58 nt ) to 10 . 6% ( WCBV; 1290 nt ) and did not correlate with genome size . Furthermore , although all lyssaviruses ( such as WCBV ) featured a high proportion of untranslated sequences ( primarily evident as a very long 3’UTR in the G gene ) , there was no consistent association between the proportion of untranslated sequences and genus assignment ( Fig . 3 ) . For example , in the genus Hapavirus , the proportion of untranslated sequences in the two largest genomes varied from 1 . 1% ( NGAV ) to 6 . 4% ( LJV ) . Similarly , in the genus Ephemerovirus the proportion of untranslated sequences varied from 1 . 2% in the smallest genome ( YATV ) to 9 . 6% in the largest genome ( KOOLV ) . The presence of long stretches of untranslated sequence , which occurred primarily within transcriptional units , suggests these regions may be functional . However , it is unclear at this time why they are present in some rhabdoviruses and not in others . Gene duplication . Previous studies have provided evidence of gene duplication in the Rhabdoviridae , involving the G and GNS genes [18 , 21] and the β and γ genes [22] in the ephemeroviruses , and the U1 , U2 and U3 genes in the hapaviruses FLAV and WONV [15 , 19 , 20] . To identify further examples of gene duplication , we conducted a BLAST analysis of all proteins in our database ( E-value <1e-3 ) and used ClustalX alignments to confirm sequence similarity . By this analysis , ORFs located between the P and M genes of most hapaviruses encode proteins which share detectable sequence similarity . This family of homologous P-M intergenic region proteins ( PMIPs ) includes the U1 , U2 and U3 proteins of LJV , WONV , PCV , ORV , LJAV , MANV , MQOV , FLAV , HPV , KAMV and MOSV ( S2 Fig . and S3 Fig . ) , as well as the U1x proteins of MANV and GLOV which are encoded in ORFs overlapping their respective U1 ORFs ( S4 Fig . ) . Although pairwise alignments provide clear evidence for homology , the hapavirus PMIPs share generally low levels of sequence identity and no universally conserved motifs , indicating considerable structural and functional divergence from their ancestral homolog . Proteins encoded in the P-M region in other hapaviruses ( i . e . , JOIV U1 , NGAV U1 , U1x and NGAV U2 ) failed to display significant similarity with the PMIPs or evidence of gene duplication but this may be due to further structural divergence . Additional evidence of gene duplication included the U2 and U3 proteins of JOIV ( encoded in ORFs located between the G and L genes ) , and the N-terminal regions of the P proteins and the upstream U1 accessory proteins of the sripuviruses CHOV and SMV , each of which share significant sequence similarity ( S5 Fig . ) . These data suggest that the U1 protein of the sripuviruses originated from a duplication of the P gene , with the downstream copy of the gene retaining the parental function . Similarly , in the curioviruses there is extensive amino acid sequence similarity between the U3 proteins of CURV and IRIRV and the N-terminal region of the G proteins , suggesting evolution of U3 through partial duplication of the G gene , which lies immediately downstream . Putative accessory genes were found to be abundant and varied greatly in number and location in each genome ( Fig . 1 ) . A complete list of ORFs >180 nt is annotated in S2 Table . In most cases , homology searches detected no significant amino acid sequence identity with entries in GenBank . However , various rhabdovirus accessory gene families were identified based on amino acid sequence identity in our custom BLAST searches , or common structural characteristics . Viroporins . Viroporins are small hydrophobic proteins that oligomerize in host cell membranes to form hydrophilic pores , disrupting various cellular processes and promoting virus replication [28] . ORFs encoding viroporin-like proteins were found in more than one-third of the rhabdoviruses in the data set , either as overlapping or consecutive ORFs within the G gene , or in additional transcriptional units following the G ( or GNS ) gene ( Fig . 1 ) . ORFs encoding putative viroporins were evident in the genomes of all ephemeroviruses , tibroviruses , hapaviruses , bahiaviruses , almendraviruses and curioviruses , as well as the unassigned species ARUV and INHV ( Fig . 4 ) . Several of these proteins have been identified previously [19 , 22 , 29–35] . Like the BEFV α1 protein for which viroporin activity has been confirmed experimentally , these proteins have the structure characteristics of class IA viroporins , including a central transmembrane and a highly basic C-terminal domain . However , although located in similar positions in the genomes , they are generally too divergent in sequence to establish orthology [22 , 36] . Other small transmembrane proteins . Small proteins with a predicted central transmembrane domain but lacking other characteristics of class 1A viroporins were identified in several other rhabdoviruses ( S6 Fig . ; S2 Table ) . Transmembrane proteins with an N-terminal ectodomain are encoded in the Gx ORF of sripuviruses and the U3 ORF of one curiovirus ( RBUV ) . However , in other curioviruses ( CURV and IRIRV ) , transmembrane proteins are encoded in the U2 ORF and are predicted to have the reverse membrane topology to the RBUV U3 protein . Sequence alignments further suggest these proteins are not orthologous . There is also a small double-membrane spanning protein with a predicted short ectodomain loop encoded in an alternative ORF in the FUKV M gene that is not present in other ledanteviruses . Other small hydrophobic ( SH ) proteins . Small highly hydrophobic proteins ( 6 . 8–10 . 8 kD ) lacking predicted transmembrane domains are encoded in all tupaviruses ( as independent transcriptional units following the M gene ) and sripuviruses ( as overlapping ORFs within the M gene ) ( S7 Fig . ; S2 Table ) . All have similar hydropathy profiles with a highly hydrophilic N-terminal domain extending to the centre of the sequence , but sequence identity indicative of orthology is restricted to closely-related viruses . Several of these SH proteins have been identified previously but their function remains unknown [37–40] . Large class I transmembrane glycoproteins . All ephemeroviruses encode a class I transmembrane glycoprotein ( GNS ) in the ORF following the G gene [18 , 21 , 30 , 31] . NGAV ( assigned to the proposed new genus Hapavirus ) also encodes a GNS protein with similar structural characteristics [35] . However , as we found no evidence to support recombination between NGAV and any ephemerovirus , the NGAV GNS gene is likely to have arisen by an independent duplication event of the upstream G gene with which it shares amino acid sequence identity . ORF U1 immediately following the MCOV G gene ( genus Hapavirus ) also encodes a large class I transmembrane glycoprotein but lacks the set of conserved cysteine residues that are characteristic of G and GNS proteins , and our homology searches failed to identify similarity with any known protein ( S8 Fig . ) . Other genus-specific accessory gene families . Orthologous sets of accessory genes occur in genus-specific patterns in each of the structural protein gene junctions ( Fig . 1; S2 Table ) . In addition to the hapavirus PMIP genes , these include genes in the N-P junction of sripuviruses CHOV and SMV ( U1 proteins ) , the M-G junction of curioviruses ( U1 and U1x proteins ) and tibroviruses ( U1 and U2 proteins ) , and the G-L junction of curioviruses ( U3x proteins ) and ephemeroviruses ( α2 , β , γ and δ proteins ) ( S9 Fig . to S11 Fig . ) . Some of these orthologous gene sets have been described previously [15] . Most encode proteins without remarkable structural characteristics and of unknown function ( S2 Table ) . Several general architectural patterns in the arrangement of ORFs were evident , implicating several mechanisms of non-canonical gene expression . Non-cannonical expression mechanisms are used commonly in other families of RNA viruses to increase genome complexity without significantly increasing genome size [41] . The patterns we observed in this data set were associated with consecutive , overlapping of alternative ORFs within individual transcriptional units . Consecutive ORFs and TURBS motifs . Consecutive long ORFs with termination and initiation codons that are either overlapping ( e . g . , UAAUG ) or separated by a short stretch of nucleotides were common in several groups of rhabdoviruses ( Fig . 5 ) . As previously observed for FLAV , this ‘stop-start’ arrangement is commonly preceded by a ‘termination upstream ribosome-binding site’ ( TURBS ) , which contains a short sequence motif that is complementary to the loop region of helix 26 of 18S ribosomal RNA [19 , 41] . The TURBS may also contain flanking anti-complementary sequence motifs that are predicted to form a stem-loop structure . This arrangement was found in the M transcriptional unit in the sripuviruses , the G transcriptional unit of several hapaviruses ( FLAV , HPV , MANV , MQOV , KAMV , MOSV and GLOV ) and the transcriptional unit between the P and M genes of GLOV . The ‘stop-start’ arrangement also occurs in the transcriptional unit between the G and L genes of ARUV , allowing expression of the U2 ORF , but in this case the TURBS appears to be further upstream of the stop-start site . Finally , the α gene transcriptional unit in most ephemeroviruses contains consecutive ORFs encoding a viroporin ( α1 ) and a second protein of unknown function ( α2 ) . In KOTV , a TUBRS is evident upstream of the stop-start site but in other ephemeroviruses the TURBS appears to be more cryptic . Overlapping ORFs and ribosomal-frame shift ( RFS ) sites . Overlapping ORFs are common in rhabdovirus genomes and represent a second common architectural arrangement requiring non-canonical gene expression . Overlapping ORFs occur within the N transcriptional unit ( WONV , ORV , PCV , MCOV , MANV ) , the G transcriptional unit ( WONV , ORV , PCV , BGV , HARV ) or within additional transcriptional units between the P and M genes ( MANV , NGAV ) or the M and G genes ( CURV , IRIRV , RBUV ) . Expression of the second ORFs in these arrangements would require either internal initiation in an alternative reading frame or another mechanism such as RNA editing or a ribosomal frame-shift ( RFS ) to extend the first ORF . Use of alternative initiation codons has been reported in the M and P genes of VSV and the P gene of RABV , and RNA editing has been described in the P gene of paramyxoviruses [23 , 42–45] . Although not described previously in mononegaviruses , potential RFS sites were identified in some of these rhabdovirus gene overlap regions , featuring the ‘slippery’ sequence motifs UARUUUUUUCA ( BGV , HARV , MSV ) or CCNUUUUUUGA ( WONV , ORV , PCV ) followed by a predicted stem-loop structure ( S12 Fig . ) . These sequence motifs and associated stem-loop structures most closely resemble the-1 RFS that allows expression of gag-pol in HIV-1 and other lentiviruses [41 , 46] . Alternative ORFs and leaky ribosomal scanning . The third architectural arrangement involves the use of alternative ORFs within a longer ORF . This arrangement was described previously in VSIV , in which two small basic proteins of 55 and 65 amino acids ( C and C’ ) are expressed from an alternative ORF within the P gene [23 , 24] . On this basis , we scanned the rhabdovirus genome data set for alternative ORFs of various size ranges and observed that the frequency varied from ~2 . 3/genome for ORFs in the range of 90–150 nt ( 30–50 amino acids ) to ~8 . 6/genome for range 150–210 nt ( 30–70 amino acids ) ( Fig . 6 ) . Alternative ORFs ≥60 amino acids occurred in each of the structural protein genes ( N , P , M , G and L ) and in the additional transcriptional units between the P and M genes . They were most common in the P and least common in the M genes . As observed in other viruses , expression of these alternative ORFs could occur by leaky ribosomal scanning , allowing initiation of transcription by a proportion of ribosomes on the alternative start codon [41] . Although , it is not known which ( if any ) of these alternative ORFs are expressed , several factors are likely to be important in determining the probability and level of expression: i ) the Kozak contexts of the first and alternative initiation codons; ii ) the length of the alternative ORF ( longer ORFs are less likely to occur by chance ) ; iii ) the location of the alternative ORF ( distally located ORFs are less likely to be expressed in long transcripts ) ; and iv ) the expression level of the transcript ( L gene transcripts are likely to be the least abundant ) . For example , short ORFs with initiation codons in poor Kozak context at the distal end of the L gene are not likely to be expressed at significant levels , if at all . However , in some cases , closely related viruses were found to contain alternative ORFs at the same genome location , with initiation codons in good context and encoding predicted polypeptides with high levels of sequence identity ( S2 Table ) . Such arrangements occurred in the N genes of HPV and FLAV , the P genes of MANV and MQOV , the U2 and M genes of KAMV and MOSV , and near the start of the G genes of the sripuviruses ( NIAV , SRIV , CHOV and SMV ) ; these proteins are considered very likely to be both expressed and functional .
We have conducted a detailed analysis of the structural organisation and genome evolution of a family of negative-sense RNA viruses—the Rhabdoviridae . Previous studies have surveyed known rhabdoviruses for biological and genomic diversity , revealed phylogenetic relationships , and considered factors that may have determined their rates of evolution [14 , 15 , 47 , 48] . In this study , we greatly expanded the repertoire of rhabdovirus genome sequences , which demonstrate extensive variation in genome size and complexity , allowing the assignment of seven proposed new genera . We also identified patterns of accessory gene evolution and expression , and showed that changes in rhabdovirus genome length and composition have occurred throughout the evolutionary history of the family , primarily through the generation and loss of new transcriptional units . This observation is especially striking given the obvious constraints on viral genome size [7] . The most remarkable aspect of this analysis is the number and variety of additional ORFs identified in rhabdovirus genomes , which provides a very different perspective of the family and its evolution than had been obtained from studies of the traditional prototype members ( VSIV and RABV ) . As many of these ORFs occur as additional transcriptional units complete with conserved transcriptional control sequences , there is a high likelihood that they would be expressed in infected cells . Expression of ORFs located in additional transcriptional units has been demonstrated previously for several ephemeroviruses and for the hapavirus WONV [18 , 21 , 30 , 31 , 36 , 49] . Others occur as either alternative or overlapping ORFs . Further studies are required to determine which of these ORFs may be expressed , but we suggest that expression is likely when both the encoded amino acid sequence and the translational context are conserved in related species . Notably , very few of the additional ORFs detected in this analysis encode proteins with identifiable sequence similarity to other known proteins . Sequence similarity , when detected , occurred only between closely related viruses assigned to a genus and , although some accessory protein families were identified , these were more commonly related by shared structural characteristics , such as charged or transmembrane domains , than by sequence . This has been observed previously for so-called orphan ( ‘ORFan’ ) proteins in other viruses and bacteria . It has been suggested that the uniqueness of orphan proteins , or their restriction to a single species or genus , is the result of creation de novo , rather than by recombination or lateral gene transfer , and that they play an ‘accessory’ role in viral pathogenicity or transmission instead of having functions in virion structure or replication [50–52] . It has also been observed that many orphan proteins are predicted to be highly disordered in structure or , when ordered , structural resolution has revealed unique folds [50] . As such , future determination of the biological activities of the plethora of novel proteins identified here will require functional studies that may well provide important insights into aspects of infection and immunity as well as fundamental cellular processes and pathways . Substantial variation in genome size and complexity was also observed in many rhabdovirus genera , suggesting that the length of the genome is not heavily constrained in all members of the family . Indeed , the presence of new ORFs and/or very long stretches of non-coding sequence within or between transcriptional units was noted frequently . Previous observations have demonstrated that foreign genes of up to ~6 kb can be inserted into the VSIV genome without significant disruption to viral replication in vitro [53 , 54] . Expanded VSIV genomes were morphologically similar but proportionally longer than wild-type viruses , suggesting that the unique morphology of the rhabdovirus particle may more readily accommodate genome expansion than other virion structures . A significant body of evidence suggests that genome size in RNA viruses is likely to be constrained by low replication fidelity [7 , 8] , and a relationship between genome size and error rate has been observed in a diverse array of organisms [55] . However , if the genome sizes of rhabdoviruses are constrained by selective pressures other than ( or in addition to ) those imposed by the background mutation rate , genome expansion may not require a concomitant reduction in polymerase error rates . As the mutation rate of rhabdoviruses has only been determined experimentally for VSIV thus far ( ~6 × 10–6 subs/nucleotide/replication ) , it is impossible to assess whether the increases in genome size observed here have been associated with concomitant reductions in mutation rate [48] . It is also striking that while some rhabdovirus genomes appear to have undergone major changes in length and complexity , others contain only the 3’ and 5’ promoter regions and five canonical transcriptional units with minimal 5’ and 3’UTRs . This suggests that the acquisition and loss of new genes and intergenic regions may be a regular feature of rhabdovirus evolution . Previous studies of RNA viruses have concluded that constraints on genome size imposed by polymerase error have led to various strategies to minimize genome size while increasing functional complexity , such as gene overlaps and protein multi-functionality [9 , 56] . Given these size constraints , it is unclear why long non-coding regions would arise both within and between transcriptional units and be maintained throughout the evolution of some rhabdovirus genera . It has been known for many years that a long 3’-UTR of unknown function ( ψ region ) in the G gene of RABV is unnecessary for efficient replication in cell culture or in mice , but may play a role in neuroinvasion [57–59] . Indeed , the retention of similar ψ regions in all lyssaviruses and the existence of long UTRs and IGRs in other rhabdoviruses suggests that they must provide some fitness advantage in vivo , such as stabilising RNA secondary structure , serving as a source of , or targets for , micro RNAs , or attenuating transcription of downstream genes to achieve the most effective balance of gene expression . Indeed , an analysis of patterns and rates of sequence evolution in the Rhabdoviridae and other families in the Mononegavirales revealed that , although non-coding regions are less conserved than those that encode proteins , their evolutionary rates are associated with relative genomic position , suggesting that they impact on gene expression [60] . Additional ORFs and non-coding sequences occurred at all junctions of the canonical structural protein genes ( i . e . , N-P , P-M , M-G , and G-L ) , although there was variation in both the frequency of insertion and the extent of expansion . Notably , insertions at the N-P junction are rare , with a single additional ORF present in the closely related sripuviruses CHOV and SMV , and short overlapping ORFs present within the N gene transcriptional unit in some hapaviruses . It has been reported previously in a study of VSIV recombinants that only the N-P gene junction was refractory to the stable expression of an inserted transcriptional unit , and resulted in a virus with significantly reduced replication efficiency [61] . In contrast , transcriptional units inserted at other gene junctions were stably expressed , maintained through repeated passages and had no effect on replication efficiency . As the insertion of additional transcriptional units attenuates expression levels of all downstream genes , this may be associated with the importance of maintaining precise control of N and P protein ratios in infected cells to ensure efficient switching between the transcription and replication modes of the ribonucleoprotein complex [62 , 63] . The relationships , locations and contexts of additional ORFs in various viruses lead us to propose a general model for rhabdovirus genome plasticity , which can account for both gains and losses in genome size and complexity ( Fig . 7 ) . In each of these viruses , small ORFs of various lengths occur within most transcriptional units; and although only those ≥180 nt have been catalogued here , there are numerous other smaller ORFs throughout most genomes . It is reasonable to assume that , although the polypeptides encoded in many of these ORFs may not be expressed at all during infection , some may be expressed through leaky ribosomal scanning . These are likely to represent a rich genetic resource for the evolution of new functional genes in RNA viruses [4] , triggering the rapid evolution of highly specialised functions . Contemporarily , the evolution of a suitable Kozak context , TURBS motifs and ribosomal frame-shift sites would allow optimal expression within the parental transcriptional unit . Ultimately , these new ORFs may become uncoupled from the parental gene through gene ( sequence ) duplication [18] . As observed previously , this process would allow unconstrained evolution of the new ORF and loss of the redundant copy of the parental ORF [4 , 64] . Alternatively , new genes may also evolve independently of existing ORFs . In some rhabdoviruses in our data set , very long non-coding regions ( up to 749 nt ) were present either within or between transcriptional units that could serve as a resource to spawn genes de novo in the absence of the evolutionary constraints imposed on alternative or overlapping ORFs . This is most likely to occur when ORFs are present in transcribed non-coding regions ( UTRs ) such as the ψ region of WCBV in which , uniquely amongst lyssaviruses , an ORF of 180 nt has been identified [65] . The creation of new genes de novo in non-transcribed IGRs , such as those present in the G-L gene junctions of LJV , KOTV and KOOLV , almost certainly would require prior or simultaneous evolution of new or modified transcriptional control sequences to allow their expression . We recognise that other mechanisms of genome expansion are also possible . In Central American isolates of VSIV , for example , imprecise reiterative insertions of up to 300 nt in the 5’-UTR of the G-gene ( variations of 3’-UUUUUAA-5’ ) have been attributed to non-templated extension by polymerase stutter at the TTP sequence [66 , 67] . Although homologous recombination appears to be very rare in mononegaviruses [68] , and we found no evidence of lateral gene transfer , we cannot exclude their involvement in rhabdovirus genome expansion . It is also evident that although there is an overall trend toward an expansion of genome size and complexity in the rhabdoviruses , gene loss is also likely to have occurred periodically throughout the evolution of the family . For example , the ephemerovirus γ proteins appear to have been lost in ARV and OBOV , and the hapavirus PMIPs are entirely absent only from MCOV ( Fig . 1 ) . Although our data suggests that gene gain is a more frequent process than gene loss , we acknowledge that , if loss is very frequent , we might not be able to observe it given the available data . This may be resolved in the future with the acquisition of significantly more genomes sampled more closely in time . Indeed , as defective-interfering particles are known to occur commonly in rhabdoviruses , a mechanism for purging redundant sequences appears to be readily available [69–71] . Nevertheless , it is evident that a remarkable capacity for genomic plasticity through the gain and loss of accessory functions has been a central theme of rhabdovirus evolution . Although our analysis was limited to the Rhabdoviridae , similar mechanisms of genome expansion appear to occur in other families of non-segmented ( - ) ssRNA viruses ( Mononegavirales ) . For example , amongst the Paramyxoviridae genome length varies by 46 . 5% from human metapneumovirus ( 13 , 113 nt ) to Beilong virus ( 19 , 212 nt ) , and paramyxoviruses also contain novel accessory genes in transcriptional units inserted at various gene junctions [72] . The apparent propensity for genome expansion in mononegaviruses may be due to their discontinuous transcription strategy which generates multiple viral mRNAs . Sequence insertions within and between the individual transcriptional units of mononegaviruses are less likely to disrupt gene expression than in ( + ) ssRNA viruses in which the genome commonly encodes a single polyprotein which is processed post-translationally . Finally , this study has also provided an important advance in rhabdovirus taxonomy , allowing the assignment of six new species to existing genera and the assignment of 37 species to seven proposed new genera as well as the identification of six new unassigned species . There are currently no formal criteria for genus demarcation in rhabdoviruses . A system of genetic classification ( DEmARC ) that allows demarcation of viral taxa based on pairwise evolutionary distances has been proposed and , for picornaviruses , was shown to be comparable to expert-based taxonomic classification [73 , 74] . However , the application of this approach to the Rhabdoviridae would likely require a larger set of sequenced genomes at lower taxonomic levels [75] , and would be compromised by extensive rate variation among lineages ( as this leads to biases in genetic distance measurements ) . In the taxonomy of higher organisms , to be descriptively useful , a genus should be monophyletic , reasonably compact , and ecologically , morphologically , or biogeographically distinct [76] . Our assignment of new genera in the Rhabdoviridae has been based primarily on the identification of well-supported monophyletic groups using unambiguously aligned regions of the L gene , together with a consideration of common features of genome organisation and known aspects of viral ecology . Genome organisation has proven here to be a useful taxonomic marker as similar arrangements of accessory genes and other conserved elements of genome architecture appear to be the result of significant evolutionary events that provide resolution between the family and species levels . For some of the new genera , host and/or vector associations have also been relatively informative but in many cases , only single isolates of a species are available and else little is known of their ecology . It is likely that the proposed assignments of viruses to genera and the placement of the proposed unassigned species will evolve into a more complete taxonomic description as more viruses are discovered and as ecological data accumulates .
Details of the viruses included in this study , including taxonomic status , sources and dates of isolation , and GenBank accession numbers of genome sequences are given in S1 Table . All but three viruses sequenced in this study were obtained from the World Reference Center for Emerging Viruses and Arboviruses ( WRCEVA ) , located at the University of Texas Medical Branch , Galveston . Of the remaining viruses , FUKV and KOOLV were obtained from the collection held at the CSIRO Australian Animal Health Laboratory , Geelong , and JOIV was obtained from the QIMR collection held at the Queensland University of Technology , Brisbane , and kindly provided by Dr John Aaskov . Viruses sequenced in this study were prepared as described previously [37] . With the exception of HPV , ITAV , CURV , GLOV , INHV , NMV , MEBV , YATV , LDV , GARV , CNTV , IRIRV , RBUV , BARV , LJAV , KEUV , MCOV , SMV , CHOV , PCV and BAV , which were sequenced directly from infected suckling mouse brain , viruses were sequenced from viral preparations grown in BHK-BSR , C6/36 or Vero cells monolayers . Sequencing was performed using either the Illumina HiSeq or MiSeq platforms . Viral RNA was fragmented by incubation at 94°C for 8 min in 19 . 5 l of fragmentation buffer ( Illumina 15016648 ) . A sequencing library was prepared from the sample RNA using an Illumina TruSeq RNA v2 kit following the manufacturer’s protocol . Samples were sequenced using the 2 × 50 paired-end protocol . Reads in fastq format were quality-filtered and any adapter sequences were removed using Trimmomatic software [77] . The de novo assembly program ABySS [78] was used to assemble the reads into contigs using several different sets of reads and k values from 20 to 40 . The longest contigs were selected and reads were mapped back to the contigs using Bowtie 2 [79] and visualized with the Integrated Genomics Viewer [80] to verify that the assembled contigs were correct . Total reads ranged from 0 . 5 to 12 million and the percentage of reads mapping to the virus genome in each sample ranged from 0 . 2% to 33% . Details are available upon request . Assembly of full genome sequences was performed as previously described [37] and predicted ORFs >30 amino acids in length were identified across each genome using Geneious 7 . 0 . 6 ( Biomatters Ltd ) . For each non-canonical ORF >60 amino acids in length , we sought to identify putative homologues by first comparing the protein sequence to the complete non-redundant protein sequence database available on GenBank using the BLASTp and PSI-BLAST search algorithms , as well as to the UniProt20 database using the hidden Markov model alignment-based algorithm HHblits[81] . For these searches , we investigated all matches with an E-value <1 . We then created a custom protein database containing all ORFs >60 amino acids in length from our data set ( 648 proteins ) and performed a custom BLAST search to identify homologues within this data set . Here , an E-value of <1e-3 was considered a significant match . Amino acid sequence alignments containing all putative matches to each ORF were then created using Clustal X and evidence of structural and sequence similarity was investigated by visual inspection . Structural predictions for proteins were conducted using Compute pI/MW , SignalP , TMHMM , TmPred , NetNES and NetNGlyc available through the ExPASy Bioinformatics Resource Portal ( http://www . expasy . org/ ) . To quantify the location and extent of variation in genome size in our data set , we compared the average length of each genomic region within and between rhabdovirus genera . For all viruses , we normalized the length of each gene region ( from the TI to TTP sequences , inclusively ) and intergenic region by dividing by the length of the corresponding L gene , which varied least across the data set ( coefficients of variation: N = 0 . 06 , P = 0 . 12 , M = 0 . 09 , G = 0 . 13 , L = 0 . 01 ) . As there was substantial variability in the proportion of the 5’ and 3’ UTRs that were included in the sequence data set , we considered each genome to begin at the first TI sequence and end at the final TTP sequence for this analysis . To infer evolutionary relationships among animal rhabdoviruses , we compiled sequences of the L ( RNA-dependent RNA polymerase ) protein , as this was the most highly conserved protein across the data set . We initially attempted to root the tree using a standard outgroup method . Members of the rhabdovirus genera that infect plants ( i . e . , Cytorhabdovirus and Nucleorhabdovirus ) were excluded as their sequences were highly divergent . We therefore utilized four members of the genus Novirhabdovirus ( Infectious haematopoietic necrosis virus ADB93801; Viral hemorrhagic septicaemia virus BAH57327; Hirame rhabdovirus ACO87999; and Snakehead rhabdovirus NP050585 ) as outgroups . Unfortunately , these novirhabdovirus sequences were also far too divergent ( >>1 amino acid change per site under multiple amino acid substitution models; results available on request ) to establish a reliable rooting for our data set , as three different basal groups were identified using different models of amino acid substitution , although overall tree topologies were similar among substitution models ( results available on request ) . In addition , the use of the novirhabdoviruses as outgroups resulted in excessive numbers of residues being removed following Gblocks pruning ( see below ) . Based on the observation that most known rhabdoviruses are either insect viruses or replicate in insect vectors , it has been reasonably argued that plant and animal rhabdoviruses may have origins in insects [82] . We therefore selected the rooting scheme that best fit this theory . To this end , we choose one of the two basal clades from the novirhabdovirus-rooted tree , comprising viruses isolated from mosquitoes ( i . e . , the almendraviruses ) , as the most divergent group . We then repeated the phylogenetic analysis ( procedure described below ) excluding the novirhabdoviruses and rooting it on the almendraviruses . Importantly , the choice of outgroup did not influence relationships either between or within the major clades demonstrating strong bootstrap support ( BSP ≥ 85 ) . The alignment used for the final tree inference ( i . e . , excluding the novirhabdoviruses ) was comprised of amino acid sequences aligned using the MUSCLE program [83] , with ambiguously aligned regions removed using the Gblocks program with default parameters [84] . This resulted in a final sequence alignment of 100 taxa , 1007 amino acid residues in length . The phylogenetic relationships among these sequences were determined using the maximum likelihood ( ML ) method available in PhyML 3 . 0 [85] employing the WAG+Γ model of amino acid substitution and subtree pruning and regrafting ( SPR ) branch-swapping . The phylogenetic robustness of each node was determined using 1 , 000 bootstrap replicates and nearest-neighbour branch-swapping . | Understanding the patterns and mechanisms of genome evolution is one of the most important , yet least understood , aspects of RNA virus biology . The evolutionary challenge faced by RNA viruses is to maximize functional diversity within severe constraints on genome size . Here we show that rhabdoviruses , a family of RNA viruses that infect hosts as diverse as plants , insects and vertebrates , have an unusual capacity for genomic plasticity . By analysing the complete or near-complete genome sequences of 99 animal rhabdoviruses , we show that genome expansion and contraction has likely occurred frequently throughout the evolution of the family . Genomic plasticity has been associated with the evolution of alternative , overlapping and consecutive ORFs within the major structural protein genes , as well as the insertion and loss of additional ORFs in each gene junction in a clade-specific manner . This has resulted in remarkable diversity in genome organisation and gene expression strategies that is reflective of the broad ecological diversity of rhabdoviruses . We conclude that genomic plasticity in rhabdoviruses may be linked to their discontinuous transcription strategy from the negative-sense single-stranded RNA genome and propose a general model that accounts for both gains and losses in genome size and complexity . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2015 | Evolution of Genome Size and Complexity in the Rhabdoviridae |
Evolution maintains organismal fitness by preserving genomic information . This is widely assumed to involve conservation of specific genomic loci among species . Many genomic encodings are now recognized to integrate small contributions from multiple genomic positions into quantitative dispersed codes , but the evolutionary dynamics of such codes are still poorly understood . Here we show that in yeast , sequences that quantitatively affect nucleosome occupancy evolve under compensatory dynamics that maintain heterogeneous levels of A+T content through spatially coupled A/T-losing and A/T-gaining substitutions . Evolutionary modeling combined with data on yeast polymorphisms supports the idea that these substitution dynamics are a consequence of weak selection . This shows that compensatory evolution , so far believed to affect specific groups of epistatically linked loci like paired RNA bases , is a widespread phenomenon in the yeast genome , affecting the majority of intergenic sequences in it . The model thus derived suggests that compensation is inevitable when evolution conserves quantitative and dispersed genomic functions .
With the complete sequencing of a large number of genomes , and with the rapid progress in the development and application of methodologies for functional annotation of whole genomes [1] , it is becoming evident that our basic concepts of genomic function must be updated . The view of genomes as “bags of genes” is challenged by multiple lines of evidence , such as the extensive transcription of short and long RNAs from a substantial fraction of the genome [2]–[4] , and the identification of a dense grid of enhancers and transcription factor binding sites in regions that could not be previously associated with genes [5] , [6] . Some of the properties of the newly emerging genomic encodings are clearly different from the prototypic example of the triplet genetic code . The direct mapping between genomic positions ( codons ) and function ( peptides ) which is a hallmark of the genetic code does not seem to hold for the majority of the genome . Instead , genomic encodings integrate small contributions from multiple positions to form complex and quantitative outcomes . These types of dispersed encodings may be involved in defining enhancer sequences , maintaining epigenomic switches , affecting widespread transcription , and contributing to chromosome structure and dynamics . The evolutionary implications of these new types of codes are still poorly understood . The classical models in molecular evolution assume fitness to be a function of a single evolving locus . Conservation of the function encoded by such a locus is quantitatively predicted to decrease its rate of evolution . What rates of evolution can be expected when each of the multiple positions have small contributions to some joint quantitative fitness ? Neutral compensatory substitutions were predicted by Kimura 25 years ago [7] to couple substitutions in pairs of interacting protein coding loci . Kimura's concept was that an evolving population trajectory may visit suboptimal fitness levels transiently , thereby invoking an adaptive corrective force that can bring the system back to optimality . Such a process will change the genomic sequence , fixating pairs of compensatory alleles . Kimura's compensatory dynamic may work in any group of loci that are associated with an epistatic ( non linear fitness function ) constraint and was quantified extensively in RNA coding loci where the epistatic coupling of paired loci has a clear structural interpretation [8]–[10] . Another important source of genomic information , transcription factor binding sites , poses evolution with a different type of epistatic constraint by forming a quantitative binding energy landscape that affects gene regulation [11] , [12] . The evolution of binding sites was shown to drive compensatory effects at the single site level [13] and also at the level of binding site clusters ( or enhancers ) [13] , [14] . Studies of enhancer evolution are continuously providing striking examples for plasticity and compensation [15]–[17] , but due to their heterogeneity , it is currently difficult to develop a general understanding of their evolutionary dynamics . A simple experimentally characterized example of a dispersed genomic encoding involves the effect of DNA sequence on nucleosome organization [18] , [19] . In-vitro and in-vivo experiments in yeast [20] , [21] and other species [22]–[24] showed that nucleosomal packaging is correlated with preferential binding of nucleosomes to specific dinucleotide periodicities , and is strongly anti-correlated with A+T content in general and with poly ( A/T ) sequences in particular [20] , [23] , [25] , [26] . The correlation between nucleosome occupancy and the underlying DNA sequence is sufficiently powerful to allow sequence based nucleosome occupancy prediction , but this prediction is not based on a strict requirement for certain nucleotides to appear at precise positions . Rather , information from multiple sequence positions along the 147bp length of the nucleosome contributes to the affinity of nucleosomes to a given sequence and consequently , to the formation of stable or semi-stable nucleosome configurations [27] . The evolution of these sequence determinants thus serves as a test case for the dynamics of dispersed genomic encodings . Analysis of substitution rates in yeast suggested that genomic sequences that are unbound to nucleosomes are evolving slower than genomic sequences that are bound to nucleosomes [20] , [28]–[30] . Whether this is an indication of classical purifying selection on nucleosome encoding sequences , increased abundance of transcription factor ( TF ) binding sites at low nucleosome occupancy loci , or nucleosome-associated mutability , is currently unclear [31] . Here we analyze patterns of divergence and polymorphisms in yeast intergenic sequences to substantiate an extended model of selection on a dispersed genomic encoding . The analysis shows that yeast low nucleosome occupancy sequences have maintained a high A+T content throughout the evolution of the Saccharomyces cerevisiae lineage . Contrary to standard evolutionary models , we show that this conservation was made possible not by pointwise sequence conservation , but by a compensatory coupling of decreased rates of A/T-losing substitutions and increased rates of corrective A/T-gaining substitutions . Theoretical analysis suggests that this type of evolutionary dynamics is largely unavoidable when the genome employs dispersed functional encodings . The evolutionary dynamics we reveal shuffle sequences continuously while preserving their encoded function , creating a dynamic yet balanced process that may be central to the evolution of gene regulation .
The global G+C content of the yeast intergenic genome is about 35% ( Fig 1A ) but there is a significant heterogeneity in the genome local nucleotide composition ( Fig S1 ) . Such heterogeneity must be the consequence of a variable evolutionary process working in G+C poor and G+C rich sequences . Recently it was shown that nucleosome occupancy patterns strongly correlate with local G+C content in yeast [32] . We define high nucleosome occupancy loci as those in the top 21% MNase-seq coverage percentiles in-vivo ( total 540 kbp , Fig 1B ) , and low nucleosome occupancy loci as those in the bottom 14% MNase-seq coverage percentiles in-vivo ( total 350 kbp ) . Overall , the intergenic G+C content at high occupancy sequences ( ∼40% G+C ) is higher than the G+C content of low occupancy sequences ( ∼28% G+C ) . This heterogeneity is even more pronounced when studying the distribution of tri-nucleotides ( Fig 1C , Fig S2 ) , showing A/T tri-nucleotides to be more abundant in low occupancy sequences , and pointing towards additional nucleosome sequence preferences . It was shown before that in-vitro nucleosome occupancy can be robustly predicted from the distribution of 5-mers or even 3-mers in the sequence [21] . This suggests that the functionality and fitness contribution of DNA-encoded nucleosome organization , if such a contribution exists , is dispersed across multiple loci in a quantitative fashion and is not encoded by a strict requirement for precise sequence elements at one or a few positions . To prove or disprove the hypothesis that yeast intergenic G+C content heterogeneity is affected by nucleosome-related selection , we studied the evolutionary dynamics of yeast sequences bound and unbound to nucleosomes . We hypothesized that through characterization of these dynamics , we may reveal , in addition to the sequence constraints affecting yeast nucleosome organization , some general principles governing the evolution of dispersed genomic encodings . To study the evolutionary dynamics that underlie G+C content heterogeneity and nucleosome occupancy in the yeast genome , we inferred substitution rates and ancestral sequences in the Saccharomyces sensu stricto clade . We performed evolutionary inference from alignments of five yeast genomes [33] , [34] for sequences that were classified as high nucleosome occupancy loci in S . cerevisiae . We separately inferred the evolutionary trajectory at low nucleosome occupancy loci . The analysis omitted exonic sequences , since the evolutionary dynamics in these involve additional sources of selection relative to those affecting intergenic sequences . Differences in locus mutability are known to be associated with the flanking nucleotides [35] , [36] , and this effect may severely bias the comparison of evolutionary dynamics between regions with different nucleotide composition . For example , A+T rich regions , like low-occupancy sequences , may exhibit slower divergence of A/T nucleotides than G+C rich regions , simply because A/T mutability is reduced in the flanking context of A/T nucleotides . To account for this effect of flanking nucleotides on substitution dynamics , we independently estimated the rate of substitution at all 16 possible combinations of flanking nucleotides . Indeed , the substitution rates estimated by our model vary significantly among flanking contexts both in high and low occupancy loci and reflect context-dependency that is consistent among phylogenetic lineages ( Fig 2A , Fig S3 ) . For example , the C to T transition rate over the S . cerevisiae lineage in low occupancy regions varies between ∼0 . 14 in the context of tCc and ∼0 . 03 in the gCg context . The estimation of context-dependent substitution rates proved essential for the unbiased comparison of evolutionary dynamics between the low occupancy , G+C poor , and the high occupancy , G+C rich sequences . As we show next , it allowed us to robustly identify and validate major differences in the evolutionary regimes of these two classes of loci . We first studied S . cerevisiae substitution rates inferred from intergenic sequences within 200 bp of annotated transcription start sites . It is known that this region in yeast promoters is enriched for transcription factor binding sites and exhibits a stereotyped nucleosome-depleted region of length ∼100–150 bp . As shown in Fig 2B–C ( see also Fig S4 and Fig S5 ) , the analysis reveals that the rates of A/T-losing transitions ( A to G , T to C ) and transversions ( A to C , T to G ) are ∼45% lower in low occupancy sequences than in high occupancy sequences . A decrease is observed for all 16 nucleotide contexts ( within an estimation variance ) , and is slightly more pronounced in A/T contexts ( AAA , AAT ) . Notably , the rates of A/T-gaining transitions ( G to A , C to T ) and transversions ( G to T , C to A ) are not decreased like the A/T-losing substitutions . In most sequence contexts , the rates of A/T-gaining substitutions are higher in low occupancy sequences or similar between the sequence classes . On the other hand , when flanked by G's or C's , the rates of A/T-gaining substitutions are four times slower in low occupancy compared to high occupancy sequences . Evolutionary theory could not predict these dynamics if the evolution of G+C content was neutral ( unless an extremely unlikely mutational regime is separating high from low occupancy regions , as we disprove below using population genetics data ) . Moreover , a simple theory assuming average stronger evolutionary constraint on low occupancy sequences [20] , [29] would predict a general decrease in the substitution rates in the region and would not explain the asymmetry between A/T-gaining and A/T-losing substitution rates . An important assumption underlying our evolutionary analysis above is that the evolutionary regime operating in regions that are occupied ( or unoccupied ) by nucleosomes in the extant S . cerevisiae genome has been the same since the divergence of S . cerevisiae from S . paradoxus . Violations of this assumption can potentially affect our substitution rate estimations . For example , if nucleosome occupancy is determined by the genomic sequence , but is not under selection , nucleosomes may drift freely following substitutions spontaneously generating new A+T rich hotspots . Following that , we may enrich for substitutions that increase A+T content in extant low occupancy sequences by assuming nucleosome organization were conserved . To verify that such a scenario has not significantly affected our analysis of TSS-proximal substitution rates , we inferred the G+C content in the common ancestor of S . cerevisiae and S . paradoxus , for 10 ranges of S . cerevisiae nucleosome occupancy levels , and compared it to the extant G+C content ( Fig 2D ) . We found that the G+C content at all levels of nucleosome occupancy did not change significantly during evolution in the S . cerevisiae lineage . Sequences proximal to TSSs therefore conserve their regional G+C content ( at least on average ) . Consequently , the different rates of substitutions in high and low nucleosome occupancy loci do not represent net divergence in the sequence features that correlates with nucleosome occupancy . This is further confirmed by recent comparative analysis of nucleosome organization in S . cerevisiae and S . paradoxus , which revealed only limited divergence in nucleosome positioning for these species [37] , [38] . The highly non symmetric substitution dynamics observed at different levels of nucleosome occupancy must therefore be explained by means of a stationary evolutionary process that conserves the underlying nucleosome-associated encoding . One intriguing possibility that may explain the asymmetry between the rates of A/T-losing and A/T-gaining substitutions in low occupancy sequences is that while A/T-losing mutations are selected against , some can be sustained in the population . Consequently , positive selection is able to push to fixation corrective A/T-gaining mutations ( possibly at different genomic positions ) . If this hypothesis is correct , we can predict that loci near sites of A/T-losing substitutions will be enriched with A/T-gaining substitutions and vice versa . Remarkably , the yeast divergence patterns confirm this prediction . The data reveal that rates of A/T-gaining substitution are accelerated next to sites of observed A/T loss ( compared to rates near conserved loci , Fig 3A ) . Furthermore , as shown in Fig 3A , this effect does not represent general spatial coupling of substitutions , since the A/T gain rate is significantly higher near sites of A/T loss than it is near sites of A/T gain . Conversely , the rates of A/T losing substitutions are higher next to sites of observed A/T gain ( Fig 3B ) . Unexpectedly , this coupling effect is observed robustly across the entire spectrum of nucleosome occupancy levels ( p<1e-5 for high nucleosome occupancy , p<0 . 04 for low nucleosome occupancy ) . The coupling between contrasting substitutions on spatially linked loci suggests the involvement of a common selective constraint , without which the dynamics at these loci must be independent of each other . The data therefore suggest that compensating A/T-losing and A/T-gaining mutations work to conserve a heterogeneous G+C content ( both high and low ) in TSS-proximal sequences . The trinucleotide distributions of low occupancy TSS-distal sequences ( over 200 bp from an annotated TSS ) are generally similar to those in TSS-proximal loci , but some important differences are notable ( Fig 4A ) . First , for low occupancy sequences , G/C trinucleotides are rarer in TSS-distal than in TSS-proximal loci . Second , poly-A/T trinucleotides are enriched relative to other A/T rich nucleotides in TSS-proximal but not TSS-distal low occupancy loci . These differences may represent a lower fraction of TF binding sites in TSS-distal regions [19] , [20] ( Fig S6 for additional analysis ) . As shown in Fig 4B–C , TSS-distal A/T-losing substitution rates are decreased in low occupancy vs . high occupancy sequences , consistent with the observations in TSS-proximal loci . Furthermore , the rates of A/T-gaining substitution in many contexts are increased in low occupancy vs . high occupancy sequences , similar to their behavior in TSS-proximal regions ( but with G/C-flanking contexts not highly conserved ) . Comparison of the ancestral and extant G+C content reveals conservation at high levels of nucleosome occupancy , but some average decrease in G+C content for low nucleosome occupancy loci ( Fig 4D ) . Analysis of compensatory spatial correlation between A/T-gaining and A/T losing substitutions reveals significant coupling at high nucleosome occupancy levels ( p<6e-4 ) . Also shown is the tendency of A/T-gaining substitutions at low nucleosome occupancy to occur in clusters ( Fig S7 ) . The data therefore support a compensatory substitution process that drives G+C content conservation in most TSS-distal loci , in a way analogous to the dynamics at TSS-proximal loci . This is demonstrated by the asymmetric rates of A/T gain and A/T loss , the conservation of G/C content and the compensatory substitution coupling at most ranges of nucleosome occupancy . An exception to this general trend is observed at some of the TSS-distal low occupancy loci . We hypothesize that during the evolution of the S . cerevisiae lineage , de-novo A/T-rich hotspots may have driven divergence of nucleosome organization in some TSS-distal loci ( possibly since these were under weaker selection [37] , [38] ) . This effect may explain the non-stationary G+C content and spatial clustering of A/T-gaining substitutions at extant TSS-distal low occupancy loci ( Fig S7 ) . Taken together , the data on TSS-distal sequences further support the idea that selection maintains heterogeneous G+C content across most yeast intergenic sequences ( and in particular at TSS-proximal sequences ) , and that this selection drives changes in substitution rates that are difficult to explain using models of selection on a single locus . To study the hypothesis that selection on dispersed nucleosome encodings drives asymmetric substitution patterns in yeasts , we devised a simple theoretical model ( Fig 5 ) . We assume that a population of 20 bp sequences ( each representing a different “genome” ) is evolving given a constant flux of mutations in some fitness landscape that depends only on the G+C content of the sequence . The mutations transform G/C nucleotides to A/T nucleotides faster than they transform A/Ts to G/Cs , driving the genomes' stationary G+C content to a neutral level of 30% . Working against this stationary G+C content , the fitness landscape defines a lower G+C content ( 20% ) as optimal , with symmetrically decreasing fitness for suboptimal values . This landscape is designed to approximate the potential selective pressure on low nucleosome occupancy sequences . We studied the model behavior at various selection intensities both analytically and using computer simulations ( Methods ) . For each intensity level , we determined the A/T gain and A/T loss substitution rates and stationary G/C content ( Fig 5A–D ) . When selection is weak , the dynamics we observed are neutral , with the rates of substitutions being equal to the rates of mutations , and the G+C content converging to the neutral stationary G+C content ( 30% ) . In contrast , when selection is strong , the rates of both A/T gain and A/T loss decrease to zero and the G+C content is optimal ( 20% ) . These two regimes are compatible with the standard evolutionary theory of selection on a single locus . More notable are the substitution rates observed at intermediate levels of selection . When selection is not sufficiently strong to purify all A/T-losing mutations , A/T-losing substitution rates are only partially decreased . Interestingly , this decrease is matched by an increase in the rate of A/T-gaining substitutions to levels higher than the neutral rate . The new balance between A/T-losing and A/T-gaining rates is sufficient to stabilize the G+C content of the regime at near-optimal levels . Detailed analysis reveals that the increase in the rate of A/T-gaining substitutions is driven by cycles of A/T-loss mutation at one position , which are corrected by an A/T-gain mutation at another position . Similar but opposite dynamics are observed when the optimal G/C content is higher than the neutral one ( modeling selection of high G+C content in high nucleosome occupancy sequences , Fig S8 ) . Furthermore , the compensatory regime is observed over a much wider range of selection intensities when the fitness landscape is more tolerant as shown , for example , in Fig 5E–I . These theoretical predictions are consistent with the empirical behavior observed in yeast , showing that weak selection can be sufficiently powerful to increase specific substitution rates over the neutral level due to a compensatory regime . Our evolutionary analysis above supports the idea that high and low nucleosome occupancy sequences in yeast evolve under a selective pressure to maintain their G+C content , or a refined nucleosome sequence potential that is approximated by the average G+C content . According to this scenario , in low occupancy sequences , which are generally A+T-rich , A/T-losing substitutions are weakly selected against , while A/T-gaining substitutions are frequently pushed to fixation by an adaptive force . According to our simulations and to the standard population genetics theory , such selection on A/T-gaining and A/T-losing mutations should affect the distribution of allele frequencies in the population . In low occupancy loci , A/T-losing single nucleotide polymorphisms ( SNPs ) are expected to have lower allele frequencies than A/T-neutral SNPs , while A/T-gaining SNPs should have higher allele frequencies . Analysis of polymorphic sites in a sample of 39 S . cerevisae strains [39] confirmed these predictions ( Fig 6 ) . We used data on 9185 SNPs in low occupancy loci and 16956 SNPs in high occupancy loci , approximating the minor allele frequency using majority voting and discarding sites with incomplete data or more than two alleles . In low occupancy loci , A/T-losing SNPs are more rare ( <20% , alternative threshold generated similar results , Fig S9 ) than A/T-gaining SNPs in non G/C flanking context ( p<2e–05 ) . A reciprocal effect is observed at high occupancy loci , where A/T-gaining SNPs are more rare than A/T-losing SNPs in non G/C flanking context ( p<3e–07 ) . The reciprocality of the effect also confirms that our conclusions are not affected by general biases in the estimation of allele frequencies due to systematic sequencing errors . We note that as expected by the low divergence of A/T nucleotides in G/C flanking contexts of low occupancy sequences , the allele frequencies of A/T-gaining SNPs in such loci are reflective of stronger selection . This may be related to the enrichment of such flanking contexts at TF binding sites , as we discuss below .
We classified yeast intergenic regions according to their nucleosome occupancy , and used evolutionary analysis of context-dependent substitution rates to reveal remarkable variability in the evolutionary dynamics of sequences bound and unbound to nucleosomes . Our analysis shows that low occupancy sequences lose A/T nucleotides slowly compared to high occupancy sequences , but gain A/T nucleotides at similar rates . We also observe spatial coupling between substitutions that gain A/Ts and substitutions that lose them , which suggests that a compensatory process preserves G+C content at both high and low occupancy loci . These observations are compatible with a model in which the local G+C content in yeast is conserved through weak quantitative selection . Such weak selection allows occasional fixation of substitutions that disrupt the optimal G+C content of the region , but then respond by adaptive evolution of corrective mutations at the mutated locus or at any of the surrounding genomic positions . Data on allele frequencies of yeast SNPs independently confirm the predictions of such a model . This set of observations proves that the G+C heterogeneity of yeast intergenic sequences is not a consequence of a neutral process and suggests that nucleosome organization may play a major role in this lack of neutrality . The role of DNA encoded nucleosome occupancy in regulating gene expression is difficult to isolate experimentally , mostly due to the challenge of separating cause and effect inside the complex system involving nucleosomes , remodeling factors and TFs . Previous analysis identified an anti-correlation between nucleosome occupancy and genomic conservation in yeast [20] , [28]–[30] putting forward the hypothesis that low occupancy regions ( nucleosome free regions , linkers ) may be under selection , either due to their increased frequency of TF binding sites , or since they serve as anchors that organize the entire nucleosome landscape . According to our analysis nucleosome occupancy is tightly correlated with substitution patterns reminiscent of selection throughout the genome and not just at low occupancy regions . The data therefore strongly support the non-negligible contribution of DNA encoded nucleosome organization to fitness and therefore to genome regulation . This is further demonstrated by contrasting the G+C content related selection patterns at TSS-proximal sequences ( Fig 2 , 3 ) , with the frequent cases of overall divergence of A/T rich hotspots and clustered A/T-gaining substitution in TSS-distal low occupancy sequences ( Fig 4 ) . The data suggest that when selection is not working , nucleosome occupancy drifts following changes in the encoding sequences [37] , [38] . We note that according to our simulations and the empirical data , the selection on nucleosomal sequences must be weak , driven by the very small ( but still specific ) fitness contribution of any individual genomic position . We predict that such selection is sufficiently powerful to contribute significantly to the heterogeneity of the yeast intergenic sequences , but it is clearly much weaker ( per base ) than the selection working to conserve classical functional elements . These theoretical considerations underline the difficulty in proving the functionality of specific nucleosome positioning sequences using direct genetics experiments , which typically require large and easily quantifiable phenotypic effects for specific genetic manipulations . One source of evolutionary constraint on yeast intergenic sequences is their interaction with transcription factors . TF binding sites are known to be conserved among yeast species [33] , [34] and their increased concentration in TSS-proximal nucleosome free regions was previously proposed to impose overall conservation at these regions . According to our inferred evolutionary dynamics at TSS-proximal DNA , selection on TF binding sites indeed contributes to the evolution of low occupancy sequences . This is indicated , for example , by a very low A/T gain rates in G/C trinucleotides ( Fig 2 ) , which are part of some of the most abundant and conserved yeast binding sites ( e . g . , Ume6 , PAC , Reb1 , MBP1 ) [11] , [12] , [40] . Nevertheless , selection on binding sites , even those that are A/T rich ( e . g . TATA boxes ) is highly unlikely to explain the nucleosome occupancy-dependent substitution rates we observed throughout the yeast genome . Specifically , the compensatory coupling of A/T-losing and A/T-gaining substitutions is not compatible with any particular binding site model . We therefore hypothesize that a combination of purifying selection on TF binding sites ( either strong [33] , [34] or weak [11] ) and composite selection on DNA encoded nucleosome organization together define a complex fitness landscape that shapes the evolution of yeast intergenic sequences . We studied here a model of evolution as manipulating sequences in a complex fitness landscape that combines contributions from multiple coupled loci into a single dispersed encoding . As shown by theoretical and empirical analysis of the model , when selection on each individual locus is weak , purifying selection is incapable of completely purging mutations that are only slightly deleterious and these are continuously challenging the overall optimality of the sequence . This suboptimality is compensated effectively by adaptive evolution at multiple other loci that participate in the dispersed encoding . In contrast to other cases of compensatory evolution ( proteins [41] or RNA molecules [8]-[10] , [42] ) , the encodings we studied here provide ample direct ways to correct a slightly deleterious substitution , thereby increasing the rate of compensation . Our study builds on earlier work on codon bias [43] , [44] , but uses the global and experimentally characterized sequence classes at high and low nucleosomes occupancy loci to establish compensatory evolution as a major driving force in evolution under multi-site selection . This type of evolutionary dynamics may be generalized to other dispersed functional encodings [45] , [46] including complex regulatory switches that typically involve a large number of TF binding sites of variable factors and specificities . The remarkably global nature of the compensatory effect we observed in yeast , which cause a measurable global increase in the substitution rate of specific mutations , supports the notion of an evolutionary process that conserves function without a strict requirement to conserve sequence . It is tempting to speculate that such a process may allow genomes to maintain diversity and continuously search the sequence space , without significantly compromising their existing regulatory circuits . Furthermore , this process may reduce , through compensation , the mutational load [47] resulting from the use of multiple loci to encode regulatory functions .
Multiple alignments of the Saccharomyces cerevisiae , Saccharomyces paradoxus , Saccharomyces mikatae , Saccharomyces kudriavzevii and Saccharomyces bayanus were downloaded from the UCSC database [48] ( sacCer2 version ) . Alignments were based on the SGD June 2008 assembly . A genome wide in-vivo nucleosome occupancy profile for S . cerevisiae was used as previously described [21] , indicating a nucleosome occupancy value for each genomic position . SNP data were downloaded from the SGRP website [39] . Gene Annotations and transcription start sites of S . cerevisiae were taken from the SGD known gene table which corresponds to sacCer2 [49] . Transcription factor binding sites were downloaded from the UCSC Genome Browser [48] and are based on the chip-chip experiments described before [50] . Our analysis focused on intergenic genome sequences which are defined based on the SGD gene annotations . Each intergenic locus was defined as TSS-proximal if it is not part of an exon , and has an annotated TSS within 200 bp of it . TSS-distal loci included the remaining non exonic loci . We defined low occupancy loci as positions with nucleosome occupancy value lower than −2 . 5 ( relative to the genomic mean , detailed description in Kaplan et al . [21] ) and high occupancy loci as positions with occupancy higher than 0 . 4 . Alternatively , we classified all loci to equal sized bins of nucleosome occupancy ( ten in analysis of ancestral G+C context and five in the analysis of spatial coupling ) . Alternative definition of low occupancy linker regions based on raw data of MNase restriction sites resulted in similar results ( data not shown ) . As described in the text , a refined context dependent substitution model is essential for the correct estimation of the different evolutionary dynamics in low G+C content , low occupancy loci and high G+C content , high occupancy loci . We therefore applied a flexible substitution model to perform ancestral inference and learn evolutionary parameters from alignment data ( details available upon request ) . The model included parameters for the substitution rates at each of 16 possible contexts parameterized by the identities of the 3′ and 5′ flanking nucleotides . Independent substitution rates were assumed for each lineage in a phylogenetic tree which was fixed throughout the process . We note that the model does not assume parametric constraints on different substitution rates , and infers substitution rates on lineages , rather than a global substitution rate matrix and branch lengths . This approach has proved more robust given that a sufficient number of loci was available to learn robustly the parameters at each lineage , and given that the substitution process in the different lineages indicated gradual changes in dynamics that a model using a universal rate matrix could not have accounted for ( for example , the extant G+C content in each of the species we used show some variability ) . To perform ancestral inference , we used a customized loopy belief propagation algorithm on a factor graph approximation of the model [51] . Parameter estimation was then performed using a generalized EM algorithm . We validated some key results using parsimony analysis ( Fig S10 and data not shown ) . For analysis of the resulted model parameters , each context dependent substitution rate was averaged with its reverse complement . For example CAT->CCT is averaged with ATG->AGG . The averaged conditional probabilities are presented in Fig 2 , 4 , Fig S3 and Fig S4 . A/T gaining is defined as any of the following substitutions in any flanking contexts: C->A , C->T , G->A , G->T . A/T loss in defined as any of the following substitutions in any flanking contexts: A->C , A->G , T->C , T->G . Analysis was generally focused on the S . cerevisiae lineage ( data on the other lineages are shown in Fig S3 , Fig S5 ) . In order to estimate the theoretical regional G+C content of S . cerevisiae intergenic sequence , we have simulated this sequence using a lineage specific evolutionary probabilistic model learned over the whole intergenic sequence ( see above ) . Specifically , the common ancestor of the sensu stricto clade was simulated first based on the learned 2-order markov model . Following this , the sequences of the descendants were simulated based on the simulated ancestor sequence and the corresponding substitution model . Iteratively , the sequences of all species in the phylogeny were simulated , including the extant species . The regional G+C content of the simulated S . cerevisiae intergenic sequence is presented in Fig S1 . To estimate the coupling between A/T gaining and A/T losing substitutions in the yeast genome , we used our probabilistic model to infer at each genomic position j the posterior probability of each type substitution in the lineage leading to species i from its ancestor ( pai ) : When sji denotes the nucleotide at the j'th genomic position of the i'th species in the phylogeny , and sjpai denotes the sequence of the ancestor of this species at the same genomic position . Given the posterior probabilities we computed for each genomic position j the expected numbers of A/T loss and A/T gain events in the sequence preceding it . This was done using a horizon parameter , which was set to 5 bp by default ( for alternative horizon values see below ) : Where the δgain , δloss functions were given by Table 1 , and the net A/T divergence of the position was defined as: We then identified all positions with A/T divergence <-0 . 9 ( A/T losing contexts ) , with A/T divergence >0 . 9 ( A/T gaining contexts ) and with conserved A/T content ( background ) . For each such set we computed the probability of A/T gain and A/T loss substitutions using the same inferred posterior probabilities . By using this approach ( conditional probability given the events in the preceding 5 bp ) we ensured each substitution is counted precisely once . By computing the probabilities for similar events ( e . g . A/T gain ) given different contexts ( A/T losing , A/T gaining , or background ) , we could robustly asses compensation patterns while controlling for the different basal rates of A/T gain and A/T loss and the general clustering of substitution in the genome . To statistically assess the coupling between A/T divergence context and A/T losing/gaining substitutions in the S . cerevisiae lineage we counted the numbers of A/T gains and A/T losses at A/T gaining and losing contexts: = number of A/T gains in A/T gaining contexts = number of A/T losses in A/T gaining contexts = number of A/T gains in A/T losing contexts = number of A/T losses in A/T losing contexts In addition we counted the numbers of A/T and C/G occurrences in these contexts: = number of A/T's in A/T gaining contexts = number of A/T's in A/T losing contexts = number of C/G's in A/T gaining contexts = number of C/G's in A/T losing contexts We wished to test whether the spatial compensation effect is significant even given the general clustering of substitutions . Our null hypothesis was therefore: We test it using bootstrapping with 100 , 000 resamples . At each resample , a set of items are sampled without replacement out of the union of two sets of sizes and ( denoted by A , B respectively ) . Similarly , we sample without replacement items out of the union of two sets of size , ( denoted by C , D respectively ) . The number of sampled items belonging either to set A or C is collected across all resamples . We end up with 100 , 000 counts representing the background distribution for the statistic . P-value for the null hypothesis is calculated by counting the fraction of iterations in which the sampled counts are bigger than . Analysis of the robustness of the observed compensation patterns for different values of the horizon parameter is shown in Fig S11 , Fig S12 , and Fig S13 . To study the hypothesis that selection on dispersed nucleosome encodings drives asymmetric substitution patterns in yeasts , we devised a simple theoretical model . For clarity we describe here the version of the model for low occupancy sequences . For nucleosome DNA the model is the same apart from the fitness function . First we used a Wright-Fischer dynamics on a population of binary sequences of size L , : In each generation there is a probability of for each site containing 0 to be flipped to 1 and for sites containing 1 to flip to 0 . The sequences are then sampled relative to their fitness , where and is equivalent to the GC content . We simulated this system for the following parameter set We note that the population expected θ parameter may be estimated from the above parameters ( in haploid population , but given the two different mutation rates the empirical theta needs to be corrected ) . The parameters we used ensured θ<0 . 04 . The simulation was based on the following procedure: Initialize: Create a population of identical sequences of length L . For simplicity sequences use a binary alphabet on A and G . We define the current reference genome sequence R using the same initial sequence . We introduce the following counters to accumulate sufficient statistics for computing the rates of A->G and G->A substitutions ( NA , NG and NA->G , NG->A , such that the rate will be estimated as NA->G /NA , NG->A /NG ) . Sample a new generation: to create a new generation , we sample times from the current population using weights that are proportional to the fitness of each individual . For each sampled individual , we introduce mutations with probability for G loci and for A loci . Starting after a minimal number of “burn-in” iterations ( at least 4 coalescent times ) we also incremented NA and NG for each sampled individual with the number of A's and G's in the respective sequence . Updating the reference genome: given the new generation population , we tested the frequency of A and G at each of the L genomic loci . Whenever the frequency in the current population is larger than 0 . 95 and the major allele is different from the reference genome R , we incremented the counter NA->G or NG->A ( after the burn-in period ) and updated the sequence R . We end up with counts of A's ( NA ) , counts of G's ( NG ) ( in units of generations X loci ) and counts of the substitutions between them ( NA->G , NG->A ) . Substitution rates are estimated by: These rates are shown in Fig 5 and Fig S8 for the different fitness landscapes we defined next . The goal landscape is defined symmetrically around an optimal number of G's denoted by nGC and the selection intensity η ( e . g . , X axis in Fig 5B , C , F , G ) : The threshold landscape is defined using similar parameters to generate an asymmetric function: Next , we studied the above model analytically in the regime of low mutation rates . In this regime , drift is the dominating mechanism and we can model the process by assuming the population is represented by a single genome ( or GC content ) . Given the definitions above , the rate at which mutations that increase the GC content enter the population is While the rate of mutations that decrease the GC content is In such drift dominating regime , the fixation probability of a new mutation is:where is the marginal fitness of the mutation [52] . Therefore , the rate at which the GC-content increases or decreases is on average Where . Thus the set equations for the dynamics of is Solving this for the steady state results in Where and is set by normalization . From this distribution of the GC-content one can calculate the average GC-content and the substitution rates As can be seen in Fig S8 , the analytical result and Wright-Fischer simulation are in good agreement . We used DNA sequences of 39 S . cerevisiae strains sequenced in the Saccharomyces genome resequencing project ( SGRP ) . Here , only intergenic 2 allele SNP's with sequence data from more than 20 strains were considered informative . For each of those SNP's , major allele was defined as the most abundant allele in the population . Minor allele is defined as the least abundant allele . A/T gaining SNPs were defined when the nucleotide of the major allele was C or G and the minor allele is A or T . A/T losing SNPs were defined reciprocally . All other SNPs were defined as A/T conserving ( see illustration in Fig 6 ) . We further subdivided SNPs into two groups: SNP's in G/C flanking context and SNP's with at least one A or T in the flanking contexts , using the reference strain for determining the context . These subgroups are again subdivided to SNP's within low occupancy sequences and SNP's within high occupancy sequences ( Fig 6 ) . We analyzed the distributions of the frequency of minor alleles of these subgroups separately . In figure 6 , shown are the fraction of rare alleles ( minor allele frequency <0 . 20 ) among A/T gain , A/T loss and A/T conserved SNP's within low or high occupancy sequences . We used a chi-squared test to reject the null of hypothesis that the fraction of rare alleles is the same between A/T gain and A/T loss SNP's . | Purifying selection is a major force in conserving genomic features . It pushes deleterious mutations to extinction while conserving the specific DNA sequence . Here we show that a large proportion of the yeast genome evolves under compensatory dynamics that conserve genomic properties while modifying the genomic sequence . Such compensatory evolution conserves the local G+C content of the genome , which influences nucleosome organization . Since purifying selection is too weak to eliminate every weakly deleterious mutation in nucleosome bound or unbound sequences , the local G+C content is frequently stabilized by compensatory G+C gaining and G+C losing mutations in proximal loci . Theoretical analysis shows that compensatory evolution is inevitable when natural selection is weak and the genomic feature is distributed over many loci . These results imply that sequence conservation may not always be equated with overall selection . They demonstrate that cycles of weakly deleterious substitutions followed by positive selection for corrective mutations , which were so far studied mostly in RNA coding genes , are observed broadly and profoundly affect genome evolution . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"genetics",
"and",
"genomics/comparative",
"genomics",
"computational",
"biology/comparative",
"sequence",
"analysis",
"computational",
"biology/evolutionary",
"modeling",
"computational",
"biology/genomics",
"evolutionary",
"biology",
"genetics",
"and",
"genomics/bioinformatics"
] | 2010 | Widespread Compensatory Evolution Conserves DNA-Encoded Nucleosome Organization in Yeast |
Genome-wide association studies ( GWAS ) have been successful in identifying single nucleotide polymorphisms ( SNPs ) associated with many traits and diseases . However , at existing sample sizes , these variants explain only part of the estimated heritability . Leverage of GWAS results from related phenotypes may improve detection without the need for larger datasets . The Bayesian conditional false discovery rate ( cFDR ) constitutes an upper bound on the expected false discovery rate ( FDR ) across a set of SNPs whose p values for two diseases are both less than two disease-specific thresholds . Calculation of the cFDR requires only summary statistics and have several advantages over traditional GWAS analysis . However , existing methods require distinct control samples between studies . Here , we extend the technique to allow for some or all controls to be shared , increasing applicability . Several different SNP sets can be defined with the same cFDR value , and we show that the expected FDR across the union of these sets may exceed expected FDR in any single set . We describe a procedure to establish an upper bound for the expected FDR among the union of such sets of SNPs . We apply our technique to pairwise analysis of p values from ten autoimmune diseases with variable sharing of controls , enabling discovery of 59 SNP-disease associations which do not reach GWAS significance after genomic control in individual datasets . Most of the SNPs we highlight have previously been confirmed using replication studies or larger GWAS , a useful validation of our technique; we report eight SNP-disease associations across five diseases not previously declared . Our technique extends and strengthens the previous algorithm , and establishes robust limits on the expected FDR . This approach can improve SNP detection in GWAS , and give insight into shared aetiology between phenotypically related conditions .
Genome-wide association studies ( GWAS ) have enabled identification of genetic variants associated with a wide range of complex phenotypes , but in many cases these variants explain only a proportion of the known heritability [4] . There is increasing evidence that this is due to the combined contribution of small effects arising from multiple distinct variants [5] . The testing of a large number of potential variants in parallel , with a comparatively low number of samples , mandates a stringent threshold for significance in order to limit false positives ( type 1 errors ) , meaning that discovery of variants responsible for small effects requires very large sample sizes . Detection of such variants by increasing numbers of samples in studies is time-consuming and expensive , particularly for rare phenotypes , but it may be possible to improve detection by re-analysis of existing data [6] . One promising strategy is to co-analyse GWAS results from similar phenotypes to exploit potential similarities in genetic aetiology . This has been attempted using several different methods [2 , 7 , 8] . The assumption that GWAS for similar diseases may yield overlapping sets of disease-associated variants is based on the phenomenon of pleiotropy , in which a genetic variant is associated with more than one trait or disease [9] . Pleiotropy is common in human genes: even when exclusively considering single nucleotide polymorphisms ( SNPs ) with strong evidence of association , around 15% of those associated with at least one trait are associated with multiple traits [10] . Elements of shared genetic aetiology may be suspected in diseases with similar symptomatology , such as bipolar disorder and schizophrenia [11] or in diseases with common risk factors , such as type 2 diabetes and obesity [12] . If two diseases are known or suspected to share associated genetic variants , a degree of association of a locus with one disease may increase the likelihood of association with the other . Use of external covariates in this way can alleviate some of the effect of multiple testing [8] , meaning that phenotypic similarity may lead to improved detection of disease-associated variants . Correspondingly , discovery and specification of shared genetic aetiology between two diseases may suggest some shared pathophysiology [12] . A technique for improved discovery of disease variants using pleiotropy between pairs of diseases has been successfully developed and applied by Andreasson et al [3 , 13 , 14] . The technique extends the empirical Bayesian false discovery rate [15] to a two-phenotype scenario , in which association with one phenotype is tested conditional on varying degrees of association with another . We denote the phenotype for which association is being tested as the ‘principal phenotype’ and the other as the ‘conditional phenotype’ . By successively restricting attention to SNPs with a given strength of association in the conditional phenotype , the number of parallel tests to perform for association with the principal phenotype is reduced . If the two phenotypes share common associated variants , this restriction will retain disease-associated SNPs at a higher rate than null SNPs , resulting in a higher proportion of disease-associated SNPs in the restricted group than in the whole . The ‘conditional false discovery rate’ ( cFDR ) , defined as the probability that a SNP is not associated the principal phenotype given its p values for the principal and conditional phenotypes are below some thresholds , exploits this effect . By computing cFDRs for schizophrenia conditioned on bipolar disorder and vice versa , Andreasson et al [3] identified multiple previously undiscovered loci for both . In a separate study computing cFDRs for hypertension conditioned on 12 related traits [13] , 42 new loci associated with hypertension were reported . These constituted considerable improvement on existing results using single GWAS , albeit using a rather relaxed threshold of estimated cFDR ≤ 0 . 01 . A major disadvantage of the algorithm developed and used by Andreasson et al is the requirement that control groups for the two GWAS be distinct , in order to ensure that observed effect sizes are uncorrelated at null SNPs . This requires splitting a pool of potential controls between studies , with the summary statistics for each GWAS computed from only the controls allocated to that study . This may be impractical as it requires access to raw genotype data . More importantly , accuracy of effect size estimates improves with larger control groups , and consequently splitting controls in this way weakens the effect size estimates for individual studies . For this reason , many researchers employ a study design in which controls are pooled into a large group; for example , the Wellcome Trust Case Control and ImmunoChip consortia [16 , 17] . Here we extend the cFDR approach to studies with overlapping control groups , exploiting an approach developed by Zaykin et al , following Lin et al [18 , 19] to adjust for the effect of shared controls . This allows the strongest available estimates of effect sizes to be used for calculation , and consequently strengthens the power of the technique . Our technique additionally allows cFDR rates to be computed from summary statistics alone , without the need to recalculate effect sizes after re-allocating controls . We demonstrate the improvement arising from sharing controls in a type 1 diabetes data set . We also identify a previously undiscussed difficulty with the technique potentially leading to a falsely low estimate of false discovery rate amongst SNPs declared non-null . Multiple overlapping sets of SNPs may be defined each of which has cFDR ≤ α . However , the union of these sets does not necessarily have an expected false-discovery rate less than α and is generally higher . An implication of this is that if we declare non-null all SNPs for which estimated cFDR is less than α , the expected overall false-discovery rate amongst SNPs declared non-null is greater than α . We describe an upper bound on the false discovery rate amongst such SNPs based on areas of regions of the unit square . We apply our method to summary SNP association statistics for ten phenotypically distinct autoimmune diseases: type 1 diabetes ( T1D ) [20] , autoimmune thyroid disease ( ATD ) [21] , coeliac disease ( CEL ) [22] , multiple sclerosis ( MS ) [23] , narcolepsy ( NAR ) [24] , primary biliary cirrhosis ( PBC ) [25] , psoriasis ( PS ) [26] , rheumatoid arthritis ( RA ) [27] , ulcerative colitis [28] , and Crohn’s disease [28] . All were genotyped using a common SNP array: the ImmunoChip , designed to provide dense genotype coverage of regions associated with autoimmune disease . Many autoimmune traits are known to have significant heritability , much of which remains unexplained [29] . We hypothesised that our method can improve detection of disease-associated variants in these diseases without the need for distinct control groups .
The unconditional false discovery rate for a set of SNPs with p values < pi is defined as the probability that a random SNP from this set is null . We denote this as uFDR ( pi ) , and our estimate as u F D R ̂ ( p i ) . The conditional false discovery rate ( cFDR ) is defined [3 , 14] as the probability that a random SNP is null for a phenotype i given that the observed p values at that SNP for phenotypes i and j are less than ( pi , pj ) ; that is , P r ( H 0 ( i ) ∣ P i ≤ p i , P j ≤ p j ) , where H 0 ( i ) is the null hypothesis that the SNP is not associated with phenotype i . We denote this quantity as cFDR ( pi∣pj ) , and call phenotype i the ‘principal phenotype’ and phenotype j the ‘conditional phenotype’ . We first apply genomic control to allow the assumption that , globally , P values for null SNPs are uniformly distributed on [0 , 1] . We compute an estimate of the cFDR , which we denote c F D R ̂ ( p i ∣ p j ) , in a similar manner to that proposed by Andreasson et al , but incorporating expected non-uniformity in the distribution of Pi due to the sharing of controls . As u F D R ̂ ( p i ) is monotonically related to pi , we set a significance cutoff at the maximum value of u F D R ̂ ( p i ) with pi < 5 × 10−8 . Correspondingly , we set a significance cutoff for c F D R ̂ ( p i , p j ) at the maximum c F D R ̂ ( p i , p j ) with pi < 5 × 10−8 . Implementation of these steps in R is available from https://github . com/jamesliley/cFDR-common-controls . If no controls are shared between studies , it is reasonable to assume that observed effect sizes for the two phenotypes are independent under a null hypothesis for the principal phenotype . This implies that the expected quantile of a given SNP’s p value for the principal phenotype is simply the p value itself regardless of its p value for the conditional phenotype . However , when control samples are shared , this assumption is invalid . Shared controls induce a positive correlation on estimated effect sizes for the principal and conditional phenotype [18 , 19] , meaning that when attention is restricted to SNPs with a given degree of association with the conditional phenotype , the p values for the principal phenotype will be falsely low; that is , the probability P r ( P i ≤ p i ∣ P j ≤ p j , H 0 ( i ) ) will not in general be equal to pi ( = P r ( P i ≤ p i ∣ H 0 ( i ) ) ) ; in fact it will usually be higher . When controls are shared , the distribution of p values for the principal phenotype given p values for the conditional phenotype depends on the underlying effect of each SNP on the conditional phenotype . For any given SNP , this underlying effect size , which we denote η , is not known . However , across all SNPs , η may be considered to be realisations of a random variable H whose distribution is mirrored by the distribution of observed effect sizes for the conditional phenotype . By integrating over this unknown true effect size for the conditional phenotype , allowance can be made for shared controls , and the ‘expected quantile’ of a p value for the principal phenotype , defined as P r ( P i ≤ p i ∣ P j ≤ p j , H 0 ( i ) ) , can be calculated , as detailed in the Methods section . We assume that H has a mixture distribution defined by two parameters ( π0 , σ2 ) , such that H = 0 with probability π0 and H ∼ N ( 0 , σ2 ) with probability 1−π0 . The parameters ( π0 , σ ) are estimated from the observed distribution of effect sizes for the conditional phenotype . In order to show the effect of our p value adjustment , we simulated p values for 20 , 000 SNPs for a principal and conditional phenotype , with controls shared between simulated studies . All SNPs were null for the principal phenotype , and were variably null or non-null for the conditional phenotype with probability 0 . 9 , 0 . 1 respectively . Z scores at non-null SNPs for the conditional phenotype were distributed as N ( 0 , σ2 ) , as per our assumption . A value of σ = 3 was used , which was similar to the values of σ in real data estimated by our E-M algorithm . We considered the set of simulated SNPs with p values for the conditional phenotype less than 0 . 05 ( Fig . 1 ) . In the absence of shared controls , we expect the distribution of pi amongst this set to be uniform , and hence expect the black dots to lie along the x-y line . However , we see the principal p values are biased downward in this set ( black dots , Fig . 1 ) . Our computed expected quantile ( blue dots ) agrees closely with the observed quantile . In a sense , this constitutes ‘adjusting’ the p values for the principal phenotype so that the expected distribution is uniform under the null hypothesis . Software to generate this simulation is available at https://github . com/jamesliley/cFDR-common-controls . Our formula can easily be adapted to arbitrary distributions of H at the cost of increased computational time , but the form of the distribution of H is not generally known . We show in S1 Text ( section A ) that , even for distributions of H which differ markedly from our assumption of normality , the error in the estimate is not large , and generally translates to a negligible difference in the set of SNPs declared non-null using the c F D R ̂ method . While our assumption has the potential to be anti-conservative if H is bimodal , nonparametric estimates for distributions of effect sizes suggest they have a uni-modal distribution centred on zero[30] . Reassuringly , our assumption is conservative if H has heavier tails than a normal . We compared Andreasson’s approach to SNP discovery which advocated splitting controls into non-overlapping subsets to our extended shared-control approach using a type 1 diabetes dataset with a total 12 , 175 cases and 15 , 171 controls . Controls and cases were each split into two sets ( control sets had size 7 , 585 and 7 , 586 , cases 6087 and 6088 ) . ‘Split’ p values were computed using one set of controls and one set of cases and corresponding ‘shared’ p values were computed using the complete set of controls . As expected , more shared p values reached genome-wide significance than did split p values ( Fig . 2 ) . We computed cFDR values by labelling one set of cases ‘conditional’ and the other ‘principal’ using the split-control p values using Andreasson’s approach and using the shared control p values using our method . For reference , we compared these to a naive application of Andreasson’s method on the shared-control p values ( Fig . 2B ) . More SNPs can be declared significant according to cFDR using the shared-control than split-control approach at all reasonable thresholds , and naive application of Andreasson’s approach to shared-control p values again increases the number declared significant . Because the quantity P r ( P i ≤ p i ∣ P j ≤ p j , H 0 ( i ) ) is systematically underestimated when using this naive method ( by assuming it is equal to pi ) as shown in S1 Text ( section C ) , it leads to a falsely low c F D R ̂ . The increase in observed number of SNPs declared significant when using the naive method shows that it can indeed lead to false discoveries . For principal phenotype p values in the range 5 × 10−6 - 5 × 10−8 - effectively the region from which ‘new’ SNPs may be discovered by cFDR rather than p value alone—the naive cFDR is frequently underestimated by 2–3 fold ( S1 Fig . , left panel ) . For lower p values , the naive cFDR may underestimate by hundreds- or thousand-fold , with the potential fold underestimation increasing with decreasing p value ( S1 Fig . , right panel ) . Because of the relatively high ratio of number of controls to number of cases , the correlation between effect sizes is lower in this constructed case ( c . 0 . 22 ) than between most phenotypes in our study ( c . 0 . 5 ) . The underestimation of cFDR using the ‘naive’ method worsens with higher correlation , so we would expect that the fold-underestimate we see here is less severe than that which would be observed if applying this to other studies . An important property of our method is the control of the expected false discovery rate ( FDR ) : the expected proportion of false positives among the SNPs found by our method . The p values at a SNP for the principal and conditional phenotype correspond to a point in the unit square . In this sense , we can define the expected FDR of a region R of the unit square as the ratio of the expected number of null SNPs whose p values are in R divided by the expected total number of SNPs whose p values are in R . From a result of Benjamini and Hochberg [31] , the expected FDR when R is a rectangle with vertices at ( 0 , 0 ) , ( pi , 0 ) , ( pi , pj ) , ( 0 , pj ) is at most α = cFDR^ ( p i ∣ p j ) . If we denote by L the closed region defined by the set of p value pairs pi , pj such that cFDR^ ( p i ∣ p j ) ≤ α , then L has the property that the FDR of any rectangle of this form contained within L is less than α . However , the expected FDR over L is not necessarily bounded by α ( Fig . 3 ) . This can be seen most easily in the extreme scenario in which all non-null SNPs are concentrated in the lower left corner of the unit square , and all null SNPs are also null for the conditional phenotype . In this case , the expected number of null SNPs in a rectangle is proportional to its area , so L is the union of all rectangles of the form above of a given area containing the lower-left corner of the unit square; that is , a hyperbola . Clearly the area of L is larger than the area of these constituent rectangles , yet it contains the same number of non-null SNPs , so it has a lower FDR . In the original method [3] , SNPs were declared significant if they were contained within any rectangular regions with a cFDR^ value of less than 0 . 01 . Our reasoning demonstrates that the expected false-discovery rate amongst all such SNPs was higher than 0 . 01 . We can derive an upper bound for the expected FDR of L by considering M* , the largest rectangle in L . We show in S1 Text ( section B ) that the bound may be expressed simply as v ( L ) v ( M * ) α * , where v ( ) denotes the expected number of null SNPs contained within L or M* ( approximately the area of L and M* ) and α* is the cFDR at the upper-right vertex of M* ( Fig . 3 ) . We obtained summary statistics in the form of p values for ten immune mediated diseases from ImmunoBase ( www . immunobase . org , accessed 19/3/14 ) . For each pair of diseases , the number of shared controls was estimated according to the description of the control samples in each paper . The numbers of cases , controls and our estimated numbers of shared controls for each study are shown in Table 1 . Uniform quality control criteria were applied to all SNPs , and the MHC region , which exhibits both strong LD and strong association with immune mediated diseases was excluded . P values were corrected within each trait for genomic inflation using a standard algorithm [32] applied to SNPs included on the ImmunoChip to replicate a GWAS study of reading and maths ability ( Steve Eyre and Cathryn Lewis , personal communication ) , unlikely to be related to any immune mediated disease studied here . P values for each principal phenotype were adjusted to p′ as described above in order to account for the effect of shared controls . For each ordered pair of phenotypes , a Q-Q plot was generated as per Andreasson et al [3] . A Q-Q plot is a graph of the observed distribution of a random variable against the expected distribution . We overlaid Q-Q plots for log10 ( p′ ) values for the principal phenotype for subsets of SNPs exhibiting successively smaller p values for the conditional phenotype . Fig . 4 shows QQ plots for T1D conditional on RA and PSO; plots for all other pairwise comparisons may be found in S4–S13 Figs . Notably , if lines shift further left with more stringent cutoffs on association with the conditional phenotype , then SNPs which are associated with the conditional phenotype are more likely to be associated with the principal phenotype , indicating pleiotropic effects of SNPs on the two phenotypes . In many cases , the Q-Q plots demonstrate considerable leftward shift with conditioning on association with a second disease , and we see strong evidence for pleiotropy for T1D conditioned on RA and little or no evidence for pleiotropy for T1D conditioned on PSO . We estimated the unconditional and conditional false discovery rates , uFDR^ ( p i ) and cFDR^ ( p i ∣ p j ) , at each SNP for each phenotype and each ordered pair of phenotypes respectively . Fig . 5 shows cFDR^ for T1D conditioned on RA . The advantage gained by cFDR^ can be seen in the left-shift of the region in which a SNP can be declared significant ( blue dots ) , corresponding to a higher p-value cutoff for significance for T1D among SNPs with low p values for RA . Indeed , if only SNPs with a p value for RA less than some threshold ζ are considered , a p value cutoff for significance for T1D is given by the leftmost border of the blue dots on the line Pj = ζ . The degree of leftward shift in the Q-Q plots clearly contains information about the degree of pleiotropy between diseases . We defined a statistic summarizing some aspects of this evidence for pleiotropy and used it to visualise the set of pairwise relationships between diseases as a network ( Fig . 6 ) . The network encouragingly reflects several pathophysiological associations: UC is linked to CRO , and T1D to ATD . Strong linkage is also seen both ways for MS and PBC , and between T1D and RA , findings which can also be seen in the Q-Q plots ( S4–S13 Figs . ) . One way relationships suggest the presence of a larger total number of associated SNPs for the disease at the start of the arrow than at the end . The numbers of SNPs deemed significant for each phenotype by analysis using unconditional and conditional approaches are shown in table 2 , with details in S2–S11 Tables . cFDR^ allows certain SNPs with p values as high as 3 × 10−6 to be declared significant while controlling the false discovery rate at a relatively low value . Fifty-one of the 59 SNPs we identify uniquely through cFDR^ have previously been reported to be associated with the relevant disease through use of alternative significance thresholds , other genomic control procedures , other GWAS or additional samples not genotyped by ImmunoChip , a useful verification of our technique . Eight of the SNPs we discover uniquely through cFDR were in regions not previously known to be associated with the corresponding disease ( table 3 ) . These will require replication in independent samples to be declared truly associated , but they contain some potentially interesting signals , such as an association for RA at SNP rs72928038 near existing MS , ATD and T1D associations in BACH2 , a transcriptional regulator involved in transcription repression and activation by MAFK [33] The SNP rs1034290 in region 1p13 . 1 , which we found to be associated with PBC , is in intron three of CD58 , which is a surface receptor involved in binding and activation of T-lymphocytes . The protective effect of the MS-associated allele is postulated to arise from upregulation of the transcription factor FOXP3 [34] and the patterns of association in the region suggest the two diseases may share a causal variant here ( http://www . immunobase . org ) .
We have extended a technique for computing conditional Bayesian False Discovery Rates to GWAS for independent diseases with shared control groups . This technique enables improved detection of disease-associated SNPs compared to conventional methods . By enabling larger control groups for each study , our method uses data more efficiently than in corresponding study designs in which control groups are independent , and is applicable to a wider range of GWAS datasets for which only summary statistics are available . Combination of GWAS by analysis of pleiotropy in this sense has several attractive advantages over single-phenotype analysis . The most obvious advantage is improved detection of disease-associated SNPs using GWAS without the need for additional samples . A secondary advantage arises from understanding of the pleiotropic structure between phenotypes: if a SNP is known to exhibit pleiotropy between two conditions , it may be causative for a shared risk factor or pre-disease state . Analysis of such SNPs has the potential to yield information on disease aetiology , with implications for preventative medicine and development of treatment . A further potential use for this technique could be the genomic analysis of diseases with complex phenotypes . In many cases , distinction between two diseases may be difficult; for instance , Crohn’s disease and Ulcerative Colitis [35] . Additionally , many diseases , including narcolepsy ( http://www . uptodate . com/contents/clinical-features-and-diagnosis-of-narcolepsy , accessed 20/6/14 ) , are definitively diagnosed on clinical grounds . This implies that these diseases may constitute a range of biochemical and genetic states . Inclusion criteria based on objective biochemical grounds , such as that used for narcolepsy in the context of this paper [24] are unlikely to characterise all patients with these diseases , and conclusions drawn from studies will not necessarily be medically applicable to the whole patient population . Given this , diseases defined phenotypically with potential genomic diversity may be better analysed by separate consideration of biochemically-defined subtypes , with a collective analysis performed by a method such as cFDR^ , avoiding the assumption that the genomic bases of disease subtypes are identical . We identify a counter intuitive property that the FDR in the union of all regions with cFDR^ less than a given α may be greater than α , and propose a method to overcome this problem . Our methods for adjusting cutoffs to control FDR and account for multiple testing demonstrate the geometrical elegance of the theory of these techniques , with the possibility for further improvements and understanding . They are complex to apply , but could be much simplified if interest was directed to SNPs with conditional p values less than some threshold p0 . Our method would ensure that the expected false discovery rate at SNPs with cFDR^ ( p i ∣ p 0 ) ≤ α would indeed be controlled at α . Our more complicated method to control FDR is necessary if the variable pj is used in place of the constant p0 . An important consideration in both our method and the original Andreasson method is that a cFDR^ ( p i ∣ p j ) value which reaches significance does not constitute genome-wide evidence of association with the conditional phenotype j; indeed , the probability of association with the conditional phenotype relates to cFDR^ ( p j ∣ p i ) and in general cFDR^ ( p i ∣ p j ) ≠ cFDR^ ( p j ∣ p i ) . In some cases , where the principal p value is very close to genome-wide significance , even conditioning on pj ≤ 0 . 5 can theoretically be enough to reach the relevant cFDR^ threshold . This is not a weakness of the cFDR^ method as such , but a consequence of using a discrete technique ( a significance cutoff ) on a variable which essentially continuous in two dimensions ( cFDR^ ) . Principal p values greater than 5 × 10−8 which can be declared significant conditioning on large conditional p value cutoffs correspond to an increase in the area of the region L ( see results section ) , which is accounted for by our FDR-controlling method . Our method enables improved detection of SNPs compared to analysis of unconditional FDR ( principal p value alone ) . However , the improvement is smaller than that reported by Andreasson et al [3 , 13 , 14] , who detected almost twice as many SNPs using cFDR as they would have detected with uFDR . This is expected for two reasons . Firstly , the gain in power from cFDR essentially comes from an increase in the total number of controls and the effective number of cases . If controls are shared , the only information gain can come from increasing the number of effective cases . Consequently , the difference in power between cFDR and uFDR will not be as large when controls are shared , although both outperform their counterparts when controls are split . Secondly , we were careful to use stringent cutoffs for FDR which were chosen to mirror the established genomewide significance threshold of p ≤ 5 × 10−8 , generally equivalent to a false discovery rate around 5 × 10−6 to 5 × 10−5 , compared to Andreasson et al who declared non-null all SNPs with cFDR^ < 0 . 01 . One alternative way to exploit pleiotropic relationships is by meta-analysing two related diseases together , as though the diseases were the same . Our method confers several advantages over this approach . The most important of these is that our method borrows strength from other SNPs according to the level of genome wide pleiotropy between diseases; that is , if the two GWAS suggest extensive pleiotropy ( such as Fig . 4 for T1D — RA ) , a low p value for a conditional phenotype will ‘sway’ our judgement of association with the principal phenotype more than the same p value for a conditional phenotype with poor pleiotropy ( such as Fig . 4 , for T1D — PSO ) . A meta-analysis would not distinguish these two scenarios . A secondary advantage of our technique is that SNP detection is not systematically weakened if the two diseases do not exhibit pleiotropy , as would be the case in meta-analysis; this arises because we are testing association with only one of the two phenotypes at a time .
This paper re-analyses previously published datasets . All patient data were handled in accordance with the policies and procedures of the participating organisations . We obtained SNP summary statistics from ten studies on autoimmune diseases from ImmunoBase ( www . immunobase . org ) . Inclusion and exclusion criteria for the studies are described in detail in the original publications ( [20–28 , 28] . Generally , some or all controls from different studies were obtained from common data sources , resulting in overlapping control groups . All studies used the ImmunoChip array [17] . P values for type 1 diabetes were from a meta-analysis of a case-control study and familial study using the transmission disequilibrium test ( TDT ) . In order to calculate the correlation between p values for different diseases , we needed to calculate effective numbers of cases and controls for the combined T1D study . For a case control study , under the assumptions of Hardy-Weinberg and the null hypothesis , the variance of the log odds ratio may be expressed as n0+n1n0n11f ( 1-f ) where n0 and n1 are the numbers of cases and controls and f is the minor allele frequency in controls . Given the standard error of a log OR for the TDT study , σ̂ , and a minor allele frequency , we estimated M = σ̂2 f ( 1−f ) for all ImmunoChip SNPs which did not show deviation from the null hypothesis ( p > 0 . 5 ) . The distribution of log ( M ) is shown in S3 Fig . By equating the median of M with n 0 + n 1 n 0 n 1 , and assuming that each TDT family contributed the equivalent information to one control in a case-control study , ie n0 = 2943 , we estimated an equivalent number of cases as 4126 . This seemed reasonable , given that there are a total of 5505 ( dependent ) cases across those families . SNPs were excluded on the basis of QC summaries calculated on 12 , 888 common controls: call rate less than 99% , minor allele frequency less than 0 . 02 , or deviation from Hardy-Weinberg equilibrium ( ∣Z∣ > 5 ) . Given the strong association of immune mediated diseases with the MHC and the extended LD in the region , we were concerned that MHC SNPs might cause inaccurate estimation of pleiotropy . We therefore excluded SNPs in a wide band around the MHC region on chromosome 6 ( co-ordinates 24500000: 34800000 , build NCBI36 ) . After quality control , genotype data was available for at least one phenotype at a total of 110677 SNPs . P values were corrected for genomic inflation using a genomic control algorithm [32] . A set of SNPs known to be unassociated with autoimmune disease was obtained from the Wellcome Trust Case Control Consortium ( WTCCC ) study on reading and mathematics ability . These SNPs were pruned so that none were in LD with r2 > 0 . 2 , and any SNPs within 500 kb of known autoimmune-associated regions were removed . The average degree of inflation was computed for each disease at the remaining 1761 SNPs , and all effect sizes and p values were adjusted accordingly . We assume that the p-values for a phenotype i across all SNPs are instances of a random variable Pi . If pi is an instance of this random variable corresponding to a SNP of interest , the unconditional false discovery rate uFDR ( pi ) is defined as uFDR ( pi ) =Pr ( H0 ( i ) |Pi≤pi ) =Pr ( H0 ( i ) ) Pr ( Pi≤pi|H0 ( i ) ) Pr ( Pi≤pi ) =Pr ( H0 ( i ) ) piPr ( Pi≤pi ) where H 0 ( i ) is the null hypothesis that the SNP of interest is not associated with phenotype i . Given a set of observed p values { p i 1 , p i 2…p i N } for a phenotype i at N different SNPs , and an observed p value pi for a SNP of interest , we estimate this quantity as uFDR^# ( pi ) =pi# ( pvaluespikwithpik≤pi ) /N=ExpectedquantileofpiunderH0 ( i ) Observedquantileofpi ( 1 ) Because we make the approximation P r ( H 0 i ) = 1 , the estimate uFDR^ is a an upwards-biased estimate of uFDR; that is , its expected value is greater than the true uFDR , making it a conservative estimator . We compute the quantity ( 1 ) for each SNP at each phenotype , declaring any SNP for which uFDR^ ( p i ) ≤ α as non-null for phenotype i . Defining V as the number of SNPs falsely declared non-null , R as the total number of SNPs declared non-null , and Q = V/R , a theorem of Benjamini and Hochberg [31] shows the expected false discovery rate E ( Q ) among SNPs with uFDR^ ≤ α is less than α . The cFDR constitutes a natural extension of this idea . We assume that the p-values for two phenotypes i and j across all SNPs are instances of a pair of random variables Pi , Pj . If pi and pj are instances of these variables corresponding to a SNP of interest then the conditional false discovery rate cFDR is defined for the set of SNPs with p values for each phenotype less than or equal to those at this SNP ( as per Andreasson et al [3] ) as cFDR ( pi|pj ) =Pr ( H0 ( i ) |Pi≤pi , Pj≤pj ) =Pr ( H0 ( i ) |Pj≤pj ) Pr ( Pi≤pi|Pj≤pj , H0 ( i ) ) Pr ( Pi≤pi|Pj≤pj ) The estimation of this quantity proceeds in a similar way to uFDR . Given a set of observed p value pairs { ( p i 1 , p j 1 ) , ( p i 2 , p j 2 ) … ( p i N , p j N ) } for two phenotypes i and j at N different SNPs , and an observed p value pair ( pi , pj ) for a SNP of interest , we define N1 as the number of p value pairs with Pj ≤ pj , and estimate the cFDR as cFDR^ ( pi|pj ) =Pr ( Pi≤pi|Pj≤pj , H0 ( i ) ) # ( pairs; ( pik , pjk ) ∈ ( Pi , Pj ) withpik≤piandpjk≤pj ) /N1=ExpectedquantileofpiunderH0 ( i ) amongstpikwithksatisfyingpjk≤pjObservedquantileofpiamongstpikwithksatisfyingpjk≤pj ( 2 ) Again , this estimate is conservative , due to the approximation P r ( H 0 ( i ) ∣ P j ≤ p j ) = 1 . We compute the quantity ( 2 ) for each SNP at each pair of phenotypes , declaring any SNP for which cFDR^ ( p i ∣ p j ) ≤ α ′ as non-null for phenotype i . However , as noted earlier , this does not guarantee that the expected false discovery rate amongst such SNPs is less than α′ . We show that the FDR is controlled at a higher level dependent on the region of the unit square defined by rectangles for which cFDR^ ( x ∣ y ) ≤ α ′ . Our method here diverges from the original method proposed by Andreasson et al , in the use of the expected quantile P r ( P i ≤ p i ∣ P j ≤ p j , H 0 ( i ) ) in place of the p-value pi . If studies share no controls , it can be reasonably assumed that , for a SNP which is null for phenotype i , the p values ( pi , pj ) are independent , so p i ′ = P r ( P i ≤ p i ∣ P j ≤ p j , H 0 ( i ) ) = p i . This is the approach taken by Andreasson et al [3] . We propose a method for computing p i ′ when controls are shared between studies , and the independence assumption above is not valid . Our approach is to compute the related quantity P r ( P i ≤ p i ∣ P j ≤ p j , H 0 ( i ) , H η ( j ) ) , where η is the ( unobserved ) effect size we would observe for a given SNP for phenotype j if the observed MAFs agreed exactly with the population MAFs for that SNP , and H η ( j ) is the hypothesis that Zj ∼ N ( η , 1 ) for that SNP . This quantity can be thought of as the ‘expected quantile’ of pi; that is , the proportion of p values we expect to be less than pi . From the first part of ( 2 ) , we have: cFDR ( pi|pj ) =Pr ( H0 ( i ) |Pj≤pj ) Pr ( Pi≤pi|Pj≤pj , H0 ( i ) ) Pr ( Pi≤pi|Pj≤pj ) ( 3 ) As per Andreasson et al [3] , the quantity P r ( H 0 ( i ) ∣ P j ≤ p j ) is set conservatively at 1 , and the quantity Pr ( Pi ≤ pi∣Pj ≤ pj ) is estimated empirically as the proportion of pairs of observed p values ( p′i , p′j ) with p′j≤pj which also satisfy p′i≤pi . For a given SNP , let η denote the standardised mean allele frequency ( MAF ) difference; that is , the Z value we would compute if the observed MAFs agreed exactly with the population MAFs . We consider η for a random SNP as being an instance of a random variable H , and that the observed z value for that SNP Z∣H = η is distributed as Z|H=η∼N ( η , 1 ) ( 4 ) We further assume that H follows a mixture distribution taking the value 0 with probability π 0 ( j ) and a normal pdf with probability 1 − π 0 ( j ) : H∼0 , p=π0 ( j ) N ( 0 , σ2 ) , p=1-π0 ( j ) ( 5 ) This implies Zj∼N ( 0 , 1 ) , p=π0 ( j ) N ( 0 , 1+σ2 ) , p=1-π0 ( j ) ( 6 ) Thus , given the observed distribution of Zj , the parameters π 0 ( j ) and σj may be estimated by an expectation - maximisation algorithm ( https://gist . github . com/chr1swallace/11421212 ) . We assume as per Zaykin [18] that the distribution of pairs of observed z values ( Zi , Zj ) for a single given SNP is bivariate normal . Denote by H η ( j ) the event that , for a given SNP , the values Zj are distributed as N ( η , 1 ) , with η depending on the SNP . Under our assumption of the null hypothesis H 0 ( i ) for the principal phenotype and a population MAF difference corresponding to η for the conditional phenotype , we have ( Zi , Zj|H0 ( i ) , Hη ( j ) ) ∼N0η , 1ρρ1 ( 7 ) The correlation ρ arises from the shared controls between groups [18 , 19] and is asymptotically equal to ρ=11+N0iN01+N0jN01+N0iNi+N0Ni1+N0jNj+N0Nj ( 8 ) where Ni and Nj are the numbers of cases , N0i and N0j are the numbers of non-shared controls , and N0 is the number of shared controls for the original GWAS for the principal and conditional phenotypes respectively . There is good agreement with the asymptotic correlation when group sizes are greater than 100 [18] . Given equations ( 5 ) – ( 8 ) , the joint distribution of Zi and Zj can be computed under only the assumption H 0 ( i ) . The value of the partial PDF of ( Z i , Z j ∣ H 0 ( i ) ) at ( x , y ) can be derived in a similar way to ( 6 ) : ( Zi , Zj|H0 ( i ) ) ~{ N ( ( 00 ) , ( 1ρρ1 ) ) , p=π0 ( j ) N ( ( 00 ) , ( 1ρρ1+σ2 ) ) , p=1−π0 ( j ) ( 9 ) We now compute the final probability in equation ( 3 ) . Define Pη ( X ) =Pr ( X|H0 ( i ) , Hη ( j ) ) ( 10 ) as the probability of observing events X for a particular SNP with true effect size η ( which may be 0 , corresponding to the general null ) . Then , Pr ( Pi≤pi|Pj≤pj , H0 ( i ) ) =Pr ( Pi≤pi , Pj≤pj|H0 ( i ) ) Pr ( Pj≤pj|H0 ( i ) ) =π0 ( j ) P0 ( Pi≤pi , Pj≤pj ) + ( 1-π0 ( j ) ) ∫-∞∞Pη ( Pi≤pi , Pj≤pj ) f ( η ) dηπ0 ( j ) P0 ( Pj≤pj ) + ( 1-π0 ( j ) ) ∫-∞∞Pη ( Pj≤pj ) f ( η ) dη . ( 11 ) If the distribution of H is estimable by other means , quantity ( 11 ) can be calculated numerically without the assumption that the non-null component of H be normally distributed , at the cost of higher computation time . Under our assumptions , equations ( 6 ) and ( 9 ) enable the fast computation of quantity ( 11 ) by normal CDFs; writing Λ ( ρ , σ2 ) ( zi , zj ) =∫|x|>|zi| , |y|>|zj|N ( 00 ) , ( 1ρρ1+σ2 ) ( x , y ) dxdyλσ2 ( zj ) =∫|y|>|zj|N ( 0 , 1+σ2 ) ( y ) dy ( 12 ) we have Pr ( Pi≤pi|Pj≤pj , H0 ( i ) ) =π0 ( j ) Λ ( ρ , 0 ) ( zi , zj ) + ( 1-π0 ( j ) ) Λ ( ρ , σ2 ) ( zi , zj ) π0 ( j ) λ0 ( zj ) + ( 1-π0 ( j ) ) λσ2 ( zj ) ( 13 ) Because the formula for P r ( P i ≤ p i ∣ P j ≤ p j , H 0 ( i ) ) is differentiable on the unit square , an expression for the expected quantile of pi given an exact value for pj can be computed by taking the partial derivative with respect to pj: Pr ( Pi≤pi|Pj=pj , H0 ( i ) ) =∂∂pjPr ( Pi≤pi|Pj≤pj , H0 ( i ) ) =A+BC where A=π0 ( j ) N ( 0 , 1 ) ( zj ) ∫|x|≥ziN ( ρzj , 1−ρ2 ) ( x ) dxB= ( 1−π0 ( j ) ) N ( 0 , 1+σ2 ) ( zj ) ∫|x|≥ziN ( ρzj1+σ2 , 1−ρ2+σ21+ρ2 ) ( x ) dxC=π0 ( j ) N ( 0 , 1 ) ( zj ) + ( 1−π0 ( j ) ) N ( 0 , 1+σ2 ) ( zj ) where Nμ , σ2 ( x ) denotes the value of the normal pdf with mean μ and variance σ2 at x . Because uFDR^ values are monotonically related to p values , the widely accepted GWAS p value cutoff of 5 × 10−8 corresponds naturally to a cutoff for uFDR^ . For each phenotype i , we set a significance threshold βi for uFDR^ ( p i ) as the lowest possible value of γ for which uFDR^ ( p i ) ≤ γ ⇔ p i ≤ 5 × 10 − 8 . We then applied an analagous approach to cFDR^ . For each pair of phenotypes ( i , j ) , we set a significance threshold αji as the lowest possible value of γ′ for which cFDR^ ( p i ∣ p j ) ≤ γ ′ ⇔ p i ≤ 5 × 10 − 8 . Given the distribution of Pj , it is possible that this could lead to declaring SNPs with pi > 5 × 10−8 , pj ≈ 1 as significant . To avoid this , if αji was larger ( less stringent ) than βi , we set αji = β i . For each ordered pair of phenotypes ( i , j ) , we declared all SNPs with cFDR^ ( p i ∣ p j ) ≤ αji as non-null for phenotype i . This included all SNPs with uFDR^ ( p i ) ≤ β i . We then used a technique described in S1 Text ( section B ) to compute upper bounds c j ( i ) on the false discovery rate amongst SNPs for which cFDR^ ( p i , p j ) ≤ αji . For each phenotype , this gave nine upper bounds , corresponding to each of the nine conditional phenotypes . We compared the degree of pleiotropy between diseases by considering how much the p-value threshold for significance for the principal phenotype changed when conditioning on a small p-value threshold for the conditional phenotype . We used the cFDR^ algorithm to compute the number p i j * such that P ( H 0 ( i ) ∣ P i ≤ p i j * , p j ≤ 5 × 10 − 6 ) = P ( H 0 ( i ) ∣ P i ≤ 5 × 10 − 8 ) ; that is , cFDR^ ( p i j * ∣ 5 × 10 − 6 ) = uFDR^ ( 5 × 10 − 8 ) . We then considered the ratio p i j * / 5 × 10 − 8; that is , the fold increase in significance cutoff after conditioning . We note that because of the fixed value of pj = 5 × 10−6 , the expected false discovery rate amongst the set of SNPs which satisfy P ( H 0 ( i ) ∣ P i ≤ p i j * , p j ≤ 5 × 10 − 6 ) ≤ P ( H 0 ( i ) ∣ P i ≤ 5 × 10 − 8 ) is bounded above by P ( H 0 ( i ) ∣ P i ≤ 5 × 10 − 8 ) , by the Benjamini-Hochberg result . Thus the expected false discovery rate amongst SNPs with P i ≤ p i j * and Pj ≤ 5 × 10−6 is bounded by the same value as the expected false discovery rate amongst SNPs with Pi ≤ 5 × 10−8 . We visualised the ratio p i j * / 5 × 10 − 8 as a heatmap ( S2 Fig . ) . We also produced a network ( Fig . 6 ) , with an edge from vertex i to vertex j if and only if , by conditioning on Pj ≤ 5 × 10−6 , the cutoff for significance for Pi could be increased from 5 × 10−8 to 4 × 10−7 . This cutoff was chosen as the smallest value such that the network was weakly connected; that is , each vertex had an arrow either to it or from it . Cutoffs βi were chosen such that uFDR^ ( p i ) ≤ β i ⇔ p i ≤ 5 × 10 − 8 . SNPs were deemed significant for a principal phenotype i if cFDR^ ( p i ∣ p j ) ≤ α j i for any conditional phenotype j and α j i ≤ β i . The expected false discovery rate amongst SNPs for which uFDR^ ( p i ) ≤ β i is less than βi due to a theorem of Benjamini and Hochberg . However , as discussed above and in S1 Text ( section B ) , the expected false discovery rate amongst SNPs for which cFDR^ ( p i ∣ p j ) ≤ α j i is not necessarily lower than α j i . For each ordered pair of phenotypes ( i , j ) , an upper bound c j i was computed for the expected false discovery rates amongst SNPs with cFDR^ ( p i ∣ p j ) ≤ α j i . The list of SNPs declared non-null for phenotype i was pruned to allow for linkage disequilibrium ( LD ) by listing all SNPs in increasing order of minj ( cFDR^ ( pi|pj ) and stepping through the list from left to right , at each stage removing all SNPs in LD with r2 ≥ 0 . 1 to the right of the current SNP . This ideally leads to the inclusion of at most one SNP from each LD block . A multiple testing problem arises from considering p values for one disease conditioned separately on nine others . Specifically , if the criterion for declaring a SNP non-null for phenotype i is that cFDR^ ( p i ∣ p j ) ≤αji for at least one of the nine possible values of j , then the FDR for all SNPs declared non-null will be greater than the FDR among the smaller set of SNPs for which cFDR^ ( p i ∣ p j ) ≤αji for only one value of j , due to multiple testing . However , this excess FDR is not enough to warrant a Bonferroni ( Sidak ) correction; the cFDR^ ( p i ∣ p j ) values for a phenotype i are highly correlated , as all are in turn highly correlated with pi . A Bonferroni correction tends to remove any advantage in SNP detection gained from cFDR^ , though an advantage may still be seen when only considering one conditional phenotype j . We opted to use a method proposed by Nyholt [36] to correct for multiple testing in SNPs with high LD . We estimated a correlation matrix Ω for potentially non-null cFDR values using Spearman’s rank correlation . The variance of the eigenvalues of Ω , Var ( λobs ) , was computed and used to estimate the effective number of variables Meff according to the equation Meff=1+91-Var ( λobs ) 9 ( 14 ) Note that Var ( λobs ) is between 9 ( completely correlated variables , effectively one test ) and 0 ( completely uncorrelated variables , essentially a Bonferroni correction ) . Denoting by n j i the number of SNPs with cFDR^ ( p i ∣ p j ) ≤ α j i , corresponding to an upper bound on the FDR of c j i , an upper bound for the FDR among all SNPs declared significant for phenotype i was then computed as c0i=Meff∑j=1 . . 10 , j≠icjinji∑j=1 . . 10 , j≠inji , ( 15 ) intuitively , multiplying the expected average number of false discoveries across conditional phenotypes ( c j i n j i ) by the effective number of tests . Values of Meff and c 0 i are shown in S1 Table . | Many diseases have a significant hereditary component , only part of which has been explained by analysis of genome-wide association studies ( GWAS ) . Shared aetiology , treatment protocols , and overlapping results from existing GWAS suggest similarities in genetic susceptibility between related diseases , which may be exploited to detect more disease-associated SNPs without the need for further data . We extend an existing method for detecting SNPs associated with a given disease by conditioning on association with another disease . Our extension allows GWAS for the two conditions to share control samples , enabling larger overall control groups and application to the common case when GWAS for related diseases pool control samples . We demonstrate that our technique limits the expected overall false discovery rate at a threshold dependent on the two diseases . We apply our technique to genotype data from ten immune mediated diseases . Overall pleiotropy between phenotypes is demonstrated graphically . We are able to declare several SNPs significant at a genome-wide level whilst controlling at a lower false-discovery rate than would be possible using a conventional approach , identifying eight previously unknown disease associations . This technique can improve SNP detection in GWAS by re-analysing existing data , and gives insight into the shared genetic bases of autoimmune diseases . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [] | 2015 | A Pleiotropy-Informed Bayesian False Discovery Rate Adapted to a Shared Control Design Finds New Disease Associations From GWAS Summary Statistics |
An increasing number of genes required for mitochondrial biogenesis , dynamics , or function have been found to be mutated in metabolic disorders and neurological diseases such as Leigh Syndrome . In a forward genetic screen to identify genes required for neuronal function and survival in Drosophila photoreceptor neurons , we have identified mutations in the mitochondrial methionyl-tRNA synthetase , Aats-met , the homologue of human MARS2 . The fly mutants exhibit age-dependent degeneration of photoreceptors , shortened lifespan , and reduced cell proliferation in epithelial tissues . We further observed that these mutants display defects in oxidative phosphorylation , increased Reactive Oxygen Species ( ROS ) , and an upregulated mitochondrial Unfolded Protein Response . With the aid of this knowledge , we identified MARS2 to be mutated in Autosomal Recessive Spastic Ataxia with Leukoencephalopathy ( ARSAL ) patients . We uncovered complex rearrangements in the MARS2 gene in all ARSAL patients . Analysis of patient cells revealed decreased levels of MARS2 protein and a reduced rate of mitochondrial protein synthesis . Patient cells also exhibited reduced Complex I activity , increased ROS , and a slower cell proliferation rate , similar to Drosophila Aats-met mutants .
A number of neurological diseases are associated with mitochondrial dysfunction . For example , mutations in the mitochondrial genome have been found in a wide range of disorders including Leber's Hereditary Optic Neuropathy ( LHON ) , Neuropathy , Ataxia and Retinitis Pigmentosa ( NARP ) , Mitochondrial myopathy , Encephalopathy , Lactic Acidosis and Stroke ( MELAS ) , Myoclonic Epilepsy associated with Ragged Red Fibers ( MERRF ) , Nonsyndromic Sensorineural Deafness ( NSSD ) , and Kearns-Sayre Syndrome [1] , [2] . All of these disorders cause some dysfunction of the nervous system . Aside from these mitochondrially encoded genes , there is a growing list of mitochondria-targeted nuclear genes that when mutated cause diseases . These include ( 1 ) components of the respiratory chain/assembly factors [3] , [4] , ( 2 ) genes required for mtDNA maintenance/replication [5] , [6] , ( 3 ) genes that regulate dNTP pools [7] , ( 4 ) genes that regulate mitochondrial morphology/cellular trafficking [8] , [9] , and ( 5 ) genes involved in mtDNA transcription and translation [10] . Mitochondria are critical for energy production and are intricately linked to numerous aspects of cellular function . For example , cell proliferation defects have been reported for several mitochondrial fly mutants [11] , [12] . It has been proposed that Complex I disruption results in reduced cell proliferation caused by the buildup of Reactive Oxygen Species ( ROS ) . ROS are short-lived oxygen radicals that are produced at low levels as a result of impaired electron transport . These ROS can react with proteins , lipids , and DNA resulting in major damage to the cell and its mitochondria [13] . Studies in Drosophila have provided insight into the function of numerous human disease genes [14] . Indeed , work on the fly homologue of the then newly discovered PARK2 gene responsible for Autosomal Recessive Juvenile Parkinson's Disease ( OMIM #600116 ) [15] provided compelling evidence that parkin mutations result in mitochondrial dysfunction and oxidative stress [16] , [17] , [18] , work that was subsequently confirmed in human cells [19] , [20] . Forward genetic screens have also been carried out to isolate genes that cause a neurodegenerative phenotype [21] , [22] . These forward genetic screens may allow us to identify novel genes and help us understand the cellular mechanisms required for neuronal survival . For example , the gene nmnat , whose loss has a strong neurodegenerative phenotype , encodes an important neuroprotective protein that may act as a chaperone [23] , [24] . Interestingly , one of its orthologues in mice has been shown to confer significant neuroprotective effects in several disease models [25] . We decided to reassess the phenotypes of numerous mutants that were isolated in a mosaic eye screen in which we screened for defective electroretinograms ( ERGs ) in mutant photoreceptors on chromosome arm 3R [24] , [26] , [27] . Here we report the isolation and characterization of the Drosophila mitochondrial gene Aats-met ( Aminoacyl-tRNA synthetase-methionine , NP_650348 . 1 ) . We show that a partial loss of Aats-met results in mitochondrial dysfunction and causes a severe and progressive neurodegenerative phenotype . We further show that rearrangements in its human homologue , MARS2 ( Methionyl Aminoacyl-tRNA Synthetase 2 , NP_612404 . 1 ) , are responsible for a human neurodegenerative disease named ARSAL , for Autosomal Recessive Spastic Ataxia with Leukoencephalopathy , or Spastic Ataxia type 3 ( SPAX3 , OMIM #611390 ) [28] .
We reexamined a collection of lethal mutants generated on chromosome 3R to identify mutations that cause a degenerative phenotype [26] . We induced large clones of homozygous mutant tissue in the eyes using the ey-FLP system and screened for flies with aberrant ERGs that significantly worsen with age as a readout for degeneration of photoreceptors [29] . As shown in Figure 1 , we isolated a lethal complementation group consisting of two alleles , HV and FB . Control flies exhibit an “on” transient ( black arrowhead ) upon a flash of light ( Figure 1A ) . A change in potential ensues ( arrow ) , which is followed by an “off” transient ( white arrowhead ) when the light is switched off . The HV and FB mutants produced ERGs with significantly reduced amplitudes ( double-headed arrow ) ( Figure 1B , D ) , suggesting a defect in phototransduction and synaptic transmission . As the flies age , the ERGs exhibit gradually smaller amplitudes in response to light ( Figure 1C , E ) . A less severe genetic combination of alleles that produces adult flies ( see below ) , HV/FB , have normal ERGs at 1 d of age , while 3-wk-old animals ( Figure 1F , G ) have severely affected ERGs . To map the HV and FB mutations we turned to meiotic recombination mapping with P-element lines [30] and deficiency mapping ( Figure S1A–B ) . This pinpointed a 120 Kb region with 18 candidate genes . One lethal mutation , a piggyBac ( PB ) transposon insertion [31] in an intron of the Aats-met gene ( Aats-metc00449 ) , failed to complement the lethality of the FB allele ( Figures 1K , S1B ) . Sequencing revealed that HV and FB affect the Aats-met gene: HV carries a c . 125T>A predicted to result in the missense mutation p . V42D , whereas FB carries a c . 671C>T predicted to result in the missense mutation p . S224L ( Figure 1L ) . Aats-met encodes the uncharacterized Drosophila mitochondrial methionyl-tRNA synthetase , with 44% identity and 75% similarity to its human orthologue MARS2 ( Figure 1L , M ) [32] . Complementation tests with the three alleles and a deficiency ( Df ( 3R ) Exel7321 ) indicate the following allelic series: Df>PB>FB>HV . Flies homozygous for HV or transheterozygous for HV and FB are semi-viable , although they exhibit reduced lifespans ( see below ) . To demonstrate that the phenotypes associated with the mutations are indeed caused by a defective Aats-met gene , we ubiquitously expressed the Drosophila Aats-met and human MARS2 cDNAs using the Gal4/UAS system in mutant backgrounds [33] . The fly and human cDNAs rescued the lethality associated with FB/Df and HV/Df , the strongest allelic combinations . Note that overexpression of these cDNAs in a wild-type background , ubiquitously or only in the eye , results in a wild-type ERG phenotype ( Figure 1J ) . Moreover , the ERGs of aged HV/Df rescued flies are normal ( compare Figure 1C with 1H–I ) , demonstrating that the mutations in Aats-met are indeed responsible for the lethality and ERG defects . These data also indicate that MARS2 and Aats-met are homologous genes as both rescue the Aats-met mutants . We also Flag-tagged the human MARS2 construct at the C-terminus and performed colocalization experiments with the mitochondrial reporter mito-GFP protein [34] in mitochondria of Central Nervous System neurons of 3rd instar larvae ( Figure 1N ) . Both proteins co-localize , indicating that MARS2 is indeed a mitochondrial protein . To assess whether a worsening of the ERG phenotype is due to progressive degeneration of photoreceptor neurons ( PRs ) in Aats-met mutant retina , we performed Transmission Electron Microscopy ( TEM ) of the retinas of flies of different ages . We focused our analysis on transheterozygous escapers ( HV/FB ) and clones of the PB allele . Both have normal ERGs ( Figure 1F ) , with no obvious developmental defects , and possess the correct number of photoreceptors per ommatidium in 1-d-old animals ( Figure 2A–C ) . They display no defects in their rhabdomeres , and the overall appearance of the PRs also appears normal . As shown in Figure 2D–E and 2G , the PRs and support cells ( glia ) progressively degenerate . By 2 wk of age , the PRs of HV/FB animals display more severe phenotypes , and some PRs are vacuolated ( arrowhead , Figure 2D ) . By 3 wk of age , most PRs are severely affected and many organelles are barely recognizable ( Figure 2E , G ) . Similarly , in mutant clones of the piggyBac ( PB ) , PRs are mostly normal at day 1 ( Figure 2C ) and become severely affected by 2 wk of age ( Figure 2F ) . In summary , different mutations cause a severe progressive degeneration of PRs and glia . A careful quantitative analysis of the TEM micrographs revealed some subtle defects in young animals . Indeed , the total mitochondrial area in mutant PRs is greater in 1-d- and 1-wk-old animals ( 2-wk-old animals were too severely affected to quantify ) ( Figure 2H ) . In addition , we also noted many grey spheres in the glia in mutants , indicating the presence of lipid droplets that are not observed in wild-type animals ( black arrowhead , Figure 2B , F ) . That these are indeed lipid droplets was confirmed with toluidine blue staining ( red arrows in Figure S1E–F ) , a possible indication of a fatty acid metabolism defect [35] . In summary , the electrophysiological and ultrastructural features indicate that the mutant photoreceptor neurons undergo progressive degeneration . HV/FB and HV/HV escapers are morphologically normal . They feed , walk , and mate , suggesting that their development and basic physiological features are relatively normal . They , however , have much shorter lifespans than wild-type flies ( Figure 2J ) and are unable to fly . In light of their inability to fly and shortened lifespans , we examined the indirect flight muscles of these flies . Interestingly , the myofibrils seem intact at 1 d of age ( Figure 3A , C ) , but the mitochondria are clearly aberrant: they are larger than normal ( Figure 3C–E ) . In 1-wk-old HV/FB flies , the myofibrils display defects ( arrowhead in Figure 3D ) , and the mitochondria are very large ( Figure 3D–E ) . At 2 wk of age the muscle is too fragmented to take TEM images . Hence , partial loss-of-function mutations in Aats-met impair longevity and mitochondrial morphology . We noted that HV/Df mutants die as late 3rd instars or small pupae , possessing small imaginal discs and larval brains ( Figures 1K , 4A–G ) . Despite their smaller size , mutant larval brains do not show any obvious differences in the immunostaining patterns and localization of neuronal and glial proteins like Elav , Bruchpilot , Fasciclin II , and Repo when compared to wild-type brains ( unpublished data ) . Mutant cells exhibit a proliferative disadvantage when compared to wild-type cells as the mutant clones are significantly smaller than their wild-type twin spots in wing imaginal discs ( Figure 4H–I ) . Moreover , anti-phosphoHistone 3 ( PH3 ) staining , a mitotic cell marker , is decreased by 23% in mutant clones when compared to wild type clones in wing imaginal discs ( Figures 4L and S2A ) , suggesting that cell proliferation is affected . However , cell growth does not seem to be significantly impaired based on staining against the cell membrane marker Dlg ( Figure 4J–K ) . We also observed no difference in the number of apoptotic cells between wild-type and mutant clones based on Caspase 3 staining ( Figure S2B–C ) , and ubiquitous overexpression of the anti-apoptotic protein P35 did not suppress the small larval brain phenotypes ( Figure S2D–G ) . In summary , these data strongly indicate that Aats-met affects cell proliferation but not cell growth and apoptosis in non-neuronal cells . A mitochondria-specific stress response ( UPRmt ) induced by the overexpression of a misfolded mitochondrial matrix protein in mammalian cells has been described [36] and confirmed to be present in C . elegans [37] . In C . elegans , many of the RNAi constructs found to activate the UPRmt correspond to mitochondrial translation factors [38] . Since Aats-met/MARS2 is a mitochondrial translation factor , and since the highly conserved mitochondrial chaperone Hsp60 is a good reporter of the UPRmt in C . elegans , we examined expression of Hsp60 [39] . We observe an elevation in Hsp60 levels in Aats-met mutant clones in the eye ( Figure S3A–B ) as well as in mutant clones in the wing imaginal discs ( Figure S3C–D ) . To determine if the cytoplasmic UPR is affected , we carried out immunohistochemical stainings with BiP/Hsc3 , which has been shown to be a reliable marker in flies for the cytoplasmic UPR [40] , [41] . Unlike Hsp60 , BiP/Hsc3 is not induced in mutant cells , indicating that the two UPR processes are uncoupled ( Figure S3E–F ) . To assess the functional consequence of mutations in Aats-met on oxidative phosphorylation , the rate of oxygen consumption of intact mutant mitochondria was measured in vitro by performing polarography [42] . In the presence of the Complex I–specific oxidizable substrates malate and glutamate , mutant mitochondria exhibit a decreased respiratory control ratio ( RCR ) , the ratio of state III ( ADP-stimulated O2 consumption rate ) to state IV ( ADP-limiting O2 consumption rate ) . The RCR for the most severe allelic combination ( FB/Df ) was significantly lower compared to control mitochondria , primarily due to a relative increase in the state IV rate , likely reflecting a partial uncoupling of oxidative phosphorylation in mutant mitochondria ( Figure 5A ) . Interestingly , the oxygen consumption rates in the presence of the Complex II–specific oxidizable substrate succinate are increased for Aats-met mutant ( FB/Df ) mitochondria compared to controls , while the RCRs remain preserved , possibly indicating a compensatory response ( Figure 5A , Table S1 ) . This is consistent with the finding in C . elegans of increased Complex II–dependent respiration activity when levels of various Complex I components are knocked down with RNAi [43] . Given that the mitochondrial genome encodes 13 polypeptides that are all components of the mitochondrial Electron Transport Chain ( ETC ) ( Table S3 ) , we investigated whether there is a respiratory chain deficiency . To directly assess the individual ETC complexes , enzyme activities of the individual respiratory chain complexes from purified and disrupted mitochondria were measured spectrophotometrically . We observed a significant decrease in Complex I activity ( Figure 5B , Table S2 ) . The partial deficiency of Complex I in mutant mitochondria is relatively mild given that 7 out of the 40 or more Complex I subunits are encoded in the mtDNA and are therefore dependent on mitochondrial protein translation ( Table S3 ) . It has been proposed that high levels of ROS ( primarily superoxide anion ) because of aberrant Complex I activity results in reduced cell proliferation ( Figure 4H–I ) , although low levels appear to promote proliferation [12] , [44] . Hence , we hypothesized that the reduced cell proliferation in Aats-met mutants may be caused by elevated levels of ROS . Since mitochondrial aconitase activity is highly sensitive to ROS [45] , [46] , we measured aconitase activity normalized to total protein levels and found it to be greatly reduced ( Figure 5C ) . Upon addition of a reducing agent , the aconitase activity is restored in the mutants , showing that aconitase is indeed more oxidized in the mutants than in the wild-type controls . One of the mutant phenotypes associated with loss of Aats-met in the eye is very similar to the loss of Pdsw , which affects Complex I [12] . Clones of Pdsw in the eye cause a glossy patch and reduce the eye size . As shown in Figure 5D , Aats-met mutant clones exhibit similar phenotypes . We therefore tested if these phenotypes can be suppressed by antioxidants and supplemented with food with the lipophilic/cell-permeable Vitamin E ( α-tocopherol ) and water-soluble N-acetylcysteine amide ( AD4 ) [47] . We scored the loss of the glossy patch and the number of ommatidia . As shown in Figure 5D–E , low levels of Vitamin E ( 20 µg/ml ) significantly improved eye morphology and size ( p<0 . 001 ) . In addition , the percentage of adult female escapers of the genotype HV/FB able to eclose at room temperature increased significantly with antioxidants , although this was not observed in males ( Figure 5F ) . Note that the doses of Vitamin E and AD4 used had no effect on wild-type eyes or eclosion rates ( unpublished data ) . We noted that the human orthologue of Aats-met , MARS2 , was located in a 3 . 33 Mb candidate interval on chromosome 2q33 . 1 . Some of the authors of this manuscript had previously mapped a neurologic disease named ARSAL to this interval [28] . ARSAL is found in a large cohort of French-Canadian families and is an autosomal recessive spastic ataxia frequently associated with white matter changes as detected by MRI [28] . To examine this region closer , we generated Single Nucleotide Polymorphism ( SNP ) haplotypes using the 300K Illumina SNP-array on selected families . This documented the existence of three different disease carrier haplotypes in French-Canadian ARSAL cases ( Figure S4 ) . Recombination events within families established a minimum candidate interval of 579 Kb ( rs16865262–rs7581202 ) ( black bar in Figure S4 ) , containing nine genes including MARS2 . MARS2 is a single exon gene that spans 3 , 528 bp of genomic DNA and encodes a 593 aa protein homologous to Aats-met [32] . Interestingly , no point mutations were uncovered by genomic or cDNA sequencing . The first mutation was identified by PCR in Family E and consists of a 268 bp deletion predicted to lead to a premature STOP codon at position 236 ( c . 681Δ268bpfs236X ) , referred to subsequently as Dup-Del ( Figure 6A ) . This deletion was confirmed by sequencing in nine patients from different families ( Tables S4 and S5 ) . As shown in Figure 6A , PCR amplification of MARS2 encompassing the first third of the coding sequence revealed the presence of a deleted fragment that segregates in ARSAL Family E ( arrow , E9 , E10 , E11 ) and can also be seen in the father ( E9 ) , who is an unaffected carrier . This deleted fragment is not observed in the mother ( E8 ) and in Family B members , who possess a different type of mutation in the MARS2 gene ( see below ) . The wild-type sequence of the MARS2 PCR products ( Figure 6B ) and DNA sequencing of the amplicon of compound heterozygous case E10 documents the deletion ( compare Figure 6C to 6B ) . This mutation was confirmed by oligonucleotide custom array Comparative Genomic Hybridization ( aCGH ) , as discussed below . Interestingly , the deletion is part of a complex duplication of MARS2 in these patients ( see below ) . In affected brothers E10 and E11 , the aCGH discriminated the presence of a duplication ( black lines/dots above the +2 copies green line in Figure 6D ) in both patients as well as a deletion ( red arrows in Figure 6D–E , compare to Figure 6I ) . Further evidence that mutations in MARS2 were causative came from the identification of a 300 bp insert in the coding sequence that segregated within Family C ( patients C6 and C8 in Figure 6F but not in Family B , which possesses a different mutation—see below ) . The insertion's sequence provided evidence of a complex 5′ mutation , since only a partial sequence of MARS2 was revealed ( Figure 6G , H ) . The presence of repetitive sequences at the 5′ end of MARS2 combined with a 250 bp GC-rich sequence immediately 5′ of the ATG hampered MARS2 full genomic sequencing . This region is 67% GC-rich and contains a 27 bp G/C stretch that is not polymorphic in controls ( [CGGGG]n in Figure 7A ) . The small size of the gene and the limited number of restriction sites prevented us from generating informative Southern blots to further investigate the breakpoints of rearrangements . Nevertheless , quantitative Southern Blot analysis using five additional restriction enzymes ( ApaI , NcoI , XhoI , KpnI , and HindIII ) confirmed the presence of the duplication ( unpublished data ) . Based on our Southern blots , we conclude that the MARS2 breakpoints are >15 Kb away from the wild-type copy of the MARS2 gene . In summary , the presence of two mutations in the MARS2 locus was documented using PCR and a Southern blot-based method . The nature of these two mutations and a third type of mutation ( e . g . , Family B ) is further documented below . To better define the rearrangements , we performed a series of experiments to identify MARS2 copy number variations ( CNVs ) . In order to circumvent the problem of low average SNP densities in the standard Illumina and NimbleGen CGHs , we designed a custom 845 Kb NimbleGen aCGH array encompassing MARS2 with an average probe density of 60 nucleotides ( nt ) to uncover small rearrangements . This high-resolution aCGH was performed on six cases from four families . Note that the MARS2 gene is surrounded by repetitive DNA , specifically Line 1 and Line 2 elements , but also AT- and TTTA-rich segments , as well as [CGGGG] repeats ( Figure 7A ) . Based on haplotype analysis ( Figure S4 ) , at least three duplication events have occurred in our ARSAL cohort ( Figure 7B–C ) . Indeed , evidence of MARS2 duplications was uncovered in all six cases that were tested by aCGH ( Figure 6D , E , I ) . The CGH data analysis established that the 268 bp deletion , described above as the c . 681Δ268bpfs236X mutation , is part of a duplication since most oligonucleotide probes covering the entire coding sequence of MARS2 have a log2 value ( Cy5/Cy3 ) of ∼0 . 5–1 . 0 ( Figure 6D–E ) , whereas compound heterozygous patients should have values of ∼0 . 2–0 . 5 . To determine whether these complex mutations were segregating in all families and were present in other ARSAL patients , we used seven pre-designed ABI-based Copy Number Assays . Four were located in the MARS2 coding region and one in each of the nearby genes PLCL1 , HSP60 , and COQ10 ( Table S4 ) . PLCL1 , HSP60 , and COQ10 do not exhibit CNVs , whereas MARS2 duplications were uncovered in all 54 ARSAL cases belonging to 38 families and were not found in 384 control chromosomes ( Table S5 ) . Similarly , a Brazilian patient with an ARSAL phenotype also carried a duplication ( patient 57 in Table S5 ) . We hypothesized that the duplications may affect MARS2 expression levels . Indeed , Northern blots show the expected mRNA size in all patients ( Figure S5A ) , but qPCR quantification assays revealed an increase in mRNA expression in two compound heterozygous and four homozygous ARSAL patients that carry the common duplication ( Figure 8A ) . In addition to the normal MARS2 mRNA band , we detected small mRNA fragments ( ∼500 bp ) in ARSAL cases but not in the controls ( Figure S5B ) . These bands are suggestive of mRNA instability or aberrant MARS2 mRNA products . Interestingly , PCR primer walking produced different amplicon lengths that are suggestive of microdeletions ranging from 1 bp to 33 bp in the 250 bp GC-rich 5′ region and interspersed L1-type repetitive elements ( Figure 7A ) . The numerous L1 and L2 elements suggest that the duplications were generated by Fork Stalling and Template Switching ( FoSTeS ) [48] . However , due to the repetitive nature of the DNA , we were unable to determine precisely where and in which orientation the MARS2 duplications were located . In summary , our mapping and CNV data convincingly show that the CNVs are responsible for the ARSAL mutations since none of the 384 non-affected individuals show a CNV in the MARS2 locus . In addition , the MARS2 rearrangements do not affect the expression of surrounding genes such as HSPD1 and PLCL1 as assessed by aCGH and quantitative PCR ( unpublished data ) . Further evidence of the rare nature of these mutations is the fact that no CNV events have been catalogued for the MARS2 region in the Database of Genomic Variants ( DGV ) track . Interestingly , a single Yoruba sequence clone from the Human Genome Structural Variation Project Discordant Clone End track was reported to be discordant from the reference sequence [49] . The discordant clone consists of a 726 bp sequence containing a 276 bp L2 element that mapped within the MARS2 coding sequence ( Figure 7A ) and shares the junction breakpoint seen in the ARSAL rearrangements . The CNVs , the quantitative Southern blots , and the Northerns indicate that the rearrangements alter both the dosage of the MARS2 gene and mRNA . Our CNV results and the presence of numerous local repetitive elements support the hypothesis that regional genomic instability has caused template switching during DNA replication ( FoSTeS ) ( modeled in Figure 7B–C ) [48] , [50] as well as recombination errors [48] , [51] , [52] . To explore the impact of the mutations on protein levels , control and ARSAL patient protein extracts were analyzed by immunoblotting with a mouse polyclonal antibody against the N-terminal end of human MARS2 . Despite increased levels of aberrant mRNA transcripts , we find a reduced level of MARS2 protein in all tested patients , ranging from 40%–80% of normal , using mitochondrial proteins encoded in the nucleus as loading controls ( Figure 8C , quantified in 8D ) . Importantly , carriers of the deletion ( but none of the other patients or controls ) produce the expected 24 kDa truncated protein in addition to the normal band ( black arrow in Figure 8C , Figure 7C ) . The level of the truncated MARS2 protein is at least three times higher than the level of the wild-type protein found in controls . The Western blot data combined with Northern blot data argue that some MARS2 transcripts are not translated , possibly because of a post-transcriptional regulatory event such as an RNA-mediated interference of translation ( Figure 7B ) . To test whether mutations in MARS2 affect mitochondrial translation , we pulse-labeled the mtDNA-encoded polypeptides in patient and control immortalized lymphoblast lines as previously reported ( Figures 8B , Table S6 ) [53] , [54] . Of the six patients tested , three showed a translation deficiency . These three patients are homozygous for the common mutation ( Dup1/Dup1 ) ( cases B4 , B5 , and P24 ) and correspond to the most severe cases diagnosed at the ages of 6 , 3 , and 9 , respectively ( Table S5 ) . Two patients with control levels of translation were compound heterozygotes for two different duplications ( EE41 , AA35 ) . These patients were clearly less severely affected and were diagnosed as adults at the ages of 36 and 26 , respectively . In addition , other clinical variables such as loss of walking ability ( Table S5 ) correlate with the extent of the translation defect in lymphoblasts . Despite the relative decrease of MARS2 levels , no effect on the steady-state levels of mitochondrial tRNAmethionine was uncovered ( Figure S6A–B ) , suggesting that the amino-acylation defect does not destabilize the cognate tRNA . To address if and how knockdown of MARS2 in cells affects translation of mitochondrial proteins , we reduced the levels of MARS2 in HEK293 cells with three different shRNAs ( Figure S6C ) . A severe knockdown ( SH-452 ) clearly affects mitochondrial protein translation ( Figure S6E ) , whereas a less severe reduction in MARS2 ( SH-152 ) does not cause an obvious reduction in mitochondrial protein translation when compared to wild-type controls . Similarly , overexpression of MARS2 had no effect on mitochondrial protein translation ( Figure S6D , F ) . Hence , unless the MARS2 protein level is reduced beyond a certain level , levels of mitochondrial translation are not obviously affected . We did not identify a significant difference in MARS2 protein levels between the patients of different genotypes , although most patients with the Dup1/Dup1 genotype have slightly lower MARS2 levels than the other patients ( unpublished data ) . Finally , consistent with our findings in Aats-met mutant flies , cultured patient fibroblasts displayed reduced Complex I activity , increased ROS levels , and concomitantly decreased cell proliferation rates ( Figure 8E–G ) . Finally , we performed an examination of the genotype-phenotype relationship using the age of symptom onset as a measure of the severity of the disease and noted that patients carrying the duplication-deletion tend to have an earlier onset ( Figure 8H ) .
ARSAL exhibits clear inter- and intrafamilial variability reminiscent of Friedreich Ataxia [28] , [65] , [66] . In the present study , we report a group of 54 affected French-Canadian cases belonging to 38 families with a mean age of onset of 24 . 4 ( 2–59 ) in which we uncovered complex genomic MARS2 rearrangements ( Table S5 , Figure 7B–C ) . The mutations are complex genomic MARS2 rearrangements that always include a gene duplication event . Duplications were found with similar breakpoints located in a GC-rich 5′ UTR sequence and in a 3′ non-coding region . The junctions created by the rearrangements are located outside the coding region of MARS2 or other known genes and do not disturb the expression of neighboring genes as demonstrated by CNV assays and quantitative PCR . The 3′ UTR of MARS2 also seems affected by putative disruptions of regulatory elements at the breakpoint junction ( Figure 7A ) . This duplication was neither detected in 384 controls , nor described in the structural variation database . Moreover , in all families for which we have affected and unaffected relatives available for genetic analysis , the presence of the rearrangement ( CNV ) segregated with the disease . These data strongly argue that mutations in MARS2 are the cause of ARSAL , and this in turn is supported by an increase in message levels of MARS2 mRNA , reduced levels of MARS2 protein , and a reduction in mitochondrially translated proteins and Complex I activity in patients . The high prevalence of repetitive sequences at both breakpoint junctions , including many long-interspersed elements ( LINES ) at the 5′ region of MARS2 , and several AT-rich repeat sequences are likely to have mediated the rearrangements ( Figure 7A ) [67] , [68] . Despite the increased mRNA levels , we observed decreased MARS2 protein levels . The increased mRNA levels may be due to the duplications of the gene as well as duplications of regulatory elements in the CpG island at the 5′ end of the MARS2 gene . Consistent with recent studies , analysis of the MARS2 genomic structure reveals a functional CpG island ( Figure 7A ) [69] , [70] . CpG islands act as constitutive promoters of housekeeping genes and are methylated to silence transcription [71] . These findings suggest that the MARS2 duplications may dysregulate transcription , possibly by affecting the size , composition , or methylation ability of these islands . The decrease in protein levels contrasts with the increase in message . The simplest hypothesis is that the gene duplications were caused by FoSTeS , and a small fragment of DNA encoding some of the 5′ or 3′ UTRs was inverted . This inverted segment may affect mRNA stability and/or translation of MARS2 via an RNAi-mediated mechanism . Indeed , FoSTeS has been shown to result in duplicated inverted segments [72] , [73] . Unfortunately , the highly repetitive nature of the DNA surrounding the MARS2 gene did not allow us to document this inversion . Our data suggest that decreased levels of Aats-met/MARS2 protein or protein function lead to a subtle reduction in mitochondrial translation in humans and problems with mitochondrial function in flies and humans . The partial loss of Aats-met protein seems to lead to the accumulation of misfolded proteins in mitochondria , triggering a mitochondrial Unfolded Protein Response ( UPRmt ) ( Figures S3A–D , S3G ) . Mutant flies and patient cells also exhibit abnormal mitochondrial physiology , most notably a rather surprisingly mild reduction in Complex I activity , as well as accumulation of ROS ( Figures 5A–C , 8D–E ) . The reduction in Complex 1 activity is consistent with the observation that 7 of the 13 mitochondrially encoded proteins are incorporated in Complex 1 . The brain tissue of Aats-met mutants contains lipid droplets that are almost never observed in wild type neurons and glia . Such an increase in lipid droplets , potentially related to a lipid metabolism defect , was also recently observed in a 12-y-old girl exhibiting progressive muscle degeneration and autoimmune polyendocrinopathy and was determined to have cosegregating mutations in MARS2's cognate tRNA , mitochondrial tRNAmethionine , as well as COX III [74] , as well as in patients with other mitochondrial diseases such as Leigh Syndrome , Alpers Disease , and Lethal Infantile Mitochondrial Disease [75] . Aats-met/MARS2 mutations do not solely affect neuronal function and survival . Indeed , severe allelic combinations affect cell proliferation , but not cell growth and apoptosis . These data are consistent with the role of increased levels of ROS in the activation of the G1-S checkpoint via the JNK signaling pathway , blocking cell cycle progression [12] . ROS has been shown to play a role in the regulation of the cell cycle , both in its promotion and blockage [44] . Importantly , several of the patient cell lines , similar to what was observed in flies , also exhibit reduced cell proliferation and increased ROS ( Figure 8F–G ) . The clinical features of ARSAL clearly argue that the neurons , glia , and muscles are more affected than other tissues or organs ( Table S5 ) [28] . Indeed , ARSAL patients exhibit ataxia , severe cerebellar and some cerebral atrophy , dystonia , and leukodystrophy . Flies that carry weak allelic combinations also exhibit a progressive demise of the muscles and brain , as can be seen in Figures 2 , 3 , and S1 . In both patient cells and flies we observe decreased levels of Complex I activity and increased levels of ROS . The ability to partially suppress the morphological defects in flies with various antioxidant compounds is noteworthy . Normally , ROS levels are tightly controlled and known to play important roles in signaling pathways , including the HIF-1α , JNK , NFκB , TNF-α , and NADPH Oxidase pathways [76] . The production of excessive levels of ROS may also play a prominent role in other neurodegenerative diseases [77] , [78] , [79] . Finally , Vitamin E deficiency as a cause of an ataxia ( AVED , OMIM #277460 ) further supports a role for ROS in hereditary cerebellar diseases [80] . In conclusion , mutations in Aats-met in flies or reduced levels of MARS2 protein in humans result in aberrant translation of the Respiratory Chain and concomitant production of ROS . These ROS are especially damaging to neurons , as evidenced by our finding that the ERG progression of the Aats-met mutants can partially be suppressed by antioxidants ( unpublished data ) . This ROS also has the effect of reducing cell proliferation , a phenotype that can also be suppressed by antioxidants ( Figure 5D–E ) . Our model is summarized in Figure S7 . It remains to be determined if antioxidants will prove beneficial for ARSAL patients .
All probands and family members underwent a detailed neurological examination by experienced neurologists . All medical records and imaging were reviewed . All families were of French-Canadian ancestry except for one Brazilian family . None of the families were known to be consanguinous . All MRIs were reviewed by J . L . This project was approved by the Institutional Ethics Committee of CRCHUM . Informed consent was obtained from all patients , all family members , and controls . Genomic DNA was extracted from blood or saliva using standard procedures ( Oragene , DNA Genotek ) . Mutagenesis of chromosome 3R was performed as described previously [26] . The genotypes of FB and HV are: y w; FRT82B Aats-metFB/TM3 and y w; FRT82B Aats-metHV/TM3 . P-element/deficiency mapping was performed as described [30] . The genotype of the Df stock is: y w; Df ( 3R ) Exel7321/TM3 , hs-hid [81] . The genotype of the piggyBac is: y w; FRT82B pBacc00449/TM3 , hs-hid [31] . The control strain used was y w; FRT82B isogenized . To generate mutant eye clones , y w eyFLP; FRT82B w+ cl/TM3 was crossed to y w; FRT82B Aats-metFB/TM3 and y w; FRT82B Aats-metHV/TM3 . Transheterozygous escapers were generated in large numbers by raising the larvae/pupae at 18°C . They were subsequently raised at room temperature and transferred and scored every 2–3 d for aging experiments . Heat-shock clones were generated using y w hsFLP; FRT82B ubi-GFPnls/TM6B . For rescue experiments , y w; Act5C-Gal4/CyO was used . The UAS-p35 stock used to inhibit apoptosis has been described [82] . Climbing assays were performed exactly as described [83] . Unless indicated , stocks were obtained from the Bloomington Drosophila Stock Center ( BDSC ) and are listed on FlyBase ( http://flybase . bio . indiana . edu ) . AD4 ( N-acetylcysteine amide ) and Vitamin E ( MP Biomedicals ) were dissolved in standard fly food . The same food batch without drug supplementation was used for the control . ERGs were recorded as described previously [26] . Images of eyes and pupae were taken with a MicroFire camera ( Optronics ) mounted on a Leica MZ16 microscope . TEM of photoreceptors was performed as described previously [24] . At least five animals were analyzed . Thick sections were prepared for inspection of sample integrity . For quantification , 18–20 photoreceptor cartridges for each genotype were analyzed . Thick sections of the optic lobe ( Figure S1 ) were visualized using a microscope ( Imager . Z1; Carl Zeiss , Inc . ) , camera ( AxioCam MRm; Carl Zeiss , Inc . ) , AxioVision release 4 . 3 software ( Carl Zeiss , Inc . ) , and the Plan-Apochromat 20× NA 0 . 75 lens . For sequencing , DNA from mutant larvae was sequenced ( Macrogen ) and analyzed ( DNAStar ) . The Aats-met cDNA ( DGC clone GH13807 ) and the human MARS2 cDNA ( Open Biosystems MHS4426-99239542 ) using iProof polymerase ( Bio-Rad ) and appropriate oligos ( with a Kozak sequence ) were subcloned into the pUAS-attB vector and injected into embryos containing the VK37 attP site [84] . Third instar larvae were homogenized in cold mitochondrial isolation buffer using a Dounce homogenizer ( Kontes ) , filtered through cheesecloth , and centrifuged at 150 G , then 9 , 000 G . Oxygen consumption of mitochondria was measured ( Clark microelectrode ( YSI Life Sciences ) ) , recorded ( PowerLab data recorder ) , and analyzed ( ADInstruments LabChart ) . Rates ( ng atomic oxygen/min/mg mitochondrial protein ) were expressed as percentage control activity . Polarography was performed for six independent mitochondrial isolations . For enzymology , 3rd instar larval mitochondria were sonicated as above . Spectrophotometric kinetic assays were performed ( monochromator microplate reader ( Tecan M200 ) ) . Complex I activity was determined by measuring NADH oxidation ( 340 nm ) , Complex II activity by measuring DCIP reduction ( 600 nm ) , Complex III activity by measuring CytC reduction ( 550 nm ) , Complex IV activity by measuring CytC oxidation ( 550 nm ) , and Citrate synthase activity by measuring DTNB reduction ( 412 nm ) coupled to acetyl-CoA reduction . All activities were calculated as nmoles/min/mg protein and expressed as percentage control . Six independent samples for each genotype were tested . The activity of mitochondrial aconitase was measured on the basis of conversion of citrate into α-ketoglutarate coupled with NADP reduction ( Sigma ) and was normalized for total protein [45] . Activity was measured in the native state and after “reactivation” by incubating mitochondria in ferrous ammonium sulfate for 5 min before performing the assay . In vitro labeling of mitochondrial translation products was performed as described previously [53] . Immunohistochemistry was performed as previously described [85] . Anti-BiP ( 1∶200 ) [41] , anti-Drosophila Hsp60 ( 1∶200 ) [39] , anti-Dlg ( 1∶50 ) [86] , anti-cleaved caspase 3 ( 1∶500 ) ( Cell Signaling ) , anti-PhosphoHistone 3 ( ab5176 ) ( 1∶1 , 000 ) ( Abcam ) , anti-Fasciclin II ( 1D4 ) ( 1∶10 ) [87] , anti-Elav ( 1∶500 ) ( 7E8A10 ) [88] , anti-Brp ( Nc82 ) ( 1∶100 ) [89] , and anti-Repo ( 8D12 ) ( 1∶10 ) [90] were used . Secondary antibodies conjugated to Cy3 , Cy5 , or Alexa 488 ( Jackson ImmunoResearch and Invitrogen ) were used at 1∶250 . For anti-PH3 quantification , homozygous FB clones were stained with anti-PH3 to mark cells undergoing DNA synthesis . The largest box possible was made of the disc , and PH3-positive cells were documented with red dots in heterozygous tissue or purple dots in the homozygous tissue . The area was then determined for both , and paired Student t tests were performed for each of five discs to compare the difference in the number of PH3-positive cells in homozygous tissue versus heterozygous tissue . A total of 20 pairs of clones and their twin spots for each genotype+temperature were measured . A SNP genome-wide scan with the Illumina HAP300 SNP chip was conducted at the Genome Quebec Innovation Center , McGill University ( Montreal , Canada ) on nine affected individuals and six non-affected family members . BeadStudio Software was used as an analysis tool for genotyping , homozygosity , and loss of heterozygosity analysis . Copy number analysis was performed using the PennCNV program . We used seven pre-designed ABI-based Copy Number Assays for human CNV screening; four were located in the MARS2 coding region , one in each coding sequence of the surrounding genes ( PLCL1 , HSPD1 , and COQ10 ) ( Table S4 ) . Each reaction was performed in quadruplicate on a 384-well PCR plate with the ABI Copy Number Reference Assay ( RNaseP ) . CopyCaller ( Applied Biosystems ) was used for data analysis , and all steps were done according to instructions . NimbleGen CGH-array was performed using a chr2 specific fine-tilling oligonucleotide ( HG18 CHR2 FT ) to detect chromosomal changes . The median probe spacing was ∼500 bp . Custom high-resolution NimbleGen's 12×135K CGH arrays ( 38 , 725 probes per array on Chr2 ) were designed to cover the entire 0 . 845 Mb surrounding MARS2 [91] , [92] . The median probe spacing was 1 bp . Primers were designed ( Table S4 ) using Primer 3 or ExonPrimer ( see URLs section below ) . Sequences were analyzed on an ABI3730 Genetic Analyzer ( Applied Biosystems ) . RNAs were treated with DNase I to avoid genomic DNA amplification . Reverse transcription was performed using 3 µg total RNA using random hexamers , OligodT , and Superscript III ( Invitrogen ) according to the vendor's protocol . We prepared cDNAs from total RNA and performed cDNA analysis by PCR with the primers as indicated in the manufacturer's protocol . Purified PCR fragments were subcloned into pCR II-TOPO TA cloning kit ( Invitrogen ) ( Table S4 ) . Quantitative real-time PCR experiments were performed using an ABI PRISM 7900 HT ( Applied Biosystems ) on genomic DNA and cDNAs . Transcript-specific primers were designed with Primer Express software ( Applied Biosystems ) . The PCR conditions and analysis of the obtained data were optimized using published protocols [93] , [94] . The cycle of threshold value ( Ct ) was normalized to the transcripts for the housekeeping genes β-globulin and GAPDH . We performed calculations as described previously [93] , [94] . Primer sequences are shown in Table S4 [95] . Cell lines were maintained under normal condition ( 37°C , 5% CO2 ) in standard culture media ( DMEM containing 10% FBS and 100 µg/ml Pen-Strep and 50 µg/ml gentamicine ) . RNA extraction was performed using TRIZOL ( Invitrogen ) . To measure the fibroblast cell proliferation rate , fibroblasts from the three control and seven patient cell lines were cultured in 12 well plates as described earlier . They were plated at the same pre-determined concentration ( 900 cells/ml ) using a hemocytometer as a guide and were counted using a Beckman Coulter Vi-Cell XR2 . 03 cell viability analyzer after 48 h and then quantified . An N-terminal mouse polyclonal antibody was obtained from Abnova ( MARS2-H00092935-Q01 ) and used at 1∶1 , 000 . We used antibodies against LRPPRC and SLIRP as loading controls . The LRPPRC polyclonal antibody was prepared by Zymed Laboratories ( #295–313 ) and used at 1∶3 , 000 . The polyclonal antibody against SLIRP was used at 1∶1 , 000 ( Abcam #ab51523 ) . Protein was extracted from cultured cells , and 20 µg were subjected to SDS-PAGE and transferred to nitrocellulose membranes ( Millipore ) . The blot was probed overnight at 4°C with the primary antibodies and then probed for 1 h at room temperature with anti-rabbit IgG-HRP secondary antibody ( 1∶10 , 000; Santa Cruz Biotechnology ) . We visualized proteins using ECL Western Blot detection reagent ( PerkinElmer ) . 10 µg of total RNA extracted from control and patient lymphoblasts were run on a 10% polyacrylamide gel containing 7 M urea , followed by transfer to Hybond N+ membrane ( GE Healthcare ) . Pre-hybridization and hybridization were carried out in EXPRESS-Hyb solution ( Clontech ) according to the manufacturer's instructions . The oligonucleotides used for the generation of the 32P-labeled probes had the following sequences: 5′-TGGTAGTACGGGAAGGGTATAACC-3′ for tRNA-Met and 5′-TGGTATTCTCGCACGGACTACAAC-3′ for tRNA-Glu . The commercial cDNA of MARS2 was digested by Xho1/Pst1 ( OriGene; SC100504 ) and oligonucleotides of complement and reversed MARS2 sequences . Southern blot analysis was performed to assess MARS2 genomic rearrangements . Southern blots were produced using standard protocol with control and mutation carrier DNA . The following restriction enzymes for DNA digestion were used: AflIII , ApaI , BamHI , BglII , HindIII , KpnI , NcoI , PstI , and XhoI . A cDNA probe was obtained from commercial human cDNA digested with XhoI/PstI ( OriGene; SC100504 ) . The blots were hybridized with a 32P-labeled MARS2 cDNA probe as described ( http://www . protocol-online . org/cgi-bin/prot/view_cache . cgi ? ID=2746 ) . Statistical analysis was performed using Excel ( Microsoft ) and Prism ( GraphPad ) . Except where otherwise mentioned , unpaired two-tailed Student t tests were used . Percentage protein similarity was determined using BlastP ( NCBI ) . We used sequences for MARS2 with accession numbers NM_138395 . 2 and NP_612404 . 1 . | Neurodegenerative diseases , as a group , are relatively common and often devastating to those who suffer from them . Key insights are emerging from the study of homologues of identified human disease-causing genes in model organisms such as fruit flies , worms , and mice . In this study , we used the fruit fly to identify novel neurodegeneration-causing mutations and identified the Aats-met gene , whose protein product is involved in mitochondrial translation . We found that mutations in this gene cause neurodegeneration , impaired mitochondrial activity , and elevated oxidative stress . We were able to attenuate these defects with antioxidants like Vitamin E . We also determined that unusual duplications in the homologous human gene , MARS2 , were responsible for a novel type of progressive ataxia found in some French Canadian families . Cells taken from these patients have many of the characteristic defects observed in flies , showing that the fly mutants can be used to further explore disease mechanisms and test potential treatments . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"animal",
"models",
"medicine",
"model",
"organisms",
"genetic",
"mutation",
"neurological",
"disorders",
"neurology",
"genetics",
"molecular",
"genetics",
"biology",
"human",
"genetics",
"neuroscience",
"genetics",
"and",
"genomics",
"clinical",
"genetics"
] | 2012 | Mutations in the Mitochondrial Methionyl-tRNA Synthetase Cause a Neurodegenerative Phenotype in Flies and a Recessive Ataxia (ARSAL) in Humans |
The causative agent of cholera , Vibrio cholerae , regulates its diverse virulence factors to thrive in the human small intestine and environmental reservoirs . Among this pathogen’s arsenal of virulence factors is the tightly regulated type VI secretion system ( T6SS ) . This system acts as an inverted bacteriophage to inject toxins into competing bacteria and eukaryotic phagocytes . V . cholerae strains responsible for the current 7th pandemic activate their T6SS within the host . We established that T6SS-mediated competition occurs upon T6SS activation in the infant mouse , and that this system is functional under anaerobic conditions . When investigating the intestinal host factors mucins ( a glycoprotein component of mucus ) and bile for potential regulatory roles in controlling the T6SS , we discovered that once mucins activate the T6SS , bile acids can further modulate T6SS activity . Microbiota modify bile acids to inhibit T6SS-mediated killing of commensal bacteria . This interplay is a novel interaction between commensal bacteria , host factors , and the V . cholerae T6SS , showing an active host role in infection .
Upon entering a human host , a bacterial pathogen is vastly outnumbered by both eukaryotic and prokaryotic cells . Pathogens , including the cholera bacterium Vibrio cholerae , overcome this numerical disadvantage with strategies to either avoid or fight the host immune system and microbiota[1 , 2] . The human pathogen V . cholerae is a marine organism and the causative agent of the diarrheal disease cholera . V . cholerae participates in both immune evasion and in predatory behavior towards phagocytic immune cells and competitor prokaryotes [2] . After being ingested , V . cholerae cells traverse the acidic environment of the stomach and penetrate the mucus layer in the small intestine to reach the epithelial lining . There , V . cholerae cells multiply and secrete cholera toxin , a protein that interferes with adenylate cyclase activity in epithelial cells to cause watery diarrhea in the host and permits dispersal of the multiplying bacteria . During diarrheal purges , high numbers of V . cholerae cells are carried out of the host and into the environment [3] . Thus , cholera toxin is crucial for the spread of V . cholerae after it infects a human host . V . cholerae strains lacking cholera toxin are still infectious . They colonize the small intestine , elicit immune responses typical of cholera , and cause mild diarrhea [4] . Thus , V . cholerae uses cholera toxin to generate severe , watery diarrhea for rapid bacterial dissemination , but relies on additional , and for the most part unidentified , virulence mechanisms to establish and maintain infections . We hypothesize that one such virulence mechanism is the tightly controlled type VI secretion system ( T6SS ) [1 , 2 , 5–7] . The T6SS equips V . cholerae to inject toxic effectors into target cells [1 , 2 , 5–7] of both the same and different species [1 , 8 , 9] . The T6SS provides V . cholerae with the means to cope with eukaryotic predators and prokaryotic competitors in the marine environment , and prokaryotic competitors and host macrophages in the human host [7] . The T6SS has structural homology to the contractile cell-puncturing device of T4 bacteriophages . The current model of the T6SS is a nanotube consisting of stacked hexameric rings of the structural Hcp tube protein , with the nanotube being capped by three VgrG effector proteins ( one VgrG1 , one VgrG2 and one VgrG3 ) at the tip the puncturing device . VgrG1 and VgrG3 contain C-terminal extensions with enzymatic activities that target prey cells [2 , 10] . VgrG proteins can carry out additional functions by binding the PAAR-repeat domain of other effector proteins to attach those effectors to the puncturing device [11] . In addition , the adaptor protein Tap-1 delivers T6SS cargo effectors to the VgrG tip[12 , 13] . The effector-decorated nanotube is believed to dock onto the cytoplasmic side of a baseplate embedded in the V . cholerae inner membrane [5 , 11 , 14–18] . VipA and VipB proteins then form an outer , contractile sheath around the nanotube . Upon contraction of the outer sheath , the Hcp tube , its VgrG tip and attached effectors are ejected from V . cholerae cells and enter neighboring target cells , thereby delivering the T6SS effectors [6] . Structural proteins , including the inner membrane protein VasK , stabilize the T6SS apparatus [18 , 19] . Deletion of vasK results in an inability to secrete Hcp and T6SS effectors [2] . A wide variety of V . cholerae T6SS effectors belong to different classes and differ from strain to strain [5 , 9 , 16 , 20] . All strains analyzed to date have three loci where unique effectors can be found . Pandemic V . cholerae strains all share a set of three cargo effectors , TseL , VasX , and VgrG3 [5 , 9 , 16 , 20] . We refer to the T6SS of pandemic strains as AAA-T6SS , indicating that they all share the same effectors at the same respective loci . AAA-T6SS effectors enzymatically degrade membrane lipids and the peptidoglycan layer of the prey , or insert themselves into the inner membrane to form a pore [10 , 16 , 21] . If neighboring V . cholerae cells engage in T6SS-mediated competition–a phenomenon termed bacterial dueling [22]–both cells can be killed when they use an active T6SS to puncture each other . To prevent the killing of sister cells , immunity proteins are encoded immediately downstream of the T6SS effectors . Immunity proteins are transported into the periplasm where they bind to their cognate T6SS effector delivered by a neighboring cell , thereby abrogating the activity of the effector [10] . Strains in which the genes for these immunity proteins are genetically removed can be used as reporters for T6SS activation , as loss of immunity results in contact-dependent killing under inducing conditions . Although all V . cholerae strains sequenced to date carry the genetic information for the T6SS , not all V . cholerae strains have an active T6SS under laboratory conditions . For example , the O37 serogroup strain V52 ( isolated from a cholera patient in the Republic of the Sudan ) has an active AAA-T6SS under laboratory conditions . Pandemic V . cholerae strains , like the O1 serogroup strain C6706 from Peru , do not express T6SS genes under laboratory conditions [23 , 24] . Zheng et al . [25] recently showed that the T6SS of C6706 is repressed at low cell densities due to the activity of the quorum-sensing regulator LuxO and the transcriptional repressor TsrA ( VC0070 ) . At low cell densities , LuxO is phosphorylated and contributes to the generation of small RNAs . These small RNAs specifically bind mRNA transcripts of the large T6SS cluster , thereby inhibiting the translation of important T6SS structural genes and activators [26] . At high cell density , an unphosphorylated LuxO allows activation of the T6SS in pandemic strains when tsrA is disrupted . This permits bacterial cells to engage in T6SS-mediated virulence , indicating that high density is critical for expression of the T6SS pandemic V . cholerae [25] . The ability to genetically activate the T6SS through luxO and tsrA deletions establishes that pandemic strains have evolved mechanisms to tightly control a fully functional T6SS [25] . The ability to genetically activate the T6SS through luxO and tsrA deletions establishes that pandemic strains have evolved mechanisms to tightly control a fully functional T6SS [25] . We currently do not know when and where pandemic V . cholerae uses the T6SS during its life cycle . Experimental evidence suggests that pandemic V . cholerae strains de-repress T6SS gene expression during infection of the small intestine of infant mice and of humans [27–29] . More recently , V . cholerae was shown to engage in T6SS-mediated killing in infant rabbits [27–29] . The biological significance of in-vivo T6SS de-repression , however , remains unknown . T6SS de-repression has also been shown in V . cholerae cells attached to chitin , which V . cholerae utilizes during colonization of copepods [30] . This activation appears to be coupled with the ability of V . cholerae to become naturally competent , by allowing V . cholerae to kill a prey cell and then take up its DNA[31] . During infection of the human host , V . cholerae is penetrates the mucus layer of the small intestine to make intimate contact with the epithelium and elicit cholera toxin-induced diarrhea [32] . The intestinal epithelium is protected by a thick layer of mucus that consists of a mix of highly glycosylated mucin proteins [33] . Mucins act as a natural barrier against pathogenic and commensal bacteria , with commensals only colonizing the luminal side of the mucus layer [34 , 35] . Unlike commensal bacteria , many pathogens have evolved strategies to overcome the antimicrobial molecules within the mucin barrier as well as to transverse the mucin barrier [36] . V . cholerae expresses the mucin-binding factor GbpA to successfully colonize the murine small intestine [37 , 38] and secretes mucinases HapA and TagA via the type II secretion system [39] to penetrate the mucus layers [40 , 41] . In addition to mucins , V . cholerae is exposed to another prominent host factor , bile . Primary bile acids , including cholic acid and its conjugates glycolic acid and taurocholic acid , are synthesized in the liver and released into the small intestine upon hormonal stimulation . Dehydroxylation of these primary bile acids by commensal bacteria in the gastrointestinal tract generates secondary bile acids; these include deoxycholic acid , glycodeoxycholic acid and taurodeoxycholic acid . Yet other commensal bacteria have the ability to deconjugate bile acids , removing the glycine or taurine and returning the conjugated compound to either cholic acid or deoxycholic acid . While some bile acids are excreted with feces , most undergo enterohepatic circulation; that is , they are absorbed from the terminal ileum ( the most distal portion of the small intestine adjacent to the large intestine ) , transported to the liver for conjugation to either glycine or taurine , and hydroxylated to regenerate primary bile acids . The cycle starts over when primary bile acids are returned to the intestine [42] . Actions of both the host and commensal bacteria contribute to the multifaceted composition of bile in the gastrointestinal tract . Experimental evidence is growing that dehydroxylation of primary bile acids can drastically alter virulence factors produced by bacterial pathogens . For example , deoxycholic acid ( but not cholic acid ) induces virulence gene expression in Campylobacter jejuni necessary for its invasion of macrophages [43] . In contrast , taurocholic acid ( a hydroxylated bile acid ) has been implicated in biofilm dispersal and motility of V . cholerae as well as in increasing tcpA expression [44] . Virulence factors thus can be regulated differently throughout the intestine depending on the bile acids present and , by extension , the presence of different commensal bacteria . In this study , we investigated the molecular basis for host control of the V . cholerae T6SS . We discovered that mucins relieve repression of the T6SS in pandemic strains . The T6SS , now de-repressed , can be then down-regulated by deoxycholic acid , a metabolite of cholic acid created by commensal bacteria . Our findings suggest that mucins , prevalent throughout the gut , activate the T6SS whereas bile acids fine-tune T6SS activity .
Animal studies used in this study ( AUP0000320 ) were reviewed and approved by the Animal Care and Use Committee–Health Sciences at the University of Alberta . This committee adheres to the policies and standards of the Canadian Council on Animal Care . The University of Alberta’s Animal Welfare Assurance Number is #A5070-01 . Bacterial strains were grown overnight on LB ( Luria Bertani ) agar plates supplemented with 100 μg/mL streptomycin ( Sm ) or 100 μg/mL rifampicin ( Rif ) . A derivative of V . cholerae strain V52 ( O37 serogroup; SmR ) lacking hlyA , rtxA , and hapA was used as the wild-type strain in all experiments . The E . coli K-12 strain MG1655 ( genotype F− , λ− , rph-1 , RifR ) was provided by Tracy Raivio ( University of Alberta ) . V . cholerae C6706 ( O1 serogroup; SmR ) and N16961 ( O1 serogroup; SmR ) were provided by John Mekalanos ( Harvard Medical School , Boston MA ) . Anaerobic bacteria Bifidobacterium bifidum ( ATCC 15696 ) and Bifidobacterium subtile ( ATCC 27683 ) were obtained from Cedarlane Laboratories , Burlington ON . Bifidobacterium adolescentis ( ATCC 15703 ) , and Bacteroides thetaiotaomicron ( ATCC 29148 ) were gifts from Alain Stintzi ( University of Ottawa , Ottawa ON ) . Anaerobic bacteria were grown on brain heart infusion ( BHI ) plates or in BHI medium under anaerobic conditions in plastic or metal jars with AnaeroGel sachets ( Oxoid; Hampshire , England ) . In-frame deletions of tsrA and luxO were performed as described in Metcalf et al . [45] . Briefly , a tsrA knockout construct was created using the following primers: A 5´-GGATCCGATTTGCGCTGCTGGTAGGG-3´ B 5´-TCGATTAGCGTTTTTGTAAGGTGGTTAGAGACATGGTG-3´ C 5´-CACCATGTCTCTAACCACCTTACAAAAACGCTAATCGA-3´ D 5´-GGATCCCTGCAAGCTCGCTGTCCC-3´ PCR products resulting from primer combinations A/B and C/D were stitched together by overlapping PCR . The resulting knockout construct was digested with BamHI and cloned into the suicide plasmid pWM91 . The luxO knockout construct was generously provided by David Raskin ( Marian University , Indianapolis IN ) . C6706ΔluxO was used as the parental strain from which tsrA was deleted to create the C6706ΔluxOΔtsrA double mutant . In-vivo competition assays were performed as described previously [46] . Pregnant CD1 mice were purchased from Charles River Laboratories . Animals were treated in accordance with protocols approved by the University of Alberta Animal Care and Use Committee . Briefly , indicated bacterial predator and prey strains were grown on selective LB agar plates overnight . 1 × 109 bacteria were resuspended in a 2 . 5% sodium bicarbonate buffer , mixed at a 1:1 ratio , and fed via oral gavage into 3-day-old CD1 infant mice . After 16 hours , the pups were sacrificed and their small intestines homogenized in 1 mL phosphate-buffered saline . In parallel , this experiment was performed in vitro by adding 50 μL of V . cholerae mixture ( in 2 . 5% sodium bicarbonate buffer ) to 5 mL LB . Cultures tubes were rolled overnight at 37°C . Serial dilutions of the in-vitro and in-vivo samples , and the bacterial inoculum , were plated on LB agar plates supplemented with Sm and 5-bromo-4-chloro-3-indolyl-β-D-galactopyranoside ( X-gal ) to count predator and prey strains . The competitive index was calculated by taking the input ratio ( mutant/wild-type ) and dividing it by the output ratio ( mutant/wild-type ) . Killing assays were performed as described previously [1] . Briefly , indicated bacterial strains were grown overnight on LB agar plates with appropriate antibiotics . Prey and predator were mixed at a 10:1 or 1:1 ratios with titers normalized by OD600 readings . Mixtures were spotted on LB agar plates with indicated supplements and incubated for 4 h at 37°C . Bacteria were harvested and serial dilutions of rifampicin-resistant prey or streptomycin-resistant predator were selectively grown on plates overnight . Sodium deoxycholate , sodium cholate , taurodeoxycholate , glycodeoxycholate , taurocholate , glycocholate , and taurine were obtained from Sigma ( St . Louis MO ) . Glycine was obtained from Thermo Fisher Scientific ( Waltham MA ) . Difco™ Bile Salts No . 3 was obtained from BD ( Mississauga ON ) . For assays of anaerobic bacteria , indicated bacteria were spotted on LB agar plates supplemented with 1 . 2 mM bile acids and incubated for 2 days under anaerobic conditions . Bacteria were scraped from the plates and plates were treated for 15 min with chloroform and incubated for an additional 15 min at 37°C [47] . This procedure ensured that all anaerobic bacteria were dead , and only their metabolites remained . Killing assays were performed on these plates . Overnight cultures of indicated bacterial strains were diluted in LB broth supplemented as indicated and grown to mid-logarithmic phase ( OD600 ~ 0 . 6 ) . Samples were subjected to SDS-PAGE and analyzed by western blotting using a mouse monoclonal antibody against DnaK ( Stressgen Bioreagents , Victoria BC ) , anti-FLAG M2 ( Sigma , St . Louis MO ) , or a rabbit polyclonal antibody against Hcp [7] . For detection , secondary antibodies goat anti-rabbit-horseradish peroxidase and goat anti-mouse-horseradish peroxidase were used ( Santa Cruz Biotechnology , Santa Cruz CA ) . Columns contained 500 μl of 3% ( wt/vol ) bovine submaxillary mucins ( Sigma , St . Louis MO ) or 3% ( wt/vol ) of Difco gelatin ( BD , Mississauga ON ) in Krebs-Ringer Tris buffer [48] . Columns were allowed to settle for 1 h at room temperature . For viability tests , columns were prepared in 1 . 5 ml reaction tubes . Approximately 1 × 108 mid-logarithmic phase bacteria ( 20 μL ) were loaded on top of each column and incubated for 1 h or 2 h at 37°C . One hour after adding supplementing bile acids/amino acids , colony-forming units ( CFUs ) were determined by plating serial dilutions on LB agar plates with appropriate antibiotics . For killing tests , columns were prepared in 1 . 5 ml reaction tubes . Indicated predator and prey strains were loaded on top of each column at a 10:1 ratio and incubated for 2 h at 37°C . CFUs/ml were determined by plating serial dilutions on LB agar plates with appropriate antibiotics . Total RNA was extracted using the TRizol reagent ( Invitrogen ) according to the manufacturer's instructions . RNA concentrations were determined using a NanoDrop spectrophotometer ( Thermo Scientific ) . 1μg of RNA from each sample was treated with DNase I ( Invitrogen ) , and transcribed into cDNA using the SuperScript III Reverse Transcriptase ( Invitrogen ) . Quantitative real-time PCR ( qPCR ) was performed with SensiFAST SYBR No-ROX Kit ( FroggaBio ) , using the CFX96 Real-Time System ( Biorad ) . Thermocycling parameters were as follows: 95°C for 2 min , followed by 40 cycles of 95°C for 15 s and 60°C for 1 min , followed by a melting curve . Primers against the different genes of interest were designed using the PrimerQuest software from Integrated DNA Technologies ( IDT ) . Primers were tested for performance in qPCR with a cDNA concentration gradient , and those with slopes between −3 . 3 and −3 . 7 , efficiency of ∼1 . 0 , and R2 of ∼1 . 0 were used in the qPCR studies ( primer sequences used in the study are summarized in Table 1 ) . The expression levels of the different targets in relation to the endogenous 16S rRNA gene control was determined by the 2−ΔΔCT method using the CFX Manager Software ( Biorad ) . The relative quantification ( RQ ) values of all samples were normalized against the expression of the 16S control for each target . Thin-layer chromatography ( TLC ) was performed as described [49] . Briefly , overnight cultures of B . subtile , B . bifidum , B . adolescentis , or Bacteroides thetaiotaomicron were harvested by centrifugation and resuspended in Ringer’s solution . Reaction mixtures were prepared by dissolving indicated concentrations of bile acids in BHI medium . Bacterial suspensions were added in a 1:1 ratio to the reaction mixture; control solutions were incubated without bacterial suspension . After anaerobic incubation overnight at 37°C , samples centrifuged and the supernatant was filtered , lyophilized and residues were redissolved in 1 mL methanol . After centrifugation , 3 μl of supernatant was spotted onto a TLC sheet ( Polygram™ SilG precoated with 0 . 25 mm silica gel; Machery-Nagel , Germany ) ( stationary phase ) and dipped into the mobile phase . The mobile phase consisted of isoamyl acetate , propionic acid , n-propanol , and water ( 40:30:20:10; Fisher Scientific , Ottawa ON ) . After the run was completed , TLC sheets were dried for 3 min at 110°C , and bile acids were located by spraying sheets with 10% phosphomolybdic acid in ethanol ( Fisher Scientific , Ottawa ON ) . Rf-values were determined by dividing the distance traveled by the bile acid by the distance traveled by the solvent front . Overnight cultures of V . cholerae were diluted 1:100 in plain LB broth or LB supplemented with various concentrations of deoxycholic acid , glycine or taurine . OD600 readings were recorded every hour . The resulting ODs were plotted versus time , and the linear portion of the graph was used to calculate a slope as described in Provenzano et al . ( 2000 ) [50] . This slope was then compared to the bacteria’s slope when grown in plain LB to determine its relative growth rate .
Once inside the host , V . cholerae has to overcome the host defense of commensal gut bacteria that secrete bacteriocins and compete for nutrients and attachment sites . We do not yet understand how V . cholerae outcompetes the commensal host flora despite V . cholerae’s numerical inferiority upon arrival in the gut . We hypothesized that the pandemic O1 strain C6706 establishes an infection by relieving repression of its T6SS to engage in T6SS-mediated competition with the commensal microbiota . To extend recent findings that V . cholerae turns on T6SS genes and engages in T6SS-mediated dueling in the infant rabbit [27 , 29] , we employed the more accessible infant mouse model in this study . The established mouse model for cholera was previously used to study regulation and anti-eukaryotic activity of the V . cholerae T6SS [46] . We co-infected infant mice with wild-type C6706 and C6706ΔtsiV1-3 , which carries in-frame deletions of three T6SS immunity genes . The immunity proteins bind to their cognate effectors . When the immunity proteins are genetically removed , the bacteria are unable to defend themselves from the killing effectors from a T6SS-positive sister cell . T6SS-mediated killing of C6706ΔtsiV1-3 was therefore used as a read-out for T6SS activity . The immunity proteins of C6706 bind the effectors TseL , VasX and VgrG3 . The relative viability of wild-type C6706 remained the same in both in vivo and in vitro assays; in contrast , we observed a significant drop in surviving V . cholerae C6706ΔtsiV1-3 in vivo compared to in vitro ( Fig 1 ) . To determine if the reduced colonization in vivo of V . cholerae C6706ΔtsiV1-3 was due to a T6SS-independent colonization defect associated with the removal of three immunity genes , we performed an additional co-infection experiment with C6706ΔvasK and C6706ΔvasKΔtsiV1-3 . In this scenario , neither strain encodes a functional T6SS . Co-infection with these two strains resulted in no difference in total bacterial numbers ( S1A Fig ) and no difference in survival for either strain between the in vivo and in vitro experiments ( S1B Fig ) . This indicates that reduced colonization by C6706ΔvasKΔtsiV1-3 in the presence of C6706 in vivo as shown in Fig 1 is not due to a T6SS-independent colonization defect , but rather a consequence of succumbing to T6SS-mediated toxicity . These results confirm in-vivo T6SS activity in the infant mouse . The observation that V . cholerae uses its T6SS in infant mouse and rabbit models prompted us to identify host cues responsible for T6SS activation . One feature of the small intestine is its low level of oxygen that fluctuates between anaerobic and microaerobic conditions [51] . To determine if V . cholerae’s T6SS is activated under anaerobic conditions , we performed a killing assay in which we mixed V . cholerae C6706 with E . coli prey under aerobic and anaerobic conditions . Anaerobic conditions did not affect E . coli viability , as the E . coli grew equally well under anaerobic and aerobic conditions in the presence of V52ΔvasK ( a V52 mutant in which the T6SS is disabled by deletion of the T6SS gene vasK ) ( S1C Fig ) . When E . coli was mixed with C6706 , no T6SS killing occurred , suggesting that C6706 maintains a repressed T6SS under anaerobic condition . To determine if an anaerobic environment permits T6SS-mediated virulence per se , we repeated the assay with V52 as the predator , because this strain employs a constitutively active T6SS . We observed that T6SS-mediated killing still occurs in the absence of oxygen , as shown in S1C Fig . This suggests that pandemic V . cholerae senses signals other than the absence of oxygen to activate the T6SS . Although the T6SS of the V . cholerae C6706 strain is inactive under laboratory conditions , RNASeq data suggest that T6SS genes are expressed in vivo [27 , 52] . The T6SS has been demonstrated to be functional and to mediate interbacterial interactions in the infant rabbit [29] and mouse ( Fig 1 ) . However , we do not yet know how the repression of the T6SS is relieved in the host . The mucus layer is the first site of contact for V . cholerae in the small intestine , therefore we hypothesized that mucins ( the main protein components of the mucus layer ) relieve the repression of the V . cholerae T6SS . Interaction with mucins increases V . cholerae motility and successful colonization of the murine small intestine [37 , 53] . To determine if mucins influence the V . cholerae T6SS , we exposed the pandemic O1 strain C6706 or an isogenic mutant lacking three immunity genes ( C6706ΔtsiV1-3 ) to a 3% mixture of mucins or gelatin ( which possesses the same viscosity as mucins ) as a negative control ( Fig 2A ) . In V . cholerae strains such as C6706 , N16961 , and V52 , the immunity proteins TsiV1 , TsiV2 , and TsiV3 confer resistance to the T6SS toxins TseL , VasX , and VgrG3 , respectively [5 , 16 , 20] . We compared numbers of bacteria recovered from either mucin or gelatin columns loaded with one of these three strains . Viable C6706 bacteria were recovered in equal numbers from gelatin or mucin columns , in contrast to the recovery of ~ 1 log fewer viable C6706ΔtsiV1-3 bacteria from mucin columns than from gelatin columns ( Fig 2A ) . With all three immunity genes missing , V . cholerae cells are unable to defend themselves against T6SS-mediated killing by siblings [54] . Thus , the failure of C6706ΔtsiV1-3 to survive in the presence of mucins suggested that mucins relieve the T6SS repression , resulting in a functional T6SS in pandemic strains . We confirmed that death of C6706ΔtsiV1-3 in the presence of mucins is due to sibling-mediated killing and not suicide due to the absence of immunity proteins by genetically disabling its T6SS . An immunity mutant with a disabled T6SS , C6706ΔtsiV1-3ΔvasK , survived in mucins ( Fig 2A ) unless challenged with parental C6706 ( Fig 2B ) , confirming that the viability defect is T6SS-dependent . Furthermore , there is no growth defect of any of these strains as demonstrated by growth curves of C6706 in LB ( S2 Fig ) . These experiments show that mucins are sufficient to coordinate assembly of a functional T6SS , and confirm that immunity genes are required for protection against sibling-mediated killing . To determine if this mucin-dependent phenomenon is a general characteristic of 7th pandemic strains , we tested the effect of mucins on an additional El Tor strain , N16961 , that was independently isolated and represses the AAA-T6SS under laboratory conditions [1] . Similarly to C6706 , N16961 killed C6706ΔtsiV1-3ΔvasK only in the presence of mucins ( Fig 2C ) . Although C6706 and N16961 differ in their degree of T6SS killing activity , this experiment shows that mucins relieve the repression of the T6SS in these two 7th pandemic O1 strains that utilize the same set of T6SS effectors [9] . Bile is abundantly present in the gastrointestinal tract and consists of water , a mixture of different primary and secondary bile acids , fats , inorganic salts , pigments , and immunoglobulins [55] . In response to bile , V . cholerae regulates the expression of the principle virulence factors cholera toxin and toxin-coregulated pilus [56–58] . We investigated whether the bile metabolism pathway ( Fig 3 ) generates bile acids that modulate T6SS activity . Recent work by the Mekalanos group showed that with the LuxO/TsrA repressing circuit disabled , pandemic V . cholerae employs a hyperactive T6SS and causes severe inflammation in the small intestine of experimental animals [25] . We hypothesized that T6SS gene expression might be transient and tightly controlled by host factors in addition to mucins , restricting the activation of the V . cholerae T6SS to specific times and/or locations during infection of the small intestine . As bile acids diffuse through the mucus layer during fat absorption [59] , V . cholerae is likely to be exposed to mucins and bile in the same locale and at the same time . Thus , we tested whether bile acids can modulate the mucin-activated T6SS of V . cholerae O1 serogroup strain C6706 . As shown in Fig 4A , wild-type C6706 and C6706ΔtsiV1–3 ( as an indicator strain for T6SS activity ) were grown either in the presence of gelatin , mucins , or mucins supplemented with cholic- , glycocholic- , taurocholic- , deoxycholic- , glycodeoxycholic- , or taurodeoxycholic acids , or glycine or taurine ( see Fig 3 for metabolic context of each bile salt ) . The only compounds that enhanced mucin-induced T6SS killing of C6706ΔtsiV1-3 kin cells were glycine and taurine ( Fig 4A ) . Deoxycholic acid , but not its precursor cholic acid , exhibited a T6SS-inhibiting role , protecting C6706ΔtsiV1–3 from mucin-induced T6SS-killing . Taurodeoxycholic acid displayed weak T6SS inhibition . To further investigate the modulating activities of bile acids , we took advantage of V . cholerae strain V52 , which has a constitutive T6SS and does not rely on mucins for activation . As shown in Fig 4B , deoxycholic acid also inhibited the V52 T6SS . In contrast to the mucin-activated T6SS of C6706 , glycodeoxycholic acid and taurodexoycholic acid stimulated the V52 T6SS at levels similar to unconjugated glycine and taurine . When deoxycholic acid was supplied in combination with either free glycine or taurine , an intermediate repression of T6SS activity was observed ( Fig 4B ) . This suggests that the carboxylic acid group on deoxycholic acid ( to which glycine and taurine are conjugated ) is important for inhibition of the T6SS of V . cholerae . Activation of the T6SS by free glycine or taurine is a feature of bile acid conjugates and not a general property of amino acids , because the related amino acids alanine and phenylalanine had a slight negative effect on the T6SS of V52 ( Fig 4B ) . We believe that this minor inhibitory phenotype is affecting a different mechanism of regulation than glycine and taurine , because the effects are different both in magnitude and direction . In conclusion , conjugation of either glycine or taurine to deoxycholic acid abolishes inhibition of the T6SS by deoxycholic acid . Next , we investigated how deoxycholic acid is able to downregulate the T6SS of V . cholerae . The T6SS of C . jejuni , an enteric pathogen that utilizes its T6SS for colonization , is also subject to downregulation by deoxycholic acid [60] . C . jejuni strains with an active T6SS display a higher susceptibility to the toxic effects of deoxycholic acid than those without a T6SS . Lertpiriyapong et al . speculated that the T6SS conduit allows diffusion of deoxycholic acid into the C . jejuni cell [60] . They went on to show that C . jejuni cells that survived the toxicity of deoxycholic acid adapted by down-regulating their T6SS [60] . Similar to the Lertpiriyapong et al . study [60] , we performed growth assays using V . cholerae in liquid LB broth with increasing concentrations of deoxycholic acid ( S3A Fig ) . We determined that toxicity did not increase for T6SS-positive bacteria compared to an isogenic mutant in which the T6SS was genetically disabled . In addition , we observed no deoxycholic acid-mediated toxicity when V . cholerae were grown on nutrient agar plates for our killing assay ( S3B Fig ) , suggesting that the mechanism of down-regulation by deoxycholic acid of the T6SS is different than the adaption showed in C . jejuni . To determine if deoxycholic acid regulates T6SS machinery on a transcriptional level , we first checked the transcript levels of hcp in bacteria grown in the absence or presence of various bile salts; we found that none of the bile salts changed Hcp transcript levels ( Fig 5A ) . This matched data demonstrating that deoxycholic acid does not affect cell-associated Hcp protein levels ( Fig 5C ) . We then performed qPCR to check for altered transcript levels of additional T6SS genes in the presence of deoxycholic acid . Again , we did not observe dramatic changes in mRNA levels for structural ( vasK ) , regulatory ( vasH ) , or effector ( tseL , vasX , and vgrG3 ) components ( Fig 5B ) . This indicates that deoxycholic acid either affects gene expression of another T6SS gene or prevents T6SS complex formation . To visualize the effect of deoxycholic acid on the T6SS , we utilized V . cholerae 2740–80 with an sfGFP labeled vipA and performed super-resolution microscopy to visualize the T6SS . V . cholerae 2740–80 is a nontoxigenic strain , which similar to V52 employs a constitutively active T6SS under laboratory conditions and has been used . We then incubated this bacterial strain on LB agar pads , LB pads mixed with deoxycholic acid , and LB pads mixed with taurine for 30 min before visualizing the T6SS . We determined the number of bacteria in the field of view and the number of T6SS tubes . We observed that tube formation in strains incubated on LB pads mixed with deoxycholic acid was reduced approximately 12 . 5-fold compared to bacteria incubated on plain LB pads ( Fig 5D ) . This indicates that although T6SS-related genes do not appear to be down-regulated in the presence of deoxycholic acid , T6SS dynamics are decreased . We observed an insignificant increase in tube formation upon incubation with taurine . We hypothesize that the T6SS of 2740–80 is already hyperactive and unable to increase its activity . We hypothesize that deoxycholic acid regulates the T6SS by destabilizing the macromolecular tube complex , which prevents delivery of toxic effectors to neighboring cells . During colonization of the small intestine , V . cholerae comes in close contact with the microbiota that inhabits the mucus layer [34] . Further , as V . cholerae is cleared from the small intestine in purges of watery diarrhea , V . cholerae enters the colon and comes in contact with other microbes . Many of these commensal bacteria , in the small intestine and the colon have the ability to modify bile acids via dehydroxylation or deconjugation [61] ( Fig 3 ) . Bacteroides and Bifidobacterium are among the major anaerobic bacteria in the gut that participate in bile acid metabolism [62–64] . As the microbiota can affect bile acid composition , and activity in the T6SS of V . cholerae is influenced by bile acids , we investigated the effect of microbiota bile acid metabolism on activation or inhibition of the T6SS . The ability of four commensal bacterial species to modify bile acids and subsequently affect the T6SS activity of V . cholerae was analyzed by thin-layer chromatography ( TLC ) and killing assays . TLC separates bile acids based on their ability to bind to the stationary phase of a chromatography plate . Using purified bile acids as positive controls , we determined the ability of bacteria to convert one bile acid to another . We mixed individual bile acids with Bifidobacterium bifidum or Bifidobacterium adolescentis and ran the cell-free mixtures on TLC plates ( Fig 6A , 6B & 6C ) . By comparing the resulting bands with our positive controls , we could see conversion of bile acids . For the killing assay , commensal bacteria of the species Bacteroides thetaiotaomicron , B . bifidum , B . adolescentis , and Bifidobacterium subtile were anaerobically grown on LB plates supplemented with either unconjugated or conjugated cholate or deoxycholate derivatives . After commensals were removed by chloroform vapor treatment [47] , killing assays were performed under aerobic conditions on the pretreated plates with V52 or V52ΔvasK predator against prey E coli MG1655 ( Fig 6D & 6E ) . B . bifidum , a Gram-positive bacterium prevalent in the human intestine [65 , 66] , dehydroxylated and deconjugated bile acids ( Fig 6A & 6D ) . As shown by TLC , B . bifidum deconjugated glycodeoxycholic or taurodeoxycholic acids to deoxycholic acid , and glycocholic or taurocholic acids to cholic acid . In addition , B . bifidum dehydroxylated cholic to deoxycholic acid . T6SS-mediated killing of E . coli on plates supplemented with glycodeoxycholic or taurodeoxycholic acids and previously treated with B . bifidum ( Fig 6D ) was reduced compared to plates not treated with B bifidum . As expected , conversion of glycocholic or taurocholic acids to cholic acid had no effect on the T6SS . A similar inhibition of T6SS-mediated killing was observed on plates supplemented with cholic acid and previously treated with B . bifidum , because cholic acid is dehydroxylated to deoxycholic acid in the presence of B . bifidum . The killing activity decreased to a level comparable to that observed in the presence of deoxycholic acid ( the end product of glycodeoxycholic acid or taurodeoxycholic acid deconjugation and cholic acid dehydroxylation ) . We conclude that B . bifidum negatively regulates T6SS activity through the metabolic conversion of each of three bile acids , glycodeoxycholic , taurodeoxycholic , or cholic acid , to deoxycholic acid . Similarly to B . bifidum , the two commensals B . adolescentis and B . subtile also modified the effects of selected bile acids on T6SS activity . B . adolescentis and B . bifidum both deconjugate glycocholic or taurocholic acids to cholic acid , and glycodeoxycholic or taurodeoxycholic acids to deoxycholic acid . However , B . adolescentis is unable to dehydroxylate cholic to deoxycholic acid . We conclude that the decrease in T6SS activity ( Fig 6E ) was due to the deconjugation of glycodeoxycholic or taurodeoxycholic acids to deoxycholic acid by B . adolescentis ( Fig 6A and 6B ) . B . subtile has a more specific metabolic activity and can only deconjugate glycodeoxycholic acid to deoxycholic acid ( S4A Fig and S4B Fig ) , resulting in a decrease in T6SS activity ( S4E Fig ) . The reduced V . cholerae killing activity on LB plates previously treated with B . adolescentis and B . subtile was analogous to what was observed in plates previously treated with B . bifidobacterium , suggesting that reduced killing is due to the bile acid-converting activities of commensals . The last commensal analyzed , B . thetaiotaomicron , did not deconjugate or dehydroxylate any of the bile acids as determined by TLC ( S4A Fig and S4D Fig ) . However , the killing of E . coli was marginally inhibited on plates containing glycodeoxycholic acid and treated with B . thetaiotaomicron ( S4D Fig ) . This observation could be explained by trace amounts of glycodeoxycholic acid conversion to deoxycholic acid undetected by TLC . We conclude that metabolic products of bile acids produced by the host inhibit the T6SS of V . cholerae .
The marine bacterium V . cholerae thrives in a wide variety of environments and has evolved mechanisms to sense cues that control host colonization and virulence factors in a spatiotemporal fashion . Together with the recent findings of the Mekalanos group , our in-vivo experiments demonstrate the importance of the host environment in the T6SS activation of pandemic V cholerae strains ( Fig 1 ) [27 , 29] . Host factors responsible for in-vivo activation of T6SS for pandemic strains were unknown . Our finding that the T6SS is functional under anaerobic conditions ( S1C Fig ) prompted us to identify host cues for T6SS activation and to investigate the role of anaerobic commensal bacteria in regulating the T6SS . Our findings that mucins activate the T6SS and that activation by mucins can further be modulated by bile acids under the metabolic control of commensal microbiota provide new insights into the complex regulation of the T6SS in vivo . We hypothesize that the activated T6SS is used by V . cholerae to counteract host defense cells and to compete with other bacteria for nutrients and space during infection of the host small intestine . These other bacteria could be commensals , other pathogens , or members of the same species . V52 and C6706 regulate their T6SS differently . V52 employs a constitutively active T6SS , whereas C6706 has a repressed T6SS that is activated in vivo . Diversity of T6SS regulation among V . cholerae strains might indicate a diversity of biological function between pandemic strains and those associated with smaller outbreaks . T6SS regulators are likely utilized differently in V52 and C6706 . If bile acids target these regulators , modulation of the T6SS by bile acids would be strain-dependent as observed in Fig 6 and S4 Fig . C6706 and other pandemic strains have a T6SS that is repressed under laboratory conditions and potentially in other stages of the lifecycle . However , as shown recently for chitin and now for mucins , pandemic V . cholerae strains can de-repress their T6SS to compete within their species and with other prokaryotes [31] . V . cholerae is often introduced to the human host by the ingestion of contaminated water , thus multiple strains may launch an infection . Activation of the T6SS in pandemic strains by mucins may allow the pandemic V . cholerae to kill competing V . cholerae strains to become the dominant agent of infection . Our findings suggest that products of bile acid metabolism in commensal bacteria have roles in regulating V . cholerae virulence factors . Such effects would be expected to differ among human hosts depending on host microbiota composition . Hsaio et al . recently shed light on how members of the microbiota modulate V . cholerae infection [48] . They showed that different members of the human microbiota were able to prevent V . cholerae infection . Our findings identified additional members of the human microbiota to develop host-based therapies that minimize the effects of a V . cholerae infection . There is precedence for a role for bile in controlling pathogen persistence . Fecal microbiota transplantation was recently proposed to prevent recurrence of C . difficile infection by correcting bile acid metabolism [67] . Thus , experimental support is emerging for the idea that the host microbiota composition determines the course of C . difficile infection [68 , 69] . Furthermore , modification of host factors such as bile and mucins might help a microbiota adapt to defend itself against pathogenic bacteria . Alternatively , as the commensal organisms utilized in this study were mainly colonic , V . cholerae may utilize bile salts as a spatial signal such that it recognizes deoxycholic acid as a signal to turn off its T6SS in the colon before being released into the environment . Locations in the gut where bile-deconjugating and-dehydroxylating commensals are absent may experience V . cholerae bacteria with higher T6SS activity than in locations where these commensal species are present . Therefore , preventive alteration of the microbiota ( through the addition of Bifidobacterium ) in people living in areas where cholera is endemic may disrupt V . cholerae’s T6SS , leading to a less fit organism and a reduction in disease outcome . Although our work does not demonstrate a necessity for the T6SS in colonization of the infant mouse , we do see competition indicating a role for the T6SS in intraspecies competition during a multi-strain infection . To test the efficacy of these probiotics , we suggest using the adult mouse model of cholera infection [70 , 71] . The adult mouse model can be used to study this hypothesis over a longer course of infection and in the presence of bile , as bile is believed to be absent from the infant mouse [65] . As commensal organisms such as Bifidobacterium and Bacteriodes are lost through diarrheal purges during acute V . cholerae infection [62] , the levels of bile acids repressing the T6SS might decrease . This would allow the pathogen to utilize its T6SS in a re-infection of patients recovering from a recent cholera episode , or during later stages of the purge , providing an opportunity for new therapies that introduce Bifidobacterium species at an early stage of V . cholerae infection to decrease pathogen T6SS activity . In conclusion , V . cholerae infection is complex and involves host factors such as bile , mucins , and a microbiota that have impact on the pathogen and regulation of its T6SS . This work describes a novel mechanism for two-pronged regulation of a virulence pathway in V . cholerae , through mucins that activate the pathway and bile metabolites that repress it . This work adds to our understanding of how pandemic V . cholerae strains have evolved as such successful pathogens . | The type six-secretion system ( T6SS ) is a molecular syringe that many Gram-negative pathogens use to kill other bacteria , including commensal bacteria of the human gut . We investigated how the environment of the intestine , specifically commensal bacteria , the mucus lining , and bile affect the T6SS of the bacterial pathogen Vibrio cholerae . First , we showed that the mucins , a family of proteins ubiquitously found in the intestine , activate the T6SS thereby allowing V . cholerae to kill other bacteria . Second , we showed that the magnitude of killing is regulated by bile acids . Certain bile acids produced by the host decrease the killing of bacteria by the V . cholerae T6SS . Last , we demonstrated that prominent members of the host microbiota metabolize these bile acids that enhance bacterial killing by V . cholerae into bile acids that diminish the bacterial killing effects of the T6SS . Our study suggests that the gut microbiota is an important first line of defense against bacterial pathogens , and that this line of defense may be impaired in individuals in poor health . Promoting a healthy microbial environment in the gut could play a role in counteracting cholera by reducing the ability of Vibrio cholerae to compete in the gut . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [] | 2015 | Bile Salts Modulate the Mucin-Activated Type VI Secretion System of Pandemic Vibrio cholerae |
Yersinia pestis appears to be maintained in multiple , geographically separate , and phylogenetically distinct subpopulations within the highlands of Madagascar . However , the dynamics of these locally differentiated subpopulations through time are mostly unknown . To address that gap and further inform our understanding of plague epidemiology , we investigated the phylogeography of Y . pestis in Madagascar over an 18 year period . We generated whole genome sequences for 31 strains and discovered new SNPs that we used in conjunction with previously identified SNPs and variable-number tandem repeats ( VNTRs ) to genotype 773 Malagasy Y . pestis samples from 1995 to 2012 . We mapped the locations where samples were obtained on a fine geographic scale to examine phylogeographic patterns through time . We identified 18 geographically separate and phylogenetically distinct subpopulations that display spatial and temporal stability , persisting in the same locations over a period of almost two decades . We found that geographic areas with higher levels of topographical relief are associated with greater levels of phylogenetic diversity and that sampling frequency can vary considerably among subpopulations and from year to year . We also found evidence of various Y . pestis dispersal events , including over long distances , but no evidence that any dispersal events resulted in successful establishment of a transferred genotype in a new location during the examined time period . Our analysis suggests that persistent endemic cycles of Y . pestis transmission within local areas are responsible for the long term maintenance of plague in Madagascar , rather than repeated episodes of wide scale epidemic spread . Landscape likely plays a role in maintaining Y . pestis subpopulations in Madagascar , with increased topographical relief associated with increased levels of localized differentiation . Local ecological factors likely affect the dynamics of individual subpopulations and the associated likelihood of observing human plague cases in a given year in a particular location .
Yersinia pestis is one of the most successful bacterial pathogens known . Its most recent common ancestor ( MRCA ) may have emerged less than 6 , 000 years ago yet it appears to have been widely dispersed throughout Eurasia during the Bronze Age . Following the acquisition of some key genetic changes only ~3000 years ago , some Y . pestis became capable of causing the deadly , flea-borne bubonic plague [1] and then swept the known world in three recognized pandemics [2] . These pandemics likely originated from Asia in multiple successive waves [3 , 4] , causing hundreds of millions of human deaths and establishing a number of global enzootic foci [2] . Indeed , Y . pestis has successfully spread to every continent except Antarctica and currently has established enzootic foci in Asia , Africa , and the Americas [2 , 5] . Y . pestis is thought to persist in these enzootic foci through low-level cycling in numerous , often cryptic , rodent species whose populations either include a mixture of relatively resistant and highly susceptible individuals or are characterized by a high replacement rate . Periodically , Y . pestis emerges from these enzootic reservoirs in large-scale epizootics involving massive die-offs of highly susceptible rodent species that serve to amplify and spread Y . pestis [2 , 6 , 7] . Alternating enzootic and epizootic cycles , human involvement , and ecology have all contributed to the observed phylogeography of Y . pestis and its consistent pattern of the spread of one to a few genotypes followed by localized differentiation . Globally , this pattern is most clearly observed with the highly successful 1 . ORI population , which was responsible for the third pandemic . In the global Y . pestis phylogeny , the basal node of this population is characterized by a large polytomy , suggesting that a rapid expansion preceded the spread of this group around the world [3] . Numerous independent lineages branch off from this initial polytomy , reflecting the independent evolution of Y . pestis in the many new enzootic foci that were established after this event [3 , 4] . Importantly , the global spread of this group would not have been possible without the inadvertent human transport , via steamship and other means , of rats and fleas infected with Y . pestis [2 , 5] . In addition , the local ecology of regions where Y . pestis was introduced strongly influenced the establishment of stable enzootic foci . During the third pandemic , new plague foci became established in locations that contained either suitable native rodent species ( e . g . , North America ) or a large enough population of non-native rodents ( e . g . , Madagascar ) that could sustain the rodent/flea transmission cycle . In contrast , locations without these conditions ( e . g . , Australia ) , did not develop enzootic foci , although they did experience outbreaks , which subsided after the number of non-native rodents was reduced , resulting in an evolutionary dead end for Y . pestis in these locations [5] . The influence of enzootic/epizootic cycling , human involvement , and ecology on Y . pestis phylogeography is also apparent on a regional level , such as in the well-studied plague foci of Madagascar . Two large areas in the central and northern highlands serve as traditional plague foci in Madagascar [8 , 9] , with the persistence of Y . pestis in these areas linked to the presence of two flea vectors , Xenopsylla cheopis and Synopsyllus fonquerniei , which are less abundant and absent , respectively , at lower elevations [8 , 10 , 11] . A third focus in the port city of Mahajanga experienced several outbreaks when Y . pestis was first introduced to Madagascar [8 , 12] and then again in the 1990s [8 , 13–16] . However , this focus does not appear to be stable , as evidenced by the 62 year gap in observed plague activity between the initial and 1990s outbreaks [8 , 13–16] and the apparent lack of current activity , based on the absence of additional confirmed human cases [8] . Within the traditional foci , Y . pestis appears to be maintained in multiple , geographically separate , and phylogenetically distinct subpopulations that are likely sustained by the black rat ( Rattus rattus ) [17–19] , the primary plague host in rural Madagascar [8–12] . There is also evidence of multiple , likely human-mediated , long-distance dispersal events of different genotypes to new locations , with at least one such event responsible for the re-emergence of the Mahajanga focus during the 1990s [18 , 19] . The spread of one to a few genotypes followed by localized differentiation is a well-established phylogeographic pattern of Y . pestis , at multiple geographic scales [3 , 4 , 18–20] . In Madagascar , there are multiple , geographically and phylogenetically distinct subpopulations that have arisen due to this localized differentiation [17–19] . However , the dynamics of these locally differentiated subpopulations through time are mostly unknown . Previous studies have suggested that some subpopulations experience extinction and/or decreases in frequency and that new subpopulations emerge and spread , potentially becoming established in new locations , either temporarily or more long-term [17–19] . In addition , a temporal study of the 1990s Mahajanga outbreaks depicted a striking cycling pattern of diversity generation and loss that occurred during and after each outbreak , consistent with severe inter-seasonal genetic bottlenecks and large seasonal population expansions [18] . However , this type of concerted temporal analysis has not been attempted in the traditional foci . Here , we investigate the phylogeography of Y . pestis in Madagascar over an 18 year period from 1995 to 2012 . We generated whole genome sequences for an additional 31 strains , enabling us to use a total of 37 Malagasy strain sequences to discover additional SNPs that we used in conjunction with previously identified SNPs and multiple-locus variable-number tandem repeat ( VNTR ) analysis ( MLVA ) to genotype 773 Malagasy Y . pestis samples from 1995 to 2012 . We then spatially map these samples through time on a fine geographic scale to examine Y . pestis phylogeographic patterns in Madagascar through time .
The DNAs used in this study ( S1 Table ) were extracted from Y . pestis cultures or complex human clinical samples originally isolated or collected , respectively , by the Malagasy Central Laboratory for Plague and Institut Pasteur de Madagascar as part of Madagascar’s national plague surveillance plan overseen by the Malagasy Ministry of Health . This program requires declaration of all suspected human plague cases and collection of biological samples from those cases . These samples and any cultures or DNA derived from those samples are all de-linked from the patients from whom they originated and analyzed anonymously if used in any research study , such as this one . The Northern Arizona University Institutional Review Board did not require additional review of this research due to the anonymous nature of the samples . DNA was obtained from 773 Y . pestis strains or complex human clinical samples collected from 1995 through 2012 ( S1 Table ) . Geographical origin data for these samples was very comprehensive , including at least commune and district of origin , with most ( N = 729 ) also including the fokontany ( i . e . , village ) of origin ( each commune is divided into fokontany ) . The samples originated from 384 fokontany , from 175 communes , in 32 districts in Madagascar ( See S1 Fig for a map of the sampled districts ) . The DNAs included 173 and 85 samples that were previously analyzed in references [18 , 19] and [17] , respectively ( S1 Table ) . The remaining 515 novel DNAs were extracted from strains selected to emphasize districts Betafo , Mandoto , Antsirabe I , Antsirabe II , and some neighboring areas ( hereafter referred to as the Betafo region ) , which experience some of the highest human plague case incidence rates in Madagascar; and district Moramanga and neighboring areas ( hereafter referred to as the Moramanga region ) , which also declares human plague cases nearly every year , but at a lower frequency than the Betafo region ( S1 Fig ) . Indeed , the analysis of these areas was very comprehensive , including all of the available samples from the Betafo , Mandoto , Antsirabe I , Antsirabe II , and Moramanga districts , and subsets of samples from the surrounding districts over the 18 year study period ( S1 Fig , S1 Table ) . DNAs consisted of simple heat lysis preparations , extracts prepared using the QIAamp DNA Mini Kit ( Qiagen , Hilden , Germany ) , or whole genome amplification ( WGA , QIAGEN , Valencia , CA ) products generated from the heat lysis or kit extraction preps . Most of the samples ( 90% ) were obtained from human plague cases with a smaller number collected from other mammals or fleas ( S1 Table ) . All DNAs were genotyped , as previously described , using 63 assorted , previously identified SNPs [18 , 19] and a 43-locus MLVA [20] . Screened SNPs included Mad-08 through Mad-48 from reference [19] and Mad-57 through Mad-78 from reference [18] ( S2 Table ) . These SNPs were screened in a hierarchical fashion , with SNP Mad-43 screened first to determine if a sample belonged in Group I or II , two previously described major groups in Madagascar [4 , 19] , and then additional , appropriate Group I or II SNPs screened to determine which previously described SNP-defined group ( i . e . , node ) a sample belonged to . MLVA was then used to provide additional discrimination within each node . A total of 31 strains were selected for whole genome sequencing to identify additional SNPs for phylogenetic analysis ( S1 Table ) . These strains were chosen based on the quality of the available DNA ( e . g . , DNAs extracted from complex human clinical samples and many of the heat lysis preps proved unsuitable for whole genome sequencing due to low concentrations ) and also to maximize the potential for new node discovery by selecting phylogenetically diverse strains with an emphasis on existing nodes containing larger numbers of samples , as determined from the above SNP and MLVA analyses ( S2 Fig ) [21] . Illumina sequence libraries were prepared as previously described [22] and the new genomes sequenced on the Illumina HiSeq platform ( Illumina , San Diego , CA ) , producing 2 × 100 bp reads . Paired-end Illumina whole genome sequence data for each newly sequenced strain and previously published whole genome sequences for six other Malagasy strains ( MG05-1020 [GenBank: AAYS00000000] , IP275 [GenBank: AAOS00000000] , 53/91 , 64/91 , 154/98 B , 17/99 B [GenBank: SRP017903] ) [4 , 18] were aligned using BWA-MEM v0 . 7 . 5 against the published genome for strain CO92 [23–25] . Duplicate regions were identified and removed based on a self-alignment of the CO92 genome using NUCmer v3 . 23 [26] . SNPs were called on the binary alignment map ( BAM ) file [27] using the UnifiedGenotyper method in GATK v2 . 7 . 5 [28 , 29] . SNPs below a minimum depth ( 10x ) or minimum allele proportion ( 90% ) were removed from subsequent analyses . Alignment and SNP calling methods were wrapped by the Northern Arizona SNP Pipeline ( NASP ) ( http://tgennorth . github . io/NASP/ ) [30] . Primers were designed targeting a ~250 bp region around each of 188 newly identified potential SNP targets ( S3 Table ) using Primer3 [31 , 32] with strain CO92 as the reference sequence [24] and with the potential SNP located at the center of each amplicon . Primer sets were ordered from IDT ( Coralville , IA ) and contained universal tails used to anneal unique indexes for sample barcoding ( forward primers , UT1 = 5’-ACCCAACTGAATGGAGC-3’and reverse primers , UT2 = 5’-ACGCACTTGACTTGTCTTC-3’ ) [33] . Assays were grouped into one of four multiplex PCRs , with 29 assays in mix 1 , 63 assays in mix 2 , 65 assays in mix 3 , and 31 assays in mix 4 ( S3 Table ) . Each multiplex was validated in singleplex using SYBR real-time PCR with the multiplex PCR used as template [34] . The optimized multiplexes were then screened across 864 total samples ( 773 of which were analyzed here ) ( S1 Table ) . A single 10 μL multiplex PCR reaction consisted of final concentrations of the following reagents: 1x 10x PCR buffer , 1 . 5 mM MgCl2 , 0 . 2 mM dNTPs , 0 . 4 μM of each primer , 1 . 5 units of Platinum Taq ( Invitrogen , Grand Island , NY ) , and 1 μL of template . Multiplex PCR cycle conditions consisted of 95°C , 10 min; ( 94°C , 30 sec; 55°C , 30 sec; 72°C , 30 sec ) × 40 cycles; 72°C , 5 min; held at 10°C . Amplicon libraries were prepared using universal tails as previously described [33] . Briefly , a cleanup was performed on the multiplex PCR products using a 1:1 bead ratio to PCR product of 1x Agencourt AMPure XP beads ( Beckman Coulter , Indianapolis , IN ) with elution in 30 μL of a 10 mM Tris-HCl 0 . 05% Tween 20 solution . Indexed barcodes were then applied to each sample , providing a unique barcode to identify each sample . The Index Extension PCR was a single 25 μL PCR containing 12 . 5 μL of 2x KAPA HiFi HotStart ReadyMix ( Kapa Biosystems , Wilmington , MA ) , 1 μL 10 μM common universal tail primer , 1 μL 10 μM specific index universal tail primer , 8 . 5 μL molecular grade water , and 2 μL cleaned up PCR product . Extension PCR parameters consisted of 98°C , 2 min; ( 98°C , 30 sec; 60°C , 20 sec; 72°C , 30 sec ) × 6 cycles; 72°C , 5 min; held at 10°C . Following index addition , the PCR product was cleaned up again using a 1:1 bead ratio with Agencourt AMPure XP beads using an elution of 40 μL of a 10 mM Tris-HCl 0 . 05% Tween 20 solution . Amplicon libraries were normalized to a concentration of 25 nM using the SequalPrep Normalization Plate Kit , 96-well ( Thermo Fisher Scientific , Waltham , MA ) according to manufacturer’s instructions . Following normalization , the amplicon libraries were pooled in sets of 96 samples using 5 μL from each of the multiplexes , resulting in nine pools of 96 uniquely barcoded samples across 188 targets . Final sample pools were generated by pooling 100 μL from each of three of the nine plate pools into a single tube , for three final pools of 288 samples each . The three final pools were concentrated by conducting another cleanup using a 1:1 bead ratio with Agencourt AMPure XP beads and an elution of 30 μL of a 10 mM Tris-HCl 0 . 05% Tween 20 solution . These final , concentrated pools were then sequenced on the Illumina MiSeq platform using 2 × 300 bp version 3 sequencing chemistry ( Illumina , San Diego , CA ) . Amplicon sequences were aligned to the reference genome of strain CO92 [24] using BWA-MEM [23] and SNPs were called with the UnifiedGenotyper method in GATK [28 , 29] in conjunction with the NASP pipeline ( http://tgennorth . github . io/NASP ) [30] . The resulting SNP matrix was filtered to focus on the SNPs to be verified . A SNP phylogeny was generated for all 773 samples using data from 42 informative SNPs from the 63 screened previously identified SNPs and 170 additional informative SNPs identified here from among the 188 potential new SNP targets ( Fig 1A , S2 and S3 Tables ) . Neighbor-joining dendrograms based upon MLVA data were then constructed using MEGA6 [35] for each node containing >1 sample to further discriminate among samples . Subgroups were identified primarily based on SNPs , but also using MLVA for unresolved samples belonging to the basal k and d nodes ( Fig 1 ) . The geographic distributions of all of the identified subgroups were then mapped through time to determine temporal phylogeographic patterns using ArcGIS 10 . 2 . 1 for Desktop ( ESRI , Redlands , CA ) and geographic point data obtained from GeoPostcodes ( http://www . geopostcodes . com/ ) for the fokontany and communes represented in the dataset ( Figs 2 , 3 and 4 ) . Additional maps illustrating the geographic distributions of all of the identified SNP determined nodes were also generated ( Figs 5 and 6 ) . To determine if evolution of Y . pestis in Madagascar is operating under a molecular clock , we reconstructed a neighbor-joining phylogeny based upon the SNPs identified among the 31 strains sequenced here ( S1 Table ) and the previously published genomes for strains IP275 [GenBank: AAOS00000000] and CO92 [24] . We then uploaded the newick file , with associated dates of isolation , into TempEst [36] , enforcing the selection of the best fitting root ( CO92 ) and the correlation function . The correlation coefficient and R2 values were calculated , and we used a permutation test ( 10 , 000 permutations of distances ) in R to determine if the observed correlation coefficient was better than would be expected by chance . To estimate divergence times for subgroups of Y . pestis in Madagascar , we employed a Bayesian molecular clock method as implemented in the BEAST v1 . 8 . 0 software package [37] . Model selection analyses were carried out in MEGA 7 . 0 . 9 for the 33 included genomes , where the corrected Akaikes’s Information Criterion [38 , 39] results were used to determine the best fitting models . The GTR model was found to be best fitting for the dataset . Because only variable sites were included in this analysis , we corrected for the invariant sites by specifying a Constant Patterns model in the Patterns List of the BEAST xml file ( A’s: 1 , 219 , 459 , C’s: 1 , 102 , 556 , T’s: 1 , 217 , 289 , and G’s: 1 , 114 , 076 ) , and then also performed an uncorrected analysis for comparison . To determine the best fitting molecular clock and demographic model combinations for this dataset , path sampling [40] and stepping stone [41] sampling marginal likelihood estimators were employed [42 , 43] . Model comparison analyses indicated that the combination of the uncorrelated lognormal molecular clock ( UCLN ) [44] and the Bayesian Skyride [45] models best fit the SNP dataset; however , timing estimates for both the TMRCA-All and TMRCA-Madagascar/Group I were starkly bimodal . We instead selected the more conservative UCLN-Constant model , which performed slightly worse in the model comparison , but incorporated fewer parameters . In addition , and without the incorporation of interior calibrations , usage of the UCLN-Constant model resulted in estimates of both the TMRCA-All and TMRCA-Madagascar/Group I clades that were historically supported ( Table 1 ) . For each dataset , four independent Markov chain Monte Carlo ( MCMC ) chains were run for 100 million generations each , with parameters and trees drawn from the posterior every 10 , 000th step . Visual trace inspection and calculation of effective sample sizes was conducted using Tracer [46] , confirming MCMC mixing within and among each of four replicate chains . LogCombiner [37] was used to merge the samples from each chain . The first 10% of each chain was discarded as burn-in , and then each chain was resampled every 40 , 000th step . FigTree [47] was used to visualize the resulting phylogenies .
There is considerable phylogenetic diversity among Y . pestis strains from Madagascar . The 170 new informative SNPs identified here ( S2 Table ) considerably expanded on previously published SNP phylogenies of Y . pestis in Madagascar [4 , 18 , 19] . The previously identified Groups I and II were still readily apparent , but with several additional lineages within these groups . Group I included the basal k node and seven lineages , five of which ( j , l , q , r , and s ) were previously described [18 , 19] and two of which ( y and z ) were identified here ( Fig 1A ) . Group II included the basal d node and six lineages , one of which ( h ) was previously described [19] and five of which ( t–x ) were identified here ( Fig 1A ) . In addition to the seven novel lineages , additional resolution within the previously described h , j , q , and s lineages [18 , 19] was also identified ( Fig 1A ) . In all , 100 individual nodes were identified , providing considerable SNP resolution among the 773 Malagasy Y . pestis samples ( Fig 1A , Table 2 ) . MLVA provided additional resolution within the SNP determined nodes , with a resolving power range of 41%– 100% ( average of 82% ) for nodes with >1 sample ( Table 2 ) . Within the basal d and k nodes , MLVA identified one ( II . D . 1 ) and four ( I . C , I . E , I . G , and I . L ) additional phylogenetic subgroups , respectively ( Fig 1B ) . These subgroups mostly corresponded to previously identified MLVA subgroups [17 , 19] from which no strains have yet been sequenced , preventing further lineage identification using SNPs . In all , we identified 18 major subgroups among the 773 analyzed samples , including 13 SNP lineages ( hereafter referred to as subgroups h , j , l , and q–z ) and 5 MLVA subgroups ( hereafter referred to as subgroups I . C , I . E , I . G , I . L , and II . D . 1 ) ( Fig 1 ) . As previously observed [18 , 19] , there was considerable congruence between the SNP and MLVA analyses . First , the previously identified congruence between lineages h , j , l , q , r , and s and the previously described MLVA subgroups II . B , I . J , I . H , I . B , I . F , and I . A , respectively , was still apparent [18 , 19] ( S2 Fig ) . Second , novel lineages t , v , and y corresponded with the previously identified MLVA subgroups II . A , II . C , and I . D , respectively [19] ( S2 Fig ) . Third , novel lineages u , x , and z were first tentatively identified as new MLVA subgroups in the initial MLVA analysis ( S2 Fig ) used to select sequencing candidates and then confirmed using the new genome sequences and SNP analyses . Finally , the new lineage w corresponds to a subdivision within the previously identified MLVA subgroup II . D [19] , with the other samples belonging to this MLVA subgroup remaining in the basal d node within identified MLVA subgroup II . D . 1 ( Fig 1 , S1 Table , S2 Fig ) . A far more robust phylogeographic analysis was possible in the heavily sampled Betafo region ( districts indicated with yellow shading and parts of districts indicated with yellow striped shading in S1 Fig ) . Several geographically distinct subgroups were previously identified in this region [17 , 19] . These subgroups , identified here as subgroups h , j , t , v , w , and II . D . 1 , were also observed in this analysis , and showed persistence in the same geographic areas over most of the 18 year study period ( Figs 2 , 3 , 4 , 5A and 6 ) . Specific geographic distributions of the individual subgroups were consistent with previous observations [17 , 19] but much expanded . In summary , subgroup j dominated in district Mandoto with some overlap with subgroup v , which dominated in the neighboring Betafo district ( Figs 2 , 3 , 4 , 5A and 6A ) . Subgroup h was most prominent southeast of subgroup v , being found mostly in the southwestern part of district Antsirabe II , but also in Antsirabe I , southeastern Betafo , and northeastern Ambatofinandrahana ( Figs 2 , 3 , 4 and 6B ) . Subgroup w was also found in district Antsirabe II and northeastern Ambatofinandrahana , but , in general , was further east than subgroup h ( Figs 2 , 3 , 4 and 6D ) . Subgroup t was mostly found further south , in districts Manandriana , Ambositra , Fandriana , and eastern Ambatofinandrahana , but also occurred in the southern part of district Antsirabe II , mostly in between subgroups h and w ( Figs 2 , 3 , 4 and 6C ) . Subgroup II . D . 1 was on the periphery of the Betafo region in eastern Ambatofinandrahana , and was , in general , further south than subgroup h and further west than subgroup t ( Figs 2 , 3 and 4 ) . Geographic patterns among the individual SNP defined nodes within subgroups h , j , t , and v were less distinct . Similar to other subgroups , the overall geographic distributions of the individual nodes within each of these subgroups were consistent with the SNP phylogeny , with phylogenetically close nodes clustered near to each other spatially ( Figs 1A , 5A , 6A , 6B and 6C ) . However , compared to other subgroups , there was far more overlap and fewer distinct geographic patterns among the individual nodes , particularly for subgroups h , j , and v . All three of these subgroups contained several nodes that were dispersed across the overall geographic distribution of their respective subgroups , including h03 , h05 , and h11 for subgroup h , j01 –j03 and j05 for subgroup j , and v09 and v11 –v13 for subgroup v ( Figs 5A , 6A and 6B ) . Similar to the geographically dispersed nodes within subgroup s , most of these nodes were more basal within the phylogeny and/or contained larger numbers of samples and considerable MLVA diversity ( Fig 1A , Table 2 ) . The other nodes within these subgroups were more spatially restricted , but were also less geographically distinct compared to the spatially restricted nodes in other subgroups ( Figs 5 and 6 ) . The individual nodes identified within subgroup t were more geographically distinct , with t02 located in southwestern Antsirabe II , t03 –t05 found mostly in northern Ambositra and western Fandriana , t06 –t08 found further south in eastern Ambatofinandrahana , northern Manandriana , and western Ambositra , and t09 –t11 found predominantly even further south in southeastern Ambatofinandrahana , southern Manandriana , and southwestern Ambositra . Basal node t01 was more geographically dispersed , but was concentrated in the northernmost part of the geographic range of subgroup t in southwestern Antsirabe II , along with t02 ( Fig 6C ) . Samples from the Moramanga region ( districts indicated with purple shading and parts of districts indicated with purple striped shading in S1 Fig ) , the other heavily sampled region in this analysis , showed much less phylogenetic diversity than samples from the Betafo region . As previously observed [17 , 19] , the Moramanga region was dominated by subgroup q , with very few samples from this region belonging to any other subgroup ( Figs 2 , 3 and 4 ) . In contrast to previous analyses , this analysis also revealed substantial phylogeographic structure among the individual SNP determined nodes within this subgroup , with q06 found predominantly in district Manjakandriana , the southern tip of Anjozorobe , and a small area in western Moramanga , q05 and q08 –q09 found mostly further north in district Anjozorobe , q10 –q11 found predominately in eastern Moramanga , a single q14 sample found in southern Moramanga , and q12 found in the northernmost part of the geographic distribution of subgroup q , in districts Andilamena , Tsaratanana , and northeastern Anjozorobe . Node q04 , a more basal node in the q subgroup , was the least geographically defined , with representatives identified in between the distributions of q06 and q10 in district Moramanga , and also in a small area in eastern Anjozorobe in the midst of some q05 , q08 , and q12 representatives ( Fig 5C ) . Previous analyses were consistent with these phylogeographic patterns , but far less defined due to the much more limited sample sizes and lower phylogenetic resolution in those analyses [17 , 19] . Similar to the other subgroups , subgroup q also displayed temporal persistence within its geographic range in the northeastern central highlands over the 18 year study period ( Figs 2 , 3 and 4 ) . Although there was an overall pattern of temporal persistence in the same geographic areas over the 18 year study period for most of the identified subgroups , many subgroups varied in the frequency of samples identified from year to year ( Figs 2 , 3 and 4 , S1 Table ) . Although the extent of this variation could not be determined for many of the subgroups due to uneven sampling in some geographic areas , the comprehensive sampling of the Betafo and Moramanga regions allowed for a closer examination of this variation for the subgroups found predominantly in these regions . These subgroups ( h , j , q , t , v , and w ) varied in the number of samples identified per year for each subgroup , with high and low sampling years observed for each subgroup . Moreover , this variation did not follow the same pattern among all of the subgroups . For example , the number of subgroup v samples identified in 1999 and 2001 were relatively high compared to the intervening year , in 2000 . In contrast , subgroups h and q experienced peaks in sample identification in 2000 and lower frequencies of sample identification in the bracketing years of 1999 and 2001 ( Fig 7 ) . In contrast to the overall pattern of geographic and temporal stability of the 18 identified subgroups , we observed some subgroup representatives in geographic areas outside their primary geographic range . Most interesting were several examples of more long distance dispersal events , such as five samples of subgroup q ( found primarily in the northeastern central highlands , Fig 5C ) that were isolated in and around the Betafo region in 1996 ( N = 1 ) , 1997 ( N = 2 ) , and 2006 ( N = 2 ) , respectively ( Figs 2 , 3 , 4 and 5C , S1 Table ) . Interestingly , these five samples included representatives of five different SNP defined nodes within this subgroup ( Fig 5C , S1 Table ) . All of the other samples assigned to these five nodes were isolated in the northeastern central highlands , suggesting that these nodes evolved there . Together , these patterns strongly suggest that the five samples of subgroup q isolated in the Betafo region were the result of independent dispersal events from the typical geographic range of subgroup q in the northeastern central highlands . Indeed , many of the occurrences of subgroups outside of their typical geographic ranges were likely due to independent dispersal events rather than an initial dispersal event followed by localized establishment of the transferred subgroup , as there was little to no evidence of persistence of a “transferred” subgroup in a non-typical geographic area . However , it should also be noted that this lack of evidence could have been due to inadequate sampling that failed to detect any low level persistence of these “transferred” subgroups . Regardless , any dispersal events , particularly over long distances , are likely human-mediated and either related to the accidental transport of rats and fleas along with legitimate shipments , or could also be related to humans who were infected in a location distant from where they sought medical attention . Indeed , the black rat has been shown to have a range of only 40–50 m during normal activities , with travel up to only ~350 m in pursuit of resources [48] , and so rat dispersal alone is unlikely to account for these types of observed transfers of Y . pestis genotypes . Other examples of potential dispersal events involved shorter distances and could have multiple causes . These included samples of subgroup j occasionally being isolated within the typical geographic range of neighboring subgroup v and vice versa , as well as other similar crossover type events that occurred among the other densely packed and phylogenetically diverse subgroups found in the highly active Betafo region ( Figs 2 , 3 and 4 ) . These examples may represent dispersal events over shorter distances that could have been rat- or human-mediated . Alternatively , these examples might not reflect dispersal events at all but could , instead , indicate subgroups that are actually established in more than one geographic area but are very rare in the area in which the “dispersal event” appears to have occurred , at least as represented by human derived samples . To determine whether or not Y . pestis in Madagascar is evolving according to a molecular clock , we performed a linear regression and found that 28% ( R2 = 0 . 28 ) of the variation in root-to-tip distances could be explained by time ( S3B Fig ) . Permuting the distances 10 , 000 times revealed that the observed correlation coefficient , R = 0 . 527 , was better than 99% of all randomly generated correlation coefficients ( S3C Fig ) . These analyses revealed that although strict molecular clock methods were not appropriate due to the limited , albeit statistically significant , variation of distance explained by time , relaxed molecular clocks were well-suited for divergence time estimation . Bayesian estimation of divergence times using the ascertainment bias correction with a relaxed molecular clock and constant population size model , revealed that CO92 and the Malagasy strains diverged from their MRCA in 1880 ( mean date; Table 1 , Fig 8 ) , which was consistent with a split following the onset of the third pandemic in 1855 in Yünnan , China [2] , and prior to the introduction of Y . pestis to Madagascar in 1898 [12] . The estimated mean TMRCA for the Malagasy strains , and therefore for the basal Group I , was 1927 ( Table 1 , Fig 8 ) , just 29 years after the introduction of plague to Madagascar . Importantly , the confidence intervals for this TMRCA ( 95% CI: 1864–1987 , Table 1 ) did not stretch back in time beyond the third pandemic , indicating that our divergence time analysis , using only tip calibrations , was supported by historical events . The mean divergence of the Group II Malagasy strains from Group I was estimated to have occurred in 1951 ( Table 1 , Fig 8 ) . Indeed , the estimated mean divergence times for most of the subgroups identified in this analysis were after 1950 , which was when several successful plague control methods were implemented in Madagascar that led to a large decrease in the numbers of human cases [9] . Likewise , the steady increase in human plague cases that began in the 1980s in Madagascar [9] was consistent with the estimated mean divergence times for the different lineages within those subgroups with multiple whole genome sequence representatives , which ranged from 1976 to 1990 ( Fig 8 ) . A final observation from this analysis was that although the mean and median divergence times for the deeper CO92 and Malagasy strains were impacted by the SNP ascertainment bias correction , the confidence intervals for all divergence time estimates , and also the mean and median divergence times for the more recent Group I and Group II divergences , were nearly identical ( Table 1 ) .
Plague continues to be a significant public health concern in Madagascar , with hundreds of human cases reported annually [9 , 49] . Human cases exhibit strong seasonality as well as spatial and temporal variation in the affected fokontany . Most human cases occur from October to April during the warm rainy season [8 , 50 , 51] and different fokontany are affected in different years , with some fokontany unaffected despite nearby fokontany having cases [48] . The seasonality of plague is linked to population dynamics of the black rat and its flea vectors , with onset of the plague season in October coinciding with the minimum abundance of rats and maximum abundance of fleas [8 , 10 , 11] . The basis for the spatial and temporal variation in affected fokontany is less clear [48] , but could be related to similar ecological factors or stochastic forces . Our temporal phylogeographic analysis of 773 Y . pestis samples from 32 districts in Madagascar , collected over 18 years , with an emphasis on the Betafo and Moramanga regions , provides insight into this and other aspects of plague ecology in Madagascar . Previous analyses have suggested that Y . pestis in Madagascar is maintained in multiple , geographically and phylogenetically distinct subpopulations likely sustained by the black rat [17–19] . Our analysis is consistent with this hypothesis and suggests that these subpopulations are spatially and temporally stable , with the same phylogenetic types persisting in the same geographic locations over a period of almost two decades ( Figs 2 , 3 and 4 ) . This observed temporal phylogeographic pattern suggests that persistent endemic cycles of Y . pestis transmission within local areas are responsible for the long term maintenance of plague in Madagascar , rather than repeated episodes of wide scale epidemic spread . Indeed , there is little evidence for frequent , widespread selective sweeps of individual genotypes . Dispersal events do occur , but seldom appear to result in the successful establishment of a new genotype in a new location ( Figs 2 , 3 and 4 ) . The failure of a dispersal event to result in successful ecological establishment may be strongly affected by the presence or absence of an existing locally established and cycling genotype . For example , Mahajanga was likely free of Y . pestis when a long distance dispersal event from the central highlands allowed for at least the temporary successful establishment of subgroup s in this city in the early 1990s [18] . Another , related factor that likely assists in plague establishment is the presence of a high abundance of susceptible hosts [5 , 52] . This very likely played a role in the Mahajanga outbreaks , which began in an area with poor hygiene and large numbers of rats and shrews [13 , 14 , 53] . Similarly to Mahajanga , the apparent spread of subgroup s from its presumed origin in district Ambositra to Antananarivo and the surrounding areas [18 , 19] may have been facilitated by a lack of locally circulating genotypes and perhaps an abundance of susceptible hosts . Supporting this idea is the fact that for ~30 years following the successful plague control campaigns of the 1950s there were only 20 to 50 human plague cases reported per year in Madagascar , and in Antananarivo specifically no cases were reported between 1953 and 1978 [9] . Following this , the number of annual human plague cases increased steadily [9] , with notable subgroup s-linked outbreaks in Antananarivo and Mahajanga in the 1990s [9 , 18 , 19] . This suggests that there may have been an open niche in Antananarivo and the surrounding areas that subgroup s was able to occupy following a fortuitous dispersal event , similar to what happened in Mahajanga . Interestingly , the estimated mean divergence time for the various lineages within subgroup s was 1976 ( Fig 8 ) , which is consistent with this timeline . Intriguingly , the rodent population in Antananarivo at this time consisted of approximately 80% Rattus norvegicus , a host not usually thought of as highly susceptible to plague [53] . By the late 1990s , R . norvegicus made up 95% of the rodent population and both R . rattus and R . norvegicus from Antananarivo displayed high levels of plague resistance [54] . Thus , if subgroup s has only been present in Antananarivo since the late 1970s , it succeeded in becoming established and persisting in a relatively resistant host population . Previous studies have also suggested that subgroup s may possess some adaptive advantage affecting its ability to become established following a dispersal event [18 , 19] . If so , this advantage does not appear to have enabled subgroup s to further expand its geographic range [17 , 19] during the 18 year period of this analysis , despite likely dispersal events of this subgroup to other areas ( Figs 2 , 3 , 4 and 5D ) . It is possible that the presence of other locally established and cycling genotypes ( i . e . , an occupied niche ) , as documented here , inhibited the establishment of this or other transferred genotypes during the 18 year period of this analysis ( Figs 2 , 3 and 4 ) . The Betafo and Moramanga regions emphasized in this analysis exhibited distinct differences in observed phylogenetic diversity that may be related to landscape differences between these two regions . The Moramanga region consists of a wide valley along the Mangoro River that contains large and fragmented forested areas and gradually decreases in elevation from north to south . The Betafo region , in contrast , consists of a much more diverse landscape . District Mandoto in the western portion of this region consists of a plateau area with rolling hills and a fairly flat relief . To the east , district Betafo is more rugged , with relatively large changes in elevation between fokontany . District Antsirabe is even more mountainous , with fokontany located in deep valleys separated by high ridges [55] . Interestingly , this landscape heterogeneity was mirrored in the observed phylogenetic diversity for these regions . The relatively homogenous and level landscape of Moramanga was dominated by a single subgroup , q ( Figs 2 , 3 , 4 and 5C ) . Similarly , the relatively flat landscape of district Mandoto within the Betafo region was also dominated by a single subgroup , j ( Figs 2 , 3 , 4 and 5A ) . In contrast , the more heterogeneous landscape of the rest of the Betafo region , characterized by much greater variation in elevation , contained much more phylogenetic diversity , with at least four distinct subgroups ( h , t , v , and w ) identified in close proximity ( Figs 2 , 3 , 4 and 6 ) . This suggests that landscape plays a role in maintaining the multiple geographically and phylogenetically distinct subpopulations of Y . pestis identified in Madagascar , likely by limiting the potential for dispersal of the black rat and its fleas . Indeed , population genetics studies of the black rat in Madagascar are consistent with this hypothesis , with rat populations from landscapes characterized by greater topographical relief showing greater genetic structure than rat populations from flatter areas [55] . There is strong evidence that Y . pestis population sizes vary through time . Population expansions and contractions related to the alternating epizootic and enzootic cycles that characterize Y . pestis are likely the basis of the highly variable molecular clock rate observed across the worldwide Y . pestis phylogeny [3] . In Madagascar , the high and low plague seasons are associated with similar Y . pestis population expansions and contractions , as indicated by higher levels of both Y . pestis seroprevalence in rats [48] and numbers of human cases [50 , 51] during the high plague season compared to the low season . A phylotemporal analysis of the Mahajanga outbreaks of the 1990s provided additional evidence , revealing a striking pattern of diversity generation and loss during and after each seasonal plague outbreak , consistent with seasonal population expansions and inter-seasonal population contractions [18] . In addition to these seasonal variations in Y . pestis population size , our analysis suggests that there is also likely variation in the magnitude of a population expansion during a given epizootic cycle in Madagascar , and that this variation is not consistent among the different Y . pestis subpopulations maintained in Madagascar . Specifically , we observed variation in sampling frequency from year to year for the subgroups found predominantly within the well-sampled Betafo and Moramanga regions . Assuming that the identified samples were representative of the underlying populations of these subgroups , this suggests that the magnitude of a seasonal Y . pestis population expansion varies from year to year . We also observed that the high and low sampling years observed for each subgroup were not consistent among subgroups . Together , these observations suggest that local , underlying ecological factors may affect the magnitude of seasonal population expansions of individual subgroups from year to year and , consequently , whether or not a subgroup was sampled in a given year . Many such potential ecological factors have been identified in Madagascar . Variation in elevation and associated temperature fluctuations are strongly associated with shifts in human plague seasonality and are thought to affect development of the flea vector and the efficiency of flea blockage by Y . pestis [8 , 51] . Rat reproductive and migration patterns are influenced by similar seasonal climatic changes and related resource availability , particularly with regards to agricultural crops [48] . Local changes in these or other factors likely affect the population dynamics of the various Y . pestis subgroups established in Madagascar , which , in turn , likely affect the likelihood of observing human cases in a particular fokontany during a particular year . Persistent endemic cycles of Y . pestis transmission within local areas of Madagascar result in strong , consistent spatial structuring that persists through time . Landscape likely influences local diversity of Y . pestis , with increased topographical relief associated with increased levels of localized differentiation , and the maintenance of multiple phylogenetically distinct subpopulations even within relatively short geographic distances . Dispersal events rarely appear to result in the establishment of a transferred genotype in a new location , possibly due to the presence of an existing locally cycling and established genotype . Local ecological factors in the geographic ranges occupied by individual Y . pestis subpopulations likely affect the dynamics of individual subpopulations , and the associated likelihood of observing human plague cases in a given year in a particular fokontany . Altogether , the ecology and epidemiology of Y . pestis in Madagascar are highly dynamic , affected by a variety of factors . | Plague exists in several geographically separate phylogenetic groups within Madagascar , but little is known about the dynamics of these groups through time . We subtyped 773 Malagasy plague samples and identified 18 phylogenetic groups that showed persistence in the same locations over a period of almost two decades . Locations with more topographical relief contained more phylogenetic groups than flatter areas and different phylogenetic groups varied considerably in the number of samples collected each year . Transfers of plague from one location to another definitely occur , but appear to seldom result in the transplanted phylogenetic group becoming successfully established in a new location . Persistent , local transmission cycles are likely responsible for the long term maintenance of plague in Madagascar , rather than repeated wide scale disease transmission events . Landscape likely plays a role in maintaining different phylogenetic groups , with increased topographical relief associated with increased numbers of phylogenetic groups . Local ecological factors likely affect the dynamics of individual subpopulations and the associated likelihood of observing human plague cases in a given year in a particular location . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"biogeography",
"taxonomy",
"ecology",
"and",
"environmental",
"sciences",
"medicine",
"and",
"health",
"sciences",
"plagues",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"population",
"dynamics",
"population",
"genetics",
"geographical",
"locations",
"micr... | 2017 | Temporal phylogeography of Yersinia pestis in Madagascar: Insights into the long-term maintenance of plague |
The physiochemical determinants of drug-target interactions in the microenvironment of the cell are complex and generally not defined by simple diffusion and intrinsic chemical reactivity . Non-specific interactions of drugs and macromolecules in cells are rarely considered formally in assessing pharmacodynamics . Here , we demonstrate that non-specific interactions lead to very slow incorporation kinetics of DNA binding drugs . We observe a rate of drug incorporation in cell nuclei three orders of magnitude slower than in vitro due to anomalous drug diffusion within cells . This slow diffusion , however , has an advantageous consequence: it leads to virtually irreversible binding of the drug to specific DNA targets in cells . We show that non-specific interactions drive slow drug diffusion manifesting as slow reaction front propagation . We study the effect of non-specific interactions in different cellular compartments by permeabilization of plasma and nuclear membranes in order to pinpoint differential compartment effects on variability in intracellular drug kinetics . These results provide the basis for a comprehensive model of the determinants of intracellular diffusion of small-molecule drugs , their target-seeking trajectories , and the consequences of these processes on the apparent kinetics of drug-target interactions .
Drug efficacy in vivo is notoriously difficult to predict owing , in part , to the complexity of the underlying biochemical processes that govern drug–target interactions in situ . Semi-empiric pharmacokinetic/pharmacodynamic ( PK/PD ) models typically describe accumulation of the drug in tissue ( s ) and , hence , do not address the question of variability in efficacy for individual cells , which is determined by the drug’s access to and interaction with its target ( s ) within a cell . Variability in drug efficacy may , therefore , be a key factor driving resistance , selection , and toxicity . Here , we investigate factors affecting drug–target interactions at the single cell level . Our model system is a monolayer cell culture that allows continuous monitoring of drug binding to its target in individual cells . While this model system is , of course , far from ideal , provided that the free drug concentration in a given tissue is fairly uniform , cell culture experiments can meaningfully address the question of heterogeneity of response in a cell population . We monitor the kinetics of 2’-[4-ethoxyphenyl]-5-[4-methyl-1-piperazinyl]-2 , 5’-bi-1H-benzimidazole trihydrochloride trihydrate ( Hoechst 33342 dye ) incorporation in individual cell nuclei by measuring the dye’s fluorescence signal intensity . Hoechst dye becomes significantly more fluorescent upon binding to the minor groove of DNA and , therefore , fluorescence intensity corresponds to the amount of bound target ( DNA ) in the nucleus . Fluorescence microscopy permits resolution of both the temporal and spatial dependence of dye incorporation . It is instructive to investigate the incorporation process on two different spatial scales . By integrating out spatial degrees of freedom , we can assess overall dye incorporation kinetics with measurement of fluorescence intensity over time , Itot ( t ) , for individual cells . At a sub-nuclear scale , we can analyze the time dependence of individual pixel intensities , I ( x → , t ) , that typically correspond to a spatial resolution two orders of magnitude smaller than the whole nucleus in our system . Individual pixel intensities are noisy , for which reason we developed a method based on moments of distribution to characterize drug diffusion and signal ‘homogenization’ within the nucleus . We introduce a physical multi-compartment model of drug diffusion and binding/dissociation that can explain our experimental findings . Within this model , we also incorporate the effects of membrane permeability and partitioning ( as recently addressed [1] ) . We further extend this reaction scheme to include diffusion [2–4] and account for non-specific interactions ( high capacity , low affinity ) between drug and macromolecules other than intended targets . Non-specific interactions are often driven by chemical reactions requiring close proximity of interacting species . In a crowded intracellular environment with high local concentrations of non-specific binders , this proximity can be achieved . We , therefore , incorporated non-specific binding and dissociation processes into our reaction-diffusion model . With this detailed model , we show computationally that owing to their spatial localization in the intracellular environment , non-specific binders act as a trap , reducing extracellular drug concentration and retarding specific drug-target kinetics . The implications of these findings for drug-target interactions and pharmacological efficacy are discussed .
We used the MFC10A cell line with the NLS-Venus ( nuclear ) reporter for microscopy . Human epitheloid cervical carcinoma cells ( HeLa cell line ) were used for spectrofluorimetric measurements . We used a spectroflurorimetric plate reader ( SpectraMax Gemini ) to monitor binding kinetics on a cell population-average level . To this end , HeLa cells were fixed with 4% formalin and resuspended in Dulbecco’s phosphate-buffered saline ( dPBS , Sigma-D5652 ) at various cell densities . Next , cells were incubated with Hoechst 33342 dye ( Invitrogen-H1399 ) , and fluorescence changes over time were monitored using the microplate reader ( excitation 350 nm , emission 461 nm ) . In order to measure free dye concentration in solution , cells were centrifuged and the collected supernatant was incubated with calf thymus DNA . Using a DNA standard ( calf thymus , Sigma-D1501 ) , we estimated the free dye concentration in the supernatant as a function of concentration and time . Fluorescent images were taken with the Operetta High Content Imaging System ( Perkin Elmer ) . The 20x objective was used throughout the experiments unless otherwise noted . Image processing and analysis were performed using customized imageJ and Matlab scripts ( S1 Text ) . In brief , the cherry-NLS signals were binarized and segmented in order to generate nuclear masks , which were applied to the Hoechst channel to obtain pixel intensities . Single-cell tracking for time-lapse experiments was archived with Python/Perl/Matlab scripts . Doxorubicin efficacy at the single cell level can be measured in terms of DNA damage biomarker ( s ) , such as histone γ-H2Ax . In order to combine kinetic measurements in live cells with antibody staining for the γ-H2Ax marker , we performed immunofluorescence microscopy experiments as follows: Live cells were incubated with both Hoechst dye and doxorubicin at different concentrations and imaged for relatively short periods ( typically three hours ) that were sufficient to detect dynamic patterns in fluorescence staining . Immediately thereafter , cells were fixed with paraformaldehyde , stained with an anti-γ-H2Ax antibody , and again imaged ( see Movie C in S1 File ) . This protocol allowed us to combine both the dynamic measurement of dye incorporation and the resulting phenotype ( extent of DNA damage ) for individual cells . Fluorescence images were analyzed using custom-designed in-house programs . Briefly , the image background was subtracted using ImageJ; and nuclear segmentation , tracking , and data analysis were performed using custom MATLAB code . Wolfram Mathematica was used to simulate reaction–diffusion model ( s ) . Since MFC10A cells are fairly symmetric and ellipsoidal in shape , we can identify principal axes and positions of the ‘center of mass’ using the nuclear localization sequence marker ( NLS-mCherry ) as a reference ( N . B . , NLS fluorescence intensity is stable and unchanging over the time course of these experiments ) . We introduced the distance r of any given pixel from the center of mass in the xy plane . The corresponding time dependent pixel intensity is I r ( θ , t ) = I ( x → , t ) and depends , of course , on the orientation θ of the pixel , as well . If the target ( DNA ) distribution were symmetric in the nucleus and the shape of the nucleus were spherical , one would expect that all pixels positioned the same distance r away from the center of the nucleus would have identical dye incorporation kinetics . Similarly , for a symmetric nuclear ellipse , pixels in the xy plane satisfy the condition: x 2 a 2 + y 2 b 2 = c o n s t = r 2 ( 1 ) and would be expected to have identical intensities at any given time ( here a , b are principal axes of the nucleus ) . In reality , owing to a non-homogeneous target distribution and other factors affecting dye mobility and dye transport , pixel intensities are not identical and are noisy . Averaging over all pixels that satisfy the geometric condition of Eq ( 1 ) yields a much more robust time-dependent observable variable Ir ( t ) = 〈Ir ( θ , t ) 〉 where averaging is performed over orientation angle θ . We note that the actual measured quantities correspond to the integrated intensity in the z-dimension within the depth of the confocal plane . We take this fact into account while matching experimental and theoretical results ( see S1 Text for more details ) . Finally , we defined moments of the pixel intensity distribution as follows: M n ( t ) = ∑ j I ( j , t ) r j n ∑ j I ( j , t ) ( 2 ) where I ( j , t ) and rj are , respectively , time-dependent fluorescence intensity and distance from the center of mass for pixel j . This representation of the front is robust and can be defined for any nuclear geometry . This method is often used in image processing and usually referred to as the image moment method . The main advantage of this method in our case is its invariance with respect to translation , scale , and rotation [5 , 6] due to movements of the cell and microscope stage .
Time traces of overall dye intensity ( incorporation ) , Itot ( t ) , for a typical experiment in live cells are depicted in Fig 1a . There are two striking features of these traces: ( i ) the characteristic time scale of drug incorporation kinetics , and ( ii ) the broad population distribution in individual cell kinetics . The dynamics of Hoechst dye incorporation for a typical cell ( population average ) is depicted in Fig 1b for various dye concentrations . The time scale of 103 sec for micromolar dye concentrations is rather unexpected based on first principles , which we next address . The simplest way to describe dye incorporation is to assume that the kinetics is driven by second order binding and first order dissociation reactions: d d t v ( t ) = - k ˜ on u ( t ) v ( t ) + k ˜ off [ c - v ( t ) ] ( 3 ) where v and u are free target and drug concentrations , respectively , and c is the concentration of available binding sites ( capacity ) . The parameters k ˜ on and k ˜ off correspond to effective association and dissociation rates , respectively . These parameters depend not only on the intrinsic reaction rates , but also on the spatial disposition of the target molecules , potential competing binding targets , obstructive barriers to free diffusion , cell membrane properties , and active transport processes in the cell . It is a straightforward exercise to demonstrate that experimentally observed values of k ˜ on and k ˜ off are very different from the corresponding intrinsic values kon and koff . Let us assume that the extracellular dye concentration is constant over time , u ( t ) = u0 ( we will see below that this is not always the case ) . Under this condition , one readily derives from Eq ( 3 ) the following equation: v ( t ) = v s t + ( c - v s t ) e - β t ( 4 ) β = k ˜ on u 0 + k ˜ off = k ˜ off ( 1 + u 0 / K d ) ( 5 ) v s t = k ˜ off β c ( 6 ) with the steady-state dye concentration vst , dissociation constant Kd , and relaxation rate β . The intrinsic dissociation rate and dissociation constant for dye-DNA complexes in vitro ( in cell free systems ) have been measured by several groups [7 , 8]: k off > 10 - 1 s e c - 1 ( 7 ) K d < 10 - 8 M ( 8 ) Based on these intrinsic parameters , one would , therefore , expect a relaxation rate β faster than 10−1 sec−1 for any dye concentration u0 . For a dye concentration in the micromolar range , u0 ∼ 1 μM , the relaxation rate is dominated by the binding reaction and would be expected to be 10 sec−1 . Experimentally , however , we observed a much slower relaxation rate , of the order of 10−3 sec−1 ( Fig 1a and 1b ) . We note that replacing the intrinsic association rate kon with a conventional diffusion-driven association rate does not explain the slowness of the observed relaxation rate . First , the exponent β is a sum of two terms [see Eq ( 5 ) ] . Second , a typical value for a diffusion-driven association rate for a small molecule the size of the dye interacting with DNA ( in water ) is 109 M−1 sec−1 , an order of magnitude faster than the intrinsic observed association rate , kon . In order to eliminate factors related to evolving cell phenotype in culture ( i . e . , cell fate ) , we also fixed cells with paraformaldehyde and measured fluorescence over an extended period of time . Of note , we observed no significant effect of fixation on the dynamics of the population average by comparing the fluorescence of live and fixed cells for time periods of less than 3 hours . The time traces of dye incorporation are shown in Figures Aa and Ab in S1 Text for dye concentrations of 8 μg/ml . Here , we used digitonin ( Fig . Aa ) selectively or in combination with Triton X-100 ( Fig . Ab ) to permeabilize either the plasma membrane alone or all cell membranes , respectively [9] . The results ( Fig . Aa ) show that mild digitonin treatment at moderate dye concentrations does not affect incorporation kinetics . Digitonin at high concentration ( 50 ug/ml or Triton X-100 ( 0 . 1% ) treatment ) , however , has a major impact on incorporation kinetics compared to the presence of an intact nuclear membrane ( Fig . Ab ) . We observed acceleration in the initial phase of the incorporation rate by 2 . 5 − 3 . 5 -fold with a high concentration of digitonin or with Triton X-100 treatment of fixed cells . Nevertheless , even under these conditions , the incorporation kinetics is very slow compared to in vitro behavior . Since it has been reported [9] that even 5 μg/ml digitonin is sufficient to permeabilize the plasma membrane in HeLa cells , we hypothesized that the reason for accelerated kinetics in the presence of higher concentrations of detergents might not only be a consequence of dissolution of limiting membrane structures , but also dissolution of other membrane structures in the cell under these conditions . We next assessed the effective dissociation rate of dye from cellular DNA by means of ‘cold chase’ experiments . After overnight incubation with dye , cells were centrifuged and the supernatant containing unbound dye aspirated and replaced with dPBS , after which fluorescence intensity was monitored over time . The resulting decay in fluorescence is depicted in Figure B in S1 Text . Here we compare the fluorescence intensity of cells that were chased with dye-free PBS ( dPBS ) ( Fig . Ba ) to cells that were maintained in dye-containing solution ( Fig . Bb ) . Note that fluorescence decay was essentially unaffected by the presence of free dye in solution . The slight and near equivalent fluorescence decay in both conditions is most likely due to dye degradation at room temperature and not dissociation from DNA . We observed that effectively irreversible tight binding of dye , resulting in fluorescence , occurs only in intact nuclei ( Figure C in S1 Text ) . Here , lysed cells were incubated with dye , and after achieving steady-state fluorescence , chased with dye-free buffer as described above . Unlike intact cells , the fluorescence intensity of the cell lysate decreases instantaneously ( on the time scale of our typical experiments ) after the chase and quickly equilibrates at a new steady-state level . In order to tease out factors contributing to the slow kinetics of dye incorporation , we studied the spatial distribution of bound dye as a function of time . Surprisingly , we observed a reaction front propagation in live cells that lasted several minutes ( cf . Fig 2a , and Movies ( A , B ) in S1 File ) . The dependence of Ir ( t ) as a function of time is shown in Figure D in S1 Text for a typical spheroidal nucleus with principal axes of the nucleus a ≈ b . It is clear from the results of Figure D that dye incorporation dynamics is non-uniform ( at least during the initial several minutes of monitoring ) . The observed front is a result of faster incorporation of the dye at the periphery of the nucleus compared with the center . While this is rather expected behavior , what is surprising , once again , is the kinetics of front propagation . Free dye diffusion in water is characterized by an estimated diffusion constant of 500 μM2 sec−1 [10] and , hence , the expected homogenization time in a nucleus of radius 20 μM is 1 sec , two to three orders of magnitude faster than what we observed experimentally . Note that the results in Figure D suggest that after an initial period of homogenization ( i . e . , completed front propagation ) , the kinetics becomes uniform across the entire nucleus . To make this observation more apparent , we compared fluorescence intensities of the whole nucleus and sub-regions of the nucleus at different time points in Figure E in S1 Text ( see S1 Text for the computational details ) . Here , a sub-region corresponds to 10% of all pixels situated around the geometric center of each individual nucleus ( sub-regions were defined by “shrinking” the nucleus’s shape in each dimension proportionately and , hence , preserving nuclear geometry ) . Comparison of sub-regional to total fluorescence intensity , indeed , demonstrates slow reaction front propagation dynamics that varies among cells . However , quantification of the dynamics based on this representation relies heavily on a uniform distribution of target density and symmetry of the nuclei . A more direct and rigorous way to quantify and characterize front propagation is to calculate moments of the pixel intensity distribution Mn , a parameter that is not dependent on symmetries in geometry and target distribution . Typical time traces of the second moment M2 are depicted in Fig 2b for individual nuclei ( [dye] = 2 μM ) and for the population average ( Fig 2c , different dye concentrations ) . Front propagation initially drives a large second moment ( only a thin shell of the nucleus incorporates dye ) towards a steady-state that depends on the DNA distribution . While the typical time scale of homogenization is significantly faster than the relaxation time for overall dye intensity , it is still much slower than the 1 sec time scale discussed above . Note that time traces of M2 depicted in Fig 2b display variability in both relaxation kinetics and the steady–state achieved , similar to total dye incorporation Itot . Furthermore , we observed excellent correlation between the relaxation rates of M2 and Itot time traces ( cf . Fig . F in S1 Text ) . The observed pattern of front propagation and incorporation suggests that the slow kinetics is driven by slow mixing of the dye in the nucleus . To confirm this hypothesis , we introduce a reaction–diffusion model that takes into account the interaction between dye and DNA , and diffusion of free dye . Assuming that DNA binding sites are largely stationary compared to dye molecules , we derive: R ( u , v ) = k on u v - k off ( c - v ) ( 9 ) ∂ t u ( x , t ) = D ∇ x 2 u - R ( u , v ) ( 10 ) ∂ t v ( x , t ) = - R ( u , v ) ( 11 ) Here , R is a local reaction rate and D corresponds to the diffusion coefficient of free dye in the nucleus . Note that the model implicit in Eqs ( 9 ) – ( 11 ) corresponds to a mean field description and , therefore , is not suitable for the study of variability in incorporation dynamics across the nucleus . Eqs ( 9 ) – ( 11 ) also need to be supplemented by the appropriate boundary condition: D ∇ x u ( x , t ) = h m [ u e x t - u ( x , t ) ] , x ∈ Ω ( 12 ) where Ω corresponds to the position of the nuclear membrane , uext is the external dye concentration , and hm is an effective mass transfer coefficient through the boundary Ω . Unlike other parameters that appear in Eqs ( 9 ) – ( 11 ) , the value of hm is difficult to estimate since it depends on multiple electrostatic and other chemical properties of the cytosol and cell membranes , such as macromolecular obstructions to diffusion , partition coefficient , dielectric properties , and specific transporter kinetics . Instead , we can attempt to determine the value of the coefficient hm by fitting experimental data to Eqs ( 9 ) – ( 12 ) . Note that under the assumption that Eqs ( 9 ) – ( 12 ) correctly describe dye incorporation kinetics , the variability among individual cells is driven by the effective mass transfer coefficient hm and nuclear radius R . Indeed , all cells are exposed to an identical dye concentration in cell culture ( even if that concentration is itself time-dependent ) , and all cells ( in the same cell cycle phase ) have a similar number of available binding sites . Upon further consideration , one realizes that the model described by Eqs ( 9 ) – ( 11 ) is inadequate . Dynamics and steady-state prediction based on Eqs ( 9 ) – ( 11 ) cannot adequately explain the experimental data ( cf . S1 Text ) . Briefly , in the steady-state , the free extracellular dye concentration will be the same as the intracellular concentration , uext = ust . Therefore , the bound dye concentration in the steady-state is completely insensitive to uext in the range of concentrations higher than Kd ∼ 0 . 01 μM; however , we observed a sensitivity to dye concentration in cell culture in the concentration range of 0 . 1 μM − 10 μM . We note that with the introduction of continuous extracellular dye depletion through the boundary condition , Eq ( 12 ) does not remedy the inadequacy of the model described by Eqs ( 9 ) – ( 11 ) ( cf . S1 Text for details ) . Other model modifications are , therefore , required to account for the observed experimental data . A local dye concentration in excess of the Kd is a principal reason for the failure of the simple passive diffusion model of Eqs ( 9 ) – ( 11 ) to recapitulate the observed experimental data . The introduction of a barrier ( such as a limiting membrane compartment ) results in slower kinetics , as we have seen for small values of the Biot number ( dimensionless transfer coefficient , B i = h m R n D ) ( cf . S1 Text ) , but by itself does not lead to a reduction in the local free dye concentration at later time points . This reduction in free dye concentration could be achieved by active transport of the dye molecules through the cell membrane boundary; however , we observed that cell fixation with formaldehyde does not qualitatively change the fluorescence kinetics . Thus , we turned to other possible explanations for the experimental observations , chief among which is non-specific binding leading to apparent anomalous diffusion . Another possible explanation for the reduction in free dye concentration is ‘buffering’ by non-specific ( i . e . , weaker ) binding to other macromolecules in the cytoplasm and nucleus . One obvious suspect in this regard is DNA itself , since dye binding to different base pair sequences occurs and results in much lower or undetectable fluorescence . If such nonspecific binding ( low affinity , high capacity ) is a correct explanation for the significant reduction in free dye concentration inside the nucleus , one would expect much higher uptake of the dye by the cells during the course of the experiment than expected from specific binding alone . Indeed , if the dye binds only to the specific high affinity sites that constitute a small fraction ( ∼ 1% ) of total DNA , the effect of dye binding to these specific sites on total dye concentration is expected to be small . In our experimental setting , the number of cells per well is ∼ 5 ⋅ 104 , and , therefore , the number of total base pairs per well bptot is ∼ 1 . 5 ⋅ 1014 . This number can serve as the basis for a rough estimate of the number of non-specific binding sites . For a typical dye concentration of 1 μM in a cell culture well of 150 μl volume , the number of available dye molecules dyetot is ∼ 9 ⋅ 1013 , which is comparable to bptot . Provided that only ∼ 1% of total DNA binds dye molecules specifically [11] , depletion of the total dye pool should be negligible with exclusive specific binding . We , however , observed a significant depletion of dye not only at [dye] = 1 μM , but also at higher dye concentrations ( vide infra ) . This finding is consistent with lower affinity binding of high capacity . The amount of non-specific binding sites that act as a dye buffer is proportional to cell density . In order to quantitate this relationship accurately , we used suspended fixed cells , which allows one to quantitate this relationship accurately and also to monitor the remaining free dye concentration in cell culture over time . The latter measurement was obtained by cell centrifugation and subsequent analysis of dye in the cell-free supernatant . Subtracting residual dye concentration from the initial concentration , we can estimate the amount of dye taken up by the cells and compare it to the amount of DNA in the cells . In order to assess different fluorescence conditions , we incubated different combinations of dye concentration and cell density . The resulting fluorescence intensity at late time point ( 18 hours of dye incubation ) is shown in Figures Ga and Gb in S1 Text , where we compare the fluorescence intensity from intact fixed HeLa cells ( Fig . Ga ) and the extrapolated signal from the calf thymus DNA ( CT ) titration data set ( Fig . Gb ) . Namely , we extrapolated a CT signal assuming 6 pg/cell DNA concentration using CT/dye titration data shown in Figure H in S1 Text . Note that high dye concentration leads to quenching of the fluorescence signal ( see also [11] ) in CT , for which reason we restricted our analysis to dye concentrations < 8 μg/ml . The results of Figures Ga and Gb suggest there may exist dye binding molecules in addition to the specific binding sites in the minor groove of DNA ( e . g . , other DNA binding sies , RNA , and/or proteins ) that would account for higher fluorescence intensity in cells compared to cell-free CT standards . The existence of buffering molecules would also explain less tight binding that manifests in significantly more gradual titration curves for cells compared to cell-free CT samples . We estimated residual ( free ) dye concentration in cell suspension samples using the standard CT method ( N . B . , we could not measure free dye by simple light absorption owing to sensitivity limits ) . Cells were centrifuged at 8000 g , and the collected supernatant was incubated with a fixed concentration of CT ( ≈ 100 μg/ml ) . The CT standard was obtained by titrating various dye concentrations in the presence of the same concentration of CT as above ( cf . Figure I in S1 Text ) . Using this approach , one can estimate the residual free dye concentration in the cell suspensions . The results are shown in Figures Gc and Gd for the corresponding raw data ( Fig . Gc ) and extrapolated values of free dye in supernatant samples ( Fig . Gd ) . Owing to the second incubation step that is necessary in this approach , the original free dye was diluted two-fold , which was taken into account in the results in Fig . Gd . We note that due to limited sensitivity of the assay , the free dye concentration could not be accurately assessed for values < 1 μg/ml . For this reason , we did not apply extrapolation to samples with initial dye concentrations less than 4 μg/ml . For high initial dye concentrations , we observed dye uptake that cannot be explained by specific DNA binding alone . Indeed , for a cell density of 2 . 5 ⋅ 105 cells/ml , there is approximately 1 . 5 μg of DNA per ml volume in solution . The dye uptake by the cells shown in Figure Gd is at least 3 times greater than the total DNA concentration , 4 . 5 μg/ml for [dye] = 8 μg/ml . Taking into account that only a fraction of DNA is available for specific binding ( cf . CT titration data , Fig . H in S1 Text ) , there must exist ( macro ) molecules with low binding affinity and much higher concentration ( capacity ) than specific DNA sites to account for the magnitude of dye uptake we observed . Before we turn to a numerical simulation of the model that takes into account non-specific binding interactions , let us demonstrate the resulting behavior in a single cell . It is intuitively clear that any binding and dissociation reactions , whether specific or non-specific , can lead to anomalous diffusion of molecule ( s ) in the cell by impairing the theoretical unimpeded diffusion of the molecule in the cytosol . ( Anomalous diffusion has been studied in some limiting cases of these interactions under the rubrics of ‘excluded volume’ and ‘fractal structure of the cell’; for review see e . g . , [12 , 13] ) . We demonstrate anomalous diffusion behavior for a ‘toy’ system: diffusion of a single particle ( drug molecule ) in bulk . In what follows , we assume that the particle undergoes a random walk on a d-dimensional lattice and can interact with particles uniformly embedded in nodes of the lattice . A diffusing particle can be in one of two possible probabilistic states , p and q ( i . e . , free or bound , respectively ) . We introduce transition rates k+ and k− between these two states ( which are the microscopic analogues to kon and koff , respectively ) . The major advantage of the model compared to a general case is linearity and , hence , the existence of an exact solution . Indeed , the continuous version of this model yields: R 1 ( p , q ) = k + p ( x , t ) - k - q ( x , t ) ( 13 ) ∂ t p ( x , t ) = D ∇ x 2 p - R 1 ( p , q ) ( 14 ) ∂ t q ( x , t ) = R 1 ( p , q ) ( 15 ) subject to boundary and initial conditions: p ( Ω , t ) = 0 ( 16 ) p ( x , 0 ) = δ ( x ) ( 17 ) q ( x , 0 ) = 0 ( 18 ) where the boundary Ω is assumed to be very far from the origin , x = 0 . In this setting , we wish to calculate the mean square displacement 〈x2〉 of the particle from its origin: ⟨ x 2 ⟩ = ∫ d x x 2 ( p + q ) ( 19 ) On very short time scales k+ t ≪ 1 , the diffusion is normal and is described by the usual rate law 〈x2〉 = 2 dDt where d is the lattice dimensionality . In the long time regime ( k+ + k− ) t ≫ 1 , one expects the following asymptotic behavior: ⟨ x 2 ⟩ ≃ 2 d D * t ( 20 ) D * = D k - k - + k + ( 21 ) The asymptotic behavior Eqs ( 20 ) and ( 21 ) is due to translational symmetry , namely , diffusion and reaction processes do not depend on the position of the particle on the lattice . Indeed , since the particle can move only while in a free state , the late time asymptotic diffusion rate is proportional to the steady-state probability that the particle is free at any given time . In order to derive an exact solution to the mean square displacement in the case of the toy model Eqs ( 13 ) – ( 18 ) , we first derive the relaxation dynamics of the free particle state p0 ( t ) in the case of d = 0 ( single site lattice , corresponds to D = 0 in Eqs ( 13 ) – ( 15 ) ) : p 0 ( t ) = k - k - + k + + k + k - + k + e - ( k - + k + ) t ( 22 ) The exact solution of Eqs ( 13 ) – ( 15 ) is , therefore , given by: ⟨ x 2 ⟩ = 2 d D ∫ 0 t d τ p 0 ( τ ) ( 23 ) ⟨ x 2 ⟩ = δ [ 1 - e - ( k - + k + ) t ] + 2 d D * t ( 24 ) δ = 2 d D k + ( k - + k + ) 2 ( 25 ) Here , the integral over time in Eq ( 23 ) corresponds to the total time the diffusing particle remains in the free state during observation time t . The numerical simulation of the mean square displacement for Eqs ( 13 ) – ( 15 ) in 1 d is presented in Fig 3a along with the exact solution , Eqs ( 23 ) – ( 25 ) . The mean square displacement in the presence of association and dissociation reactions for our model exhibits anomalous diffusion in the transient time regime , Eqs ( 23 ) – ( 25 ) . One expects that if interacting particles are embedded on the lattice in a non-uniform fashion , this pattern will persist much longer since under these conditions a diffusing particle will explore different regions of space , and microscopic reaction rates k+ and k− will become position-dependent . We next considered a locally non-uniform distribution of interacting particles ( which is the case for DNA binding sites ) and compared the time dependence of 〈x2〉 to the case of a uniform distribution ( with identical average k+ values in both cases ) . One may expect that after sufficient space exploration time ( late time limit ) , 〈x2〉 would exhibit similar asymptotic behavior for both uniform and non-uniform local distributions of interacting particles . In order to demonstrate this fact , we introduced a local perturbation to the binding rate , k + = k + 0 + Δ ( x ) . For periodic local perturbation , Δ ∝ sin ( ωx + θ ) , simulated mean square displacements 〈x2〉 are shown in Fig 3b . Even for uniformly distributed interacting particles , the diffusion is anomalous if more than a single particle performs a random walk on the lattice . This anomalous diffusion occurs because binding and dissociation rates become time-dependent . Indeed , for a given walker , the state of interacting particles at any site on the lattice depends on whether other walkers are engaged at that site . Disregarding spatial fluctuations , we can formulate a mean-field approximation for the multiple walkers problem by substituting k+ ∼ kon ρf ( t ) , where ρf ( t ) describes the time-dependent concentration of available ( i . e . , not bound ) reactive species interacting with the walker . For low dye ( walker ) concentrations , we can estimate the effective diffusion rate using Eqs ( 20 ) and ( 21 ) : D * = D k o f f k o f f + k o n ρ ( 26 ) D * = D K d K d + ρ ( 27 ) where it is assumed that the free reactive species concentration does not change significantly , ρf ( t ) ≈ ρ . For the sub-micromolar dissociation constant Kd of non-specific binding reported in [11] and high intracellular concentrations with lower affinity binding sites ρ ≥ 100 μM , one may expect a 102 − 103 times slower effective diffusion rate D* compared to the diffusion of dye in water , D . This slow effective diffusion constant is consistent with the time scale we observed in our experiments . We note here that the mechanism of retardation of dye transport through membrane ( s ) most likely is also driven by non-specific interaction between dye and lipid molecules or dye and ( membrane ) protein molecules present in high local concentration . Using Eq ( 27 ) we can approximate the time-dependent changes in effective diffusion constant in the bulk phase by assuming D e f f ( t ) = D K d K d + ρ ( t ) ( 28 ) where ρ ( t ) is a time-dependent spatial average concentration of available binding sites . This is , of course , a crude approximation that completely ignores spatial fluctuations in interacting particle density . In order to estimate the time dependence of ρ ( t ) , we ( i ) assume that all cells are identical , and ( ii ) once again ignore spatial fluctuations in the distribution of interacting particles . Under these assumptions , we derive the autonomous evolution equation for ρ ( t ) : d d t ρ = - k o n ( u 0 + ρ - ρ t o t ) ρ + k o f f ( ρ t o t - ρ ) ( 29 ) ρ ( 0 ) = ρ t o t ( 30 ) where we define ρtot as a total amount ( capacity ) of DNA and u0 is an initial amount of dye available for each cell . The solution of the nonlinear equation , Eqs ( 29 ) and ( 30 ) , is: ρ ( t ) = ρ 1 + ρ 2 tanh ( 1 2 β t + ρ 3 ) ( 31 ) β = [ k o f f + k o n u 0 + k o n ρ 0 ] 2 - 4 [ k o n u 0 ] [ k o n ρ 0 ] ( 32 ) where all parameters , ρ1 , ρ2 , ρ3 , and β , depend on reaction rates and initial conditions . In order to estimate the value of rate β , we consider a case wherein u0 ≈ ρ0 ∼ 1 μM . ( Note that here ρ0 corresponds to the average concentration of DNA in culture media , not in the individual cell ) . In this case one derives: β ∼ 2 k o f f ρ 0 / K d ( 33 ) Experimentally , we observed a very slow effective dissociation rate koff ≲ 10−5 sec−1 , ( see Fig . B in S1 Text ) . Hence , the dye depletion rate can be approximated from the above as β ≲ 10−4 sec−1 for sub-millimolar non-specific dissociation constant Kd . The derivation of the solution Eqs ( 31 ) and ( 32 ) and its generalization to the case of multiple binding species can be found in S1 Text; also note Figure Ja for a comparison of analytical and numerical solutions for this case ( cf . S1 Text section , Mean-field Solution to Autonomous Binary Reaction Model ) . We also used a numerical simulations scheme that allows us to trace a single “molecule” ( walker ) displacement during a stochastic reaction-diffusion process implemented in 3d space . The typical time traces of 〈x2〉 for mobile species in the absence and presence of interactions with stationary interacting species are shown in Figs . Jc-Je in S1 Text . A conceptual diagram of the time dependence of D ˜ is shown in Fig 3c . In order to incorporate nonspecific binding in the model defined by Eqs ( 9 ) – ( 11 ) , we introduced an additional term that corresponds to an average ( lower affinity , relatively ) non-specific binding site . We further assume that this non-specific binding site is immobile compared to free dye in the time course of the experiment: R ( u , v ) = k on u ( c - v ) - k off v ( 34 ) R n ( u , v n ) = k on n u ( c n - v n ) - k off n v n ( 35 ) ∂ t u ( x , t ) = D ∇ x 2 u - R ( u , v ) - R n ( u , v n ) ( 36 ) ∂ t v ( x , t ) = R ( u , v ) ( 37 ) ∂ t v n ( x , t ) = R n ( u , v n ) ( 38 ) Here , the superscript ns refers to a ( generic ) non-specific binding site . We include an estimate of two additional parameters in the model in Eqs ( 34 ) – ( 38 ) , k on n and k off n , from reference [7] and assume that the concentration of non-specific binding sites is ∼ 100-fold greater than specific sites , i . e . , cn ∼ 100c . The results of numerical simulation of the model described by Eqs ( 34 ) – ( 38 ) are shown in Fig 4a–4d . The corresponding experimental results are shown in Fig 5a and 5b . Numerical simulations of the full non-specific interaction model support the prediction of the qualitative estimate above that both dye incorporation and front propagation are consistent with a slow diffusion process . The front dynamics is not described by a simple exponent as expected in the case of normal diffusion but , rather , consistent with the anomalous behavior discussed above . We turn next to the heterogeneity in incorporation kinetics . To begin , we examine the steady-state levels of dye incorporation . A typical histogram of steady-state fluorescence intensity is presented in Figure L in S1 Text . This multi-modal distribution is likely to be driven by the cell cycle , with the two largest peaks corresponding to G1 and G2 phases . The coefficient of variation ( CV ) for cells in G1 and G2 states is similar and has typical values of 0 . 1 , excluding outliers ( e . g . , segmentation errors in quantifying nuclear fluorescence ) . We observed that low dye concentrations result in very slow kinetics ( cf . Fig 1b ) for dynamics of the population average . This slow kinetics is difficult to study experimentally , especially with live cells ( tracking individual cells becomes difficult with cell motion over long times ) . In order to explore a possible link between the variability in kinetics and DNA target state ( such as cell cycle phase ) , we performed a timed double ( sequential ) addition experiment , viz . , dye was added to cell culture in two sequential steps . If , at the first step of the experiment , the dye concentration is low and it is experimentally impossible to achieve the steady-state , adding high dye concentrations to the cells in the second step allows us to achieve steady-state equilibration . Even though conditions at final equilibrium are different from conditions after the first addition of dye , the DNA binding capacity can be resolved using this method . Using these sequential addition experimental data , it is straightforward to confirm the existence of the buffering molecules discussed above . If the dye were not depleted from cell culture , one would expect that adding less dye in the second phase of the experiment would result in a decrease or , in the best case , no change in final fluorescence intensity . This is not the case , as shown in Figures Ma and Mb in S1 Text . As an example , consider changes in the average fluorescence intensity for the experimental conditions [dye1] = 0 . 25 μM , [dye2] = 0 . 12 μM shown in Figure Ma ( brown curve ) . Despite adding a lower concentration of dye , the average fluorescence intensity increases . This behavior persists for higher dye concentrations . For example , the experimental conditions [dye1] = [dye2] = 1 μM also result in an increase of fluorescence intensity ( Fig . Mb , red line ) . The steady-state dependence of mean and CV of individual nuclear intensities on dye concentration are shown in Figure N in S1 Text . We classified cells into two cycle phases based on the final fluorescence intensity observed in the sequential addition experiment . While steady-state intensities display significant variation for different fluorescence conditions ( Figures Na and Nb ) , their degree of variability ( CV ) remains roughly constant for a broad range of dye concentrations ( Figures Nc and Nd ) . In contrast to the narrow distribution of dye incorporation in the steady-state , relaxation kinetics toward equilibrium exhibit a much greater variance . To demonstrate this point graphically , we introduce the normalized time-dependent variable I t o t * defined as: I t o t * = I t o t ( t ) I t o t ( T ) , ( 39 ) where T is a final dye incubation time point . Time traces of live cells’ raw intensity Itot and normalized intensity I t o t * are shown in Fig 6a and 6b , respectively . The estimated half-life of relaxation ranges from the fastest relaxation rate , τ1/2 ≈ 10 min , to the slowest relaxation rate , τ1/2 > 60 ( min ) , for [dye] = 1 μM , a 6-fold difference . It is clear from Fig 6b that the relaxation rate of incorporation correlates with the cell cycle , namely , cells in G1 phase achieve equilibrium faster than those in G2 phase . Therefore , the variability in relaxation rates is actually smaller if one takes into account cell cycle state . Even allowing for this distinction , the variability in relaxation rate is still several fold higher ( CV ∼ 0 . 6 ) than the variability in the steady-state fluorescence intensity ( CV ∼ 0 . 15 ) ( Figs . Nc , Nd , Oa and Ob in S1 Text ) . In order to determine the factors controlling the variability in dye kinetics , we performed experiments on fixed cells with permeabilized membranes ( using Triton X-100 ) . The resulting kinetics is depicted in Fig 6c and 6d . Cells permeabilized with Triton X-100 display fluorescence dynamics that is initially significantly faster ( about 3-fold on average ) than intact cells , consistent with the microplate reader data discussed above for the HeLa cell line , ( cf . Fig 5a and 5c ) . The variability in intensity of permeabilized cells appears significantly lower compared to that of intact cells . As a result , late time behavior becomes almost uniform for permeabilized cells . In addition and importantly , the data from the sequential addition experiments show that variability in kinetics among cells persists after the first addition using non-permeabilized fixed cells ( Fig 6e and 6f ) . Therefore , the factor ( s ) that cause variability do not “saturate” during the incubation phase . Since the interaction of the dye with membrane ( s ) is most likely driven by non-specific association/dissociation reactions , one would expect that saturation of binding sites would result in more uniform dynamics during the second addition phase of the experiment . This result suggests that there may exist factor ( s ) other than transport through the cell membrane that control ( s ) variability in incorporation kinetics . Some of the fixed cells’ time traces exhibit other interesting behaviors . Namely , a few traces reach peak fluorescence intensity during the incubation period after which their fluorescence intensity decreases with time . This biphasic behavior is especially apparent for low dye concentrations ( Fig 6e and 6f ) . We examined images of cells that exhibit this behavior and discovered that the nuclei of these cells have region ( s ) that incorporate dye very quickly compared to the rest of the nucleus . The very same region is responsible for a decrease in fluorescence signal after it peaks . We hypothesized that the regions with fast reaction kinetics correspond to micro-damaged areas of the nucleus ( i . e , exposed/accessible DNA binding sites ) owing to fixation . The effective diffusion and , hence , mixing , of the dye is , therefore , enhanced . Under this assumption , the peak fluorescence intensity is caused by a decrease in the extracellular dye concentration during the time course of the experiment ( due to depletion of the free dye by cells discussed above ) . This observation supports the hypothesis that the reason for accelerated kinetics in the presence of detergents might not only be a consequence of membrane dissolution , but also of the presence of other binding species and compartments within the cell . We next investigated whether anomalous and slow diffusion in cells is unique to Hoechst dye . To this end , we studied the incorporation dynamics of another DNA binding drug , doxorubicin , a potent cancer chemotherapeutic agent . In order to characterize doxorubicin incorporation , we employed an indirect method based on doxorubicin-DNA intercalation competition with Hoechst dye 33342 [14 , 15] . Since the total pool of DNA sites specific for binding to doxorubicin and Hoechst dye is limited , one may expect that dye fluorescence in cells would depend on the local concentration of doxorubicin . We , indeed , observed this antagonistic ( competitive ) effect at the single cell level . If doxorubicin is delivered at the same time or later than dye to cultured cells , we observed a peak pattern in time traces shown in Fig 7a and 7b . The peak position corresponds to the point at which doxorubicin concentration in the nucleus becomes high enough to compete effectively with bound dye for specific DNA binding sites . The timing of the peak fluorescence depends on relative dye and doxorubicin concentrations in cell culture , as can be seen in the case of high or low dye concentrations shown in Fig 7a and 7b , respectively . ( A similar pattern is observed in fixed cells , Figs . Qa and Qb in S1 Text ) . If cells are pre-treated with doxorubicin several hours prior to dye addition , however , traces exhibit simple plateau saturation ( which is [Dox]-dependent ) . This observation leads to the conclusion that it takes a fairly long period of time for doxorubicin to achieve sufficient intracellular concentrations to compete effectively with Hoechst dye . As in the dye case , this time period is [Dox]-dependent ( see timing of peaks in Fig 7a ) . Thus , slow incorporation is most likely a common feature of DNA binding drugs for exactly the same reasons as for Hoechst dye: ( i ) high local DNA concentrations , and ( ii ) non-specific interactions with other macromolecules in cells . Since these factors affect both dye and doxorubicin molecules similarly , one may expect that the kinetics of dye incorporation can be used as a proxy for doxorubicin kinetics . Surprisingly , dye homogenization in cells does not seem to be affected by co-incubation with doxorubicin . This conclusion is supported by the time traces of either moment of inertia M2 introduced above or another proxy for homogenization , the coefficient of variation in individual nuclear pixel intensities CVp . The observed dynamics of CVp is shown in Fig 7c and 7d , and unlike total intensity of incorporation ( Fig 7a and 7b ) , is largely [Dox]-independent . ( A similar pattern is seen in fixed cells , Figs . Qc and Qd in S1 Text ) . The most likely explanation for this behavior is the very similar effective diffusion properties of dye and doxorubicin , since one would otherwise expect non-uniform displacement of bound dye molecules throughout the nucleus . Doxorubicin is , of course , a clinically used chemotherapeutic agent and , hence , one can quantify drug efficacy in individual cells by assessing the time course of DNA damage after incubation . We used γ-H2Ax antibody intensity as a proxy for DNA damage in cells . To simplify phenotype characterization , we dichotomized DNA damage by introducing an assay threshold . The threshold was set based on a comparison of γ-H2Ax antibody intensity in doxorubicin-treated and untreated conditions . First , we observed that dye acts as a buffer at high dye concentration by competing for binding with doxorubicin in the DNA minor groove ( Fig 8a and 8b ) . For high dye concentration ( 16 μM ) , the extent of DNA damage is below the threshold ( corresponding to an intensity of 100 arbitrary units of γ-H2Ax antibody ) for most cells . By contrast , incubation with low dye concentration ( 0 . 5 μM ) leads to extensive DNA damage for a large fraction of cells . This result is consistent with the peak pattern for dye and doxorubicin co-incubation discussed above , which is also driven by competition for DNA binding . In addition , slow dynamics of drug incorporation leads to a higher extent of DNA damage , which is a non-trivial effect . To demonstrate this phenomenon , we plotted time traces of dye fluorescence intensity in individual cells treated with doxorubicin , as depicted in Fig 8c and 8d . Most of the cells that undergo DNA damage are in G2 phase , which is typically characterized by slower incorporation kinetics compared to cells in G1 phase; however , cells that exhibit a lesser degree of DNA damage in G2 phase typically achieve peak dye fluorescence intensity faster . The temporal position of the peak is related to the rate of intracellular doxorubicin accumulation . Hence , counterintuitively , cells are more likely to escape DNA damage if doxorubicin incorporation dynamics is rapid .
We observed several striking features of binding kinetics in our model system: First , both binding and dissociation of dye are much slower ( by three orders of magnitude ) in cells than in cell-free systems . In fact , the effective dissociation rate is so slow that binding is essentially irreversible . We show that this dye “trapping” in the nucleus is due to ( i ) high local DNA concentrations; ( ii ) higher capacity , lower affinity interactions with other macromolecules; and ( iii ) lipid membrane ( s ) partitioning and permeability characteristics . Second , we observed reaction front propagation by monitoring the spatial distribution of the dye in the nucleus over time . Temporal dynamics of front propagation is also slow compared to the dye diffusion rate in water , and is most likely controlled by the same factors as mentioned above . Third , slow drug intake/extrusion is not unique to the dye . We demonstrate that a clinically used drug ( doxorubicin ) that has a binding mechanism similar to the Hoechst dye also exhibits slow binding kinetics . Finally , we demonstrate that drug incorporation dynamics varies significantly among individual cells . On the characteristic time scales of our experiments ( minutes to hours ) , some of the heterogeneity is due to the effects of the cell membrane compartments in the cell and their kinetic effects on dye entry into the cytosol and nucleus . We observed a correlation between the dynamics of drug incorporation and its efficacy in causing DNA damage using doxorubicin as a drug and dye dynamics as a proxy for the kinetic properties of individual cells . Effectively irreversible binding has a very interesting implication in terms of distribution of incorporated drug between cells . For sub- or even micromolar drug concentrations , one expects that cells with fast incorporation kinetics would effectively serve as a sink reducing drug availability to cells with slower kinetics . This behavior might be interpreted as “passive” drug resistance in subpopulations of cells . There might be nothing biologically unique about this cell subpopulation; however , the existence of cells that can take up drug rapidly is a driving factor for the drug-resistant subpopulation . The possible clinical solution in this case might be completely counterintuitive . Instead of improving targeting of passively resistant cells , the drug-sensitive subpopulation of resistant cells needs to be treated with reagents that decrease their drug incorporation rate . A similar notion of the effective sink might be applicable on a larger spatial scale to cells in solid tumors . Some cells ( e . g . , those in outer layers ) may act as a shield , taking up the drug , which , in turn , may facilitate drug resistance of the inner layers of cells in the tumor . Non-specific interactions are often short range , driven by chemical reaction requiring close proximity of interacting species . Owing to a crowded intracellular environment , these interactions can effectively trap drug molecules in subcellular regions with high local concentrations of non-specific binders . Hence , non-specific interactions between drug and macromolecules present in the cell may result in slow and anomalous intracellular diffusion of drug molecules . Since the spatial organization of the intracellular micro-environments depends on cell cycle phase , one may expect that drug incorporation kinetics will also be cell cycle-dependent . The heterogeneity of drug incorporation is not driven exclusively by cell cycle state . We observed a high degree of variability in kinetics for both G1 and G2 subpopulations of cells . While active transport has been shown to be an important factor contributing to drug incorporation efficacy on long time scales , we have not detected significant changes in short-term kinetics between live and fixed cells at the population average level ( at least not in HeLa and MFC10A cell lines ) . Hence , other factors , such as relative spatial organization of drug targets and non-specific interacting molecules , likely drive variability in incorporation kinetics and account for anomalous diffusion characterization of the drug . Slow drug transport through the plasma membrane is often empirically taken into account during drug design and optimization stages . We observed , however , that a slow diffusion process occurs within a cell , as well , at least for cationic DNA-binding small molecules , such as Hoechst dye and doxorubicin . The immediate consequence of this slow diffusion is a dramatic mismatch between kinetic reaction rates in vivo and in vitro , which we observed experimentally . Hence , we believe that non-specific interactions have to be taken into account in order to describe drug kinetics adequately . By so doing , it is likely that different strategies will be needed to optimize drug efficacy and minimize drug resistance . | Small-molecule drug design assumes target binding of high affinity . Most small molecules can interact with other macromolecules in the cell ‘nonspecifically , ’ i . e . , with significantly lower affinity . The extent to which these nonspecific interactions influence the availability and action of the drug for its specific target depends upon the relative concentrations of drug , the specific target , and nonspecific targets . The structure of the cell is quite crowded with a highly non-uniform distribution of macromolecules that can interact with the drug of interest both specifically and nonspecifically . Thus , some compartments or micro-domains within the cell may have a comparatively high concentration of nonspecific targets , sufficient to trap the drug and retard its diffusion toward the specific target . Here , using small-molecule binding to DNA and single cell monitoring , we demonstrate that this effect results in apparently anomalous small molecule-DNA binding kinetics in cells at rates that are 1000-fold slower than in a homogeneous , dilute , aqueous environment . This slow intracellular diffusion , however , has an advantageous consequence: it leads to virtually irreversible binding of the small molecule ( drug ) to specific DNA targets in cells . We study and quantify the effect of nonspecific interactions between small DNA-binding molecules , including known DNA-binding drugs , in different cellular compartments in order to identify factors that account for the variability in binding kinetics among individual cells . | [
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] | [
"cell",
"binding",
"cell",
"physiology",
"medicine",
"and",
"health",
"sciences",
"biological",
"cultures",
"pigments",
"dna",
"damage",
"materials",
"science",
"cell",
"cultures",
"pharmacology",
"dna",
"dyes",
"cellular",
"structures",
"and",
"organelles",
"chemical... | 2018 | Determinants of drug-target interactions at the single cell level |
Worldwide hundreds of millions of schistosomiasis patients rely on treatment with a single drug , praziquantel . Therapeutic limitations and the threat of praziquantel resistance underline the need to discover and develop next generation drugs . We studied the antischistosomal properties of the Medicines for Malaria Venture ( MMV ) malaria box containing 200 diverse drug-like and 200 probe-like compounds with confirmed in vitro activity against Plasmodium falciparum . Compounds were tested against schistosomula and adult Schistosoma mansoni in vitro . Based on in vitro performance , available pharmacokinetic profiles and toxicity data , selected compounds were investigated in vivo . Promising antischistosomal activity ( IC50: 1 . 4–9 . 5 µM ) was observed for 34 compounds against schistosomula . Three compounds presented IC50 values between 0 . 8 and 1 . 3 µM against adult S . mansoni . Two promising early leads were identified , namely a N , N′-diarylurea and a 2 , 3-dianilinoquinoxaline . Treatment of S . mansoni infected mice with a single oral 400 mg/kg dose of these drugs resulted in significant worm burden reductions of 52 . 5% and 40 . 8% , respectively . The two candidates identified by investigating the MMV malaria box are characterized by good pharmacokinetic profiles , low cytotoxic potential and easy chemistry and therefore offer an excellent starting point for antischistosomal drug discovery and development .
With hundreds of millions of people living at risk of infection and 207 million people infected with schistosomes worldwide , schistosomiasis is one of the most devastating parasitic diseases in tropical countries and remains a major public health problem , especially in Sub-Saharan Africa [1] , [2] . Schistosoma haematobium , S . japonicum and S . mansoni are the main schistosome species , responsible for the largest number of infections [3] , [4] . A major cornerstone of schistosomiasis control programs is the treatment of at risk populations with praziquantel , with the aim of controlling morbidity and preventing associated mortality [5]–[7] . Praziquantel , discovered in the 1970's , is the only drug available for the treatment of schistosomiasis [7]–[9] . Despite many benefits of praziquantel , most notably its high efficacy and excellent tolerability , the drug has major drawbacks , most importantly its inefficacy against juvenile schistosomes [10] , [11] . Furthermore the increasing administration of praziquantel to millions of people annually [12] results in high drug pressure , and thus drug-resistant parasites are likely to evolve [13] . These facts underline the urgent need to discover and develop the next generation of antischistosomals . Only a few compounds are presently being studied in the preclinical phase [14]–[17] and none of the candidates evaluated in clinical trials in the past years ( e . g . mefloquine [18] or the artemisinins [19] ) ( Figure S1 ) met the target product profile for a novel antischistosomal drug [20] . Interestingly many of the chemical scaffolds that revealed promising activity against schistosomes had their origin in antimalarial research and discovery [21] . The blood-feeding characteristic that both parasites have in common forms the basis for the dual antimalarial and antischistosomal activity of drugs interfering with the parasites' hemoglobin degradation pathway [22] , [23] . The aim of the present study was to investigate the antischistosomal properties of the Medicines for Malaria Venture ( MMV ) malaria box containing 200 diverse drug-like compounds ( which fit in the “Lipinski space” or rule of five ) , as a starting point for oral drug discovery and development , and 200 diverse probe-like compounds ( no filters applied ) . Note that all of the compounds in the box have confirmed activity against the blood-stage of Plasmodium falciparum in vitro and are commercially available [24] . Studying this diverse set of molecules might reveal an entirely new chemical scaffold for antischistosomal drug discovery and therefore fill up the empty antischistosomal drug pipeline . At the Swiss Tropical and Public Health Institute ( Swiss TPH ) , drugs were first studied against schistosomula in vitro followed by a re-evaluation of successful hits on adult S . mansoni . In parallel all the drugs were independently tested at the London School of Hygiene and Tropical Medicine ( LSHTM ) in an in vitro adult worm assay . Possible class effects and structure-activity relationships are discussed . The onset of action and IC50/IC90 ratios were studied . Based on in vitro performance and available pharmacokinetic profiles as well as toxicity data , selected compounds were investigated in vivo .
The MMV Box [24] , containing 400 compounds as stock solutions dissolved in dimethylsulfoxide ( DMSO ) , concentration 10 mM , was kindly provided by MMV/SCYNEXIS , Inc . ( Geneva , Switzerland; Durham , USA ) . For the in vitro studies on adult worms at the Swiss TPH and the in vivo studies in mice 5–100 mg of 1: MMV000963 , 2: MMV665852 , 3: MMV665807 , 4: MMV019555 , 5: MMV019918 , 6: MMV000445 , 7: MMV019780 , 8: MMV665927 , 9: MMV665941 , 10: MMV000634 , 11: MMV665830 , 12: MMV666054 , 13: MMV009063 , 14: MMV007591 , 15: MMV665969 , 16: MMV666070 , 17: MMV007224 , 18: MMV665794 , 19: MMV666057 , and 20: MMV665799 were purchased from Specs ( Delft , Netherlands ) , and MolPort ( Riga , Latvia ) . Praziquantel was purchased from Sigma-Aldrich ( Buchs , Switzerland ) GmbH . Compounds 1–20 were dissolved in DMSO for drug stock solutions of 10 mg/ml for in vitro evaluations . Culture medium for newly transformed schistosomula ( NTS ) was made by supplementing Medium 199 ( Lubioscience , Lucerne , Switzerland ) with 5% heat-inactivated fetal calf serum ( iFCS ) , penicillin ( 100 U/ml ) , and streptomycin ( 100 µg/ml ) ( Lubioscience , Lucerne , Switzerland ) . Culture medium for adult worms was prepared by supplementing RPMI 1640 with 5% iFCS , penicillin ( 100 U/ml ) , and streptomycin ( 100 µg/ml ) . S . mansoni cercariae ( Liberian strain ) were harvested from infected intermediate host snails ( Biomphalaria glabrata ) following in-house standard procedures . Collected cercariae were mechanically transformed to NTS as described previously [25] , [26] . The obtained NTS suspension was adjusted to a concentration of 100 NTS per 50 µl using supplemented Medium 199 . NTS suspensions were incubated ( 37°C , 5% CO2 in ambient air ) for a minimum of 12 to 24 hours until usage to ensure completed conversion into schistosomula [27] . In vivo studies were conducted at the Swiss TPH , Basel , and approved by the veterinary authorities of the Canton Basel-Stadt ( permit no . 2070 ) based on Swiss cantonal and national regulations . Experimentation at LSHTM was carried out under the UK Animals Scientific Procedures Act 1986 with approval from the LSHTM Ethics committee . Animals ( female NMRI , 3-week old , weight ca . 14 g ) were purchased from Charles River ( Sulzfeld , Germany ) and allowed to adapt under controlled conditions ( temperature ca . 22°C; humidity ca . 50%; 12-hour light and dark cycle; free access to rodent diet and water ) for one week . Mice were infected by subcutaneous injection with ∼100 S . mansoni cercariae each , harvested from infected snails . For in vitro studies on adult flukes , schistosomes were collected from the hepatic portal and mesenteric veins of infected mice 7–8 weeks post infection [28] . Freshly harvested schistosomes were placed in supplemented RPMI culture medium , quickly rinsed , and stored at 37°C , 5% CO2 until usage . Initially , all compounds were tested at a concentration of 100 µM on S . mansoni NTS . Active compounds progressed into a secondary screening at 33 . 3 µM . For this purpose drug stock solutions were diluted in 96-flat bottom well plates ( BD Falcon , USA ) with supplemented Medium 199 and 50 µl of prepared NTS suspension ( 100 NTS/well ) to the desired final concentration of 100 µM or 33 . 3 µM , respectively . Each drug was tested at least in triplicate and the highest concentration of DMSO served as control . Plates were incubated at 37°C , 5% CO2 . NTS were evaluated by microscopic readout ( Carl Zeiss , Germany , magnification 80–120× ) using a viability scale as previously described with regard to death , changes in motility , viability , and morphological alterations 72 hours post drug exposure [25] , [26] . To ensure the accuracy of our assay , 45 compounds that lacked activity at one of the tested concentrations , were randomly selected and retested at 33 . 3 µM . Compounds that killed the NTS at 72 hours after exposure in at least one well were deemed active and selected for further testing . In the next step , the IC50 was determined for active compounds from the preceding screens . Drug dilution series were prepared in 96-flat bottom well plates with concentrations 2 . 1 , 4 . 2 , 8 . 4 , 16 . 7 , and 33 . 3 µM using supplemented culture medium . The prepared NTS suspension was then added to each well and plates were incubated at 37°C , 5% CO2 . NTS incubated in the presence of the highest DMSO concentration and praziquantel served as control . Drug effects on NTS were evaluated 72 hours post exposure , using a viability scale , as described above . Each concentration was tested in duplicate and experiments were repeated once . Compounds presenting IC50 values ≤10 µM were then tested at a concentration of 33 . 3 µM on adult worms in duplicate . Drug stock solutions ( 10 mM ) were diluted in supplemented RPMI 1640 culture medium reaching a final concentration of 33 . 3 µM in 24-flat bottom well plates ( BD Falcon , USA ) within a final volume of 2 . 4 ml . At least three schistosomes of both sexes were added to each well . Schistosomes incubated in the presence of the highest concentration of DMSO served as control . Plates were incubated for 72 hours at 37°C , 5% CO2 . Seventy-two hours post drug exposure S . mansoni were examined phenotypically by microscope using the motility scale described before [29] . Drugs leading to the death of schistosomes 72 hours post exposure were characterized further and their IC50 ( IC90 ) values were determined . Specifically , drug dilution series were prepared in 24-flat bottom well plates ( BD Falcon , USA ) with concentrations of 0 . 31 , 0 . 93 , 2 . 78 , 8 . 33 , and 25 . 0 µg/ml using supplemented RPMI culture medium and freshly prepared drug stock solutions ( 10 mg/ml ) . At least three schistosomes of both sexes were added to each well and plates were incubated at 37°C , 5% CO2 . Parasites incubated in the highest DMSO concentration and praziquantel served as controls . Drug effects were evaluated 72 hours post exposure as described above . Each concentration was tested in duplicate and trials were repeated once . Adult worm drug testing was performed as previously reported [29] with some modifications as described . Worms of a Puerto Rican strain of S . mansoni were obtained by portal perfusion of CD1 mice ( Charles River , UK ) 6 weeks post-infection . Three pairs of worms were added to the wells of 48-well plates ( Nunc , UK ) in 1 ml complete DMEM medium supplemented with 10% fetal calf serum , 2 mM L-glutamine , 100 U/ml penicillin , and 100 µg/ml streptomycin ( cDMEM ) . Compounds were tested at 15 µM containing 0 . 15% DMSO in single wells . Negative controls contained worms cultured in cDMEM alone and in cDMEM with 0 . 15% DMSO . Positive control wells contained worms cultured in praziquantel ( Sigma-Aldrich , UK ) at 10 µM . Cultures were incubated at 37°C and 5% CO2 . Effects were assessed on day 5 of culture using an inverted microscope ( Leitz Diavert Wetzlar , Germany ) . Any compounds producing complete immotility or ≥70% worm motility inhibition plus severe morphological damage were considered hits in the primary screen [29] . Active compounds were then tested for IC50 value determination at a concentration range from 0 . 55–15 µM in single wells . The onset of action ( length of time needed before an antischistosomal effect was visible ) was determined for selected compounds in vitro by evaluating the IC50 at a time-range of 1–72 hours ( 1 , 2 , 4 , 7 , 10 , 24 , 48 , and 72 hours ) post drug exposure , as described above . The onset of action of praziquantel was also studied . Additionally , possible protein binding effects were studied for three lead candidates and praziquantel . For that purpose RPMI medium was supplemented with two different iFCS concentrations ( 0% and 50% ) and IC50 values were calculated for the different conditions . Furthermore , IC50 values were determined after varying drug exposure times ( 1 , 2 , or 4 hours ) followed by incubation in drug free RPMI medium for 72 hours . Groups of 3–4 NMRI mice characterized by a patent S . mansoni infection ( 49 days post-infection ) were treated orally with the test drug using either single oral doses of 400 mg/kg or 80 mg/kg administered on four consecutive days . An additional dosage regimen of 100 mg/kg administered four times every 4 hours was tested for the 2 most active compounds ( 2 , 17 ) . Compounds were freshly prepared in an aqueous hydroxypropyl methyl cellulose ( HPMC ) ( 1% ) : DMSO ( 95∶5 ) formulation . Eight to sixteen untreated mice served as controls . Fourteen days post-treatment animals were killed by the CO2 method and were dissected and the worms were sexed and counted [28] . Mean worm burdens of treated mice were compared to the mean worm burden of untreated animals and worm burden reductions were calculated . Parasite viability values of NTS and adult schistosomes obtained from microscopic evaluation were averaged ( means ( +/− standard deviation ) ) using Microsoft Excel . IC50 and IC90 values of test compounds were determined using the CompuSyn software ( Version 3 . 0 . 1 , 2007; ComboSyn Inc . , USA ) and Microsoft XLfit version 5 . 1 . 0 . 0 ( 2006–2008 ID Business Solutions Ltd ) . Selectivity indices were calculated by dividing the IC50 of the MRC-5 cells-fibroblast cytotoxicity data by the IC50 of the adult worm assay . The Kruskal-Wallis test was applied for in vivo studies , comparing the worm burden of the treated animals and control animal groups . A difference in worm burden was considered to be significant at a significance level of 5% ( StatsDirect , version 2 . 7 . 2 . ; StatsDirect Ltd . , UK ) .
Exposing schistosomula to the test drugs ( n = 400 ) at a concentration of 100 µM resulted in death of NTS for 45% of the tested compounds ( n = 179 ) . Schistosomicidal effects were observed for 18% of these active compounds ( n = 72 ) at the lower concentration of 33 . 3 µM ( Figure 1 ) . A diverse range of chemical scaffolds was observed amongst active compounds . Successful candidates were characterized further on NTS . Promising antischistosomal activity ( IC50: 1 . 4–9 . 5 µM ) was observed for 34 compounds , two of which were identified during our quality control re-evaluation of 45 compounds and nine of which showed comparable or increased activity ( IC50: 1 . 4–2 . 4 µM ) to praziquantel ( IC50: 2 . 2 µM ) . All hits ( IC50<10 µM ) ( n = 34 ) were next tested at a concentration of 33 . 3 µM on adult S . mansoni . Seventy-two hours post drug exposure , 16 ( 1–16 ) of these compounds ( Table S1 ) killed the adult worms . Four of the ten compounds with high activities ( IC50<2 . 5 µM ) on NTS lacked antischistosomal activity on adult worms . The 16 active candidates were further characterized by IC50 value determination . The highest in vitro activities were observed for the diaminoquinazoline derivative 1 ( IC50: 0 . 8 µM ) the diarylurea 2 and diarylamide 3 , presenting IC50 values of 0 . 8 and 1 . 3 µM , respectively ( PZQ: 0 . 2 µM ) . IC50 values ranging from 2 . 6–9 . 2 µM were calculated for compounds 4–11 , whereas only moderate activity ( IC50 values >10 µM ) was determined for five compounds ( 12–16 ) . Compounds with IC50>10 µM were excluded from further consideration , meaning only eleven compounds were considered as hits . Forty-four compounds were classified as hits ( compounds producing complete immotility or ≥70% worm motility inhibition plus severe morphological damage ) against adult S . mansoni in vitro at a concentration of 15 µM . These compounds were further tested for IC50 values ( Table S1 ) . Twelve compounds showed IC50 values >15 µM . Fourteen compounds revealed IC50 values between 10–15 µM . Eighteen compounds had IC50 values <10 µM . To provide a comparison with the Swiss TPH assays , the 32 hits were subsequently tested using the schistosomula assay at LSHTM [30] . This showed generally good concordance with the LSHTM adult assay , in that all adult hits with IC50<10 µM were also hits in the larval assay ( Table S1 ) . Based on in vitro performance on the adult worms ( Table S1 ) , toxicity , pharmacokinetic ( PK ) properties and availability of the compounds , five lead candidates ( 1 , 2 , 5 , 8 , 17 ) ( Figure 2 ) were selected for in vivo testing and in depth characterization in vitro . In more detail , 11 compounds were excluded after comparing their IC50 values and PK parameters ( Cmax , tmax , t1/2 , AUC ) . Four compounds showed poor antischistosomal activity ( IC50>10 µM ) and four compounds showed poor bioavailability ( Cmax<IC50 of the corresponding compound ) . Ten compounds were characterized by low selectivity indices ( SI<1 ) and two were not commercially available . Four active compounds were derivatives belonging to the class of diarylureas and two compounds were characterized as dianilinoquinoxalines . Only the most active candidate of each chemical group , compound 2 and compound 17 , was selected for in vivo studies . A summary of the IC50 values , toxicity and pharmacokinetic parameters of the lead candidates is provided in Table 1 . The onset of action was studied in compounds selected for in vivo testing ( n = 5 ) and compared to the onset of action for praziquantel ( Figure 3 ) . Compound 2 was the fastest acting drug , presenting an IC50<5 µM already after 1 hour of in vitro exposure , followed by compound 17 with an IC50<10 µM , 1 hour post incubation . Compound 1 was intermediate in speed with an onset time of 7 hours post-incubation . Compound 8 had fully exerted its antischistosomal properties 24 hours following incubation , while compound 5 was slow acting ( exposing its full antischistosomal activity only 72 hours post treatment ) . In comparison , praziquantel exposed its entire antischistosomal activity already after 1 hour of drug exposure ( IC50: 0 . 2 µM ) . The determined IC90 values of the lead candidates were 2–5 fold higher than the observed IC50 values 72 hours post exposure and thus the concentration-response curves for these compounds are quite steep ( Table 2 ) . Comparatively , praziquantel even showed a 13-fold difference between the two values . Praziquantel lead very quickly to a strong motility inhibition and morphological changes , whereas higher concentrations ( IC90: 2 . 0 µM ) were necessary to actually kill the worms . Compound 2 and 17 revealed the highest in vivo activity with worm burden reductions ( WBR ) of 52 . 5% ( dosage 1×400 mg/kg; p<0 . 005 ) and 53 . 4% ( dosage 4×100 mg/kg; p<0 . 005 ) , respectively ( Table 3 ) . In addition , both treatment regimens using multiple doses of compound 2 resulted in significant worm burden reductions of 46 . 0% ( 4×80 mg/kg; p<0 . 005 ) and 31 . 2% ( 4×100 mg/kg; p<0 . 05 ) . Treatment with a single 400 mg/kg dose of compound 17 resulted in a significant worm burden reduction of 40 . 8% ( p<0 . 05 ) , while multiple treatment courses of 80 mg/kg over four consecutive days achieved a lower effect ( WBR: 25 . 5% , p<0 . 05 ) . Compounds 1 , 5 , and 8 lacked in vivo activity ( WBR 0–18 . 7% ) . No significant differences were observed between total and female worm burden reductions . Compound 17 showed a 7-fold increase in activity in iFCS-free medium ( IC50: 0 . 3 µM ) versus incubation in 50% serum supplemented medium ( IC50: 2 . 1 µM ) ( Table S2 ) . A strong increase in activity in serum free medium was observed for praziquantel ( IC50: 0 . 02 µM ) . No altered activities were detected for compound 2 within varying iFCS-concentrations . Short-term exposure of schistosomes to compound 2 or praziquantel ( 1–4 hours ) followed by incubation in drug free medium for 72 hours resulted in high IC50 values , ranging from 51 . 1 µM ( 1 hour ) to 24 . 6 µM ( 4 hours ) for compound 2 and from 96 . 1 µM to 7 . 7 µM for praziquantel ( Figure S3 ) . These values are much higher than the IC50 values determined when the worms are continuously exposed to the drugs for 72 hours ( 2: IC50: 0 . 8 µM; PZQ: IC50: 0 . 2 µM ) . Incubation of schistosomes for 4 hours with compound 17 achieved similar effects ( IC50: 1 . 3 µM ) ( Figure S3 ) as described for the 72 hours exposure time ( IC50: 0 . 8 µM ) ( Table S2 ) .
The aim of this study was to investigate the antischistosomal potential of 200 drug-like and 200 probe-like compounds assembled in the MMV Malaria Box . The MMV Malaria Box provided a unique opportunity: commercially available compounds with confirmed in vitro activity against P . falciparum serve as good starting material for antischistosomal R&D , as many antimalarials have antischistosomal activity [16] , [23] , [31] . In addition , and in line with the target characteristics of a trematocidal lead candidate [20] , properties of the drug-like compounds are commensurate with oral absorption and the presence of known toxicophores is minimized . NTS were used as a prescreening tool at Swiss TPH , since their use greatly reduces the need for laboratory animals and thus is a major contributor to the 3 R rules ( replace , reduce , refine ) [25] . Nearly half of the tested compounds ( 45% ) presented schistosomicidal effects on the schistosomular stage at a concentration of 100 µM . Given this high hit rate , compounds which were not lethal on NTS did not progress further . This might be a limitation of the Swiss TPH screening , since many effective anthelmintics ( including praziquantel at low concentrations ) cause paralysis rather than death of worms [32] . Thirty-four of the active compounds had IC50 values ranging from 1 . 4 to 9 . 5 µM , suggesting that both parasites , P . falciparum and S . mansoni , have a similar drug sensitivity profile . About half of the compounds active against NTS ( n = 16 ) revealed good to moderate activity on the adult stage ( IC50: 0 . 8–22 . 3 µM ) . Several compounds that showed high antischistosomal effects on schistosomula lacked activity on adults . This phenomenon , where the hit rates were higher against the larval stages than against the adult stages , has been previously reported [14] , [33] . A higher sensitivity of the larval stage , or mode of action dependent effects might partially explain this higher hit rate: for example recent studies with various peroxide classes documented less activity on the adult stage than on the NTS stage [15] . The parallel screening at LSHTM screened all compounds directly on adult schistosomes . Thirteen additional compounds active against adult worms ( IC50<10 µM ) were identified at LSHTM . Nine of these lacked activity against NTS at Swiss TPH ( Figure S2 ) . Interestingly , these compounds showed activity against NTS at LSHTM ( Table S1 ) . On the other hand , four compounds with activity ( IC50<10 µM ) against NTS and adult worms identified at Swiss TPH lacked activity in the LSHTM screen . Overall , 22 compounds had an IC50<10 µM against adult worms in at least one of the screens . Only five compounds were characterized by an IC50<10 µM in both screenings . Strain differences but also different ways of assay set up and readout might offer an explanation for these results . Nonetheless , follow up studies to clarify these issues are warranted . Compound 2 , a diarylurea , revealed the highest activity against adult S . mansoni in vitro . In addition , our onset of action studies revealed that it was the fastest acting compound , comparable to praziquantel . The compound is characterized by an intriguingly simple chemistry and can be easily synthesized . The class of N , N′-diarylureas was recently found to activate heme-regulated inhibitor kinase which inhibits translation initiation and plays a central role in cancer initiation [34] . Additionally various N , N′-diarylureas , including compound 2 , have been investigated as potential anti-cancer agents and were proposed as promising lead compounds [35] . Significant worm burden reductions of 52 . 5% , 46 . 0% , and 31 . 2% were observed with compound 2 following single oral dosing with 400 mg/kg , 80 mg/kg on four consecutive days and 4×100 mg/kg every four hours , respectively . This might indicate that in vivo activity follows a time over threshold model rather than it being Cmax driven . However , based on the in vitro performance and pharmacokinetic data , a better in vivo outcome was expected . Our follow-up in vitro studies , which studied protein binding and the short-term drug exposure , might offer an explanation for this discrepancy . Short incubation times ( 1 to 4 hours ) were not sufficient to kill the worms , since most of the parasites recovered 3 days later . Note that compound 2 is characterized by a half-life ( t1/2 ) of 4 . 7 hours and Cmax of 4 . 4 µM at 46 . 3 mg/kg ( po ) . Additionally compound 17 , a 2 , 3-dianilinoquinoxaline derivative , showed high in vitro ( IC50: 0 . 83 µM ) and significant in vivo activity with WBRs between 53 . 4% ( multiple po dose of 100 mg/kg every four hours ) and 40 . 8% ( single po dose 400 mg/kg ) . This series has been reported to show antimycobacterial activity [36] . The order of in vivo activity of the five selected candidates is in line with the onset of action observed in vitro . The fastest acting compound 2 exhibited the highest activity in vivo followed by compound 17 ( WBR: 40 . 8% ) . The discrepancy of excellent in vitro performance of compound 17 , but only moderate in vivo activity might be explained by protein binding effects . Increased activities were observed when incubated sans serum proteins in vitro . Notably , short-term incubation of 4 hours was sufficient to exhibit high antischistosomal effects for both drugs . Compounds 1 , 5 , and 8 acted slower ( only 7–10 hours post exposure ) , and lacked activity in vivo . This finding is in line with PK properties of these drugs . Since the half-lives of the compounds are rather short ( 2 . 4–5 . 2 hours ) plasma concentrations remain insufficiently long above the IC50 values for the slow acting compounds to exert in vivo activity . Since a series of related derivatives was present in the MMV Malaria Box , we carried out an initial structure-activity relationship study by sourcing commercially available near neighbors for compounds 1 , 2 , and 5 ( Table S3 ) . Exchanging the phenyl-group of 1 with an ethanol group revealed a stage specific sensitivity with activity on NTS , but lacked schistosomicidal effects on adult worms . The substitution pattern on phenyl-residues of compound 2 influenced activity . For example , exchanging the para-chloro to a para-fluro on one of the phenyl rings led to a two-fold decrease in activity on NTS . Such subtle changes in activity require further investigation with a larger set given the easy chemical accessibility of derivatives . In conclusion , by screening the MMV malaria box on S . mansoni we underlined the potential of compounds with an antimalarial background on schistosomes . We identified two entirely new chemical scaffolds: the N , N′-diarylurea ( 2 ) and 2 , 3-dianilinoquinoxaline derivatives ( 17 ) with antischistosomal in vitro activity in the sub micromolar range and moderate in vivo activity . The compounds offer promising drug characteristics such as a good pharmacokinetic profile and low cytotoxic potential . Their easy chemistry simplifies further drug optimization steps and offers an excellent starting point for antischistosomal drug discovery and development . | To date , praziquantel is the only available drug for the treatment of the tropical neglected disease schistosomiasis and is widely used in morbidity control programs . To discover new chemical scaffolds for the treatment of schistosomiasis , we investigated the Medicines for Malaria Venture malaria box containing 200 diverse drug-like and 200 probe-like compounds with known antimalarial activity against Schistosoma mansoni . Compounds were first investigated on the larval stage of S . mansoni , followed by testing against adult worms in vitro and by in vivo studies of lead candidates . We identified two entirely new chemical scaffolds: the N , N′-diarylurea and 2 , 3-dianilinoquinoxaline derivatives with antischistosomal in vitro activity in the sub micromolar range and significant activity in the mouse model . Since both compounds offer a good pharmacokinetic profile , low cytotoxic potential and easy chemistry , structure-activity relationship studies should be launched . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"medicine",
"neglected",
"tropical",
"diseases",
"infectious",
"diseases",
"parasitic",
"diseases"
] | 2014 | Orally Active Antischistosomal Early Leads Identified from the Open Access Malaria Box |
In the United States , roughly 10% of the population is exposed daily to hazardous levels of noise in the workplace . Twin studies estimate heritability for noise-induced hearing loss ( NIHL ) of approximately 36% , and strain specific variation in sensitivity has been demonstrated in mice . Based upon the difficulties inherent to the study of NIHL in humans , we have turned to the study of this complex trait in mice . We exposed 5 week-old mice from the Hybrid Mouse Diversity Panel ( HMDP ) to a 10 kHz octave band noise at 108 dB for 2 hours and assessed the permanent threshold shift 2 weeks post exposure using frequency specific stimuli . These data were then used in a genome-wide association study ( GWAS ) using the Efficient Mixed Model Analysis ( EMMA ) to control for population structure . In this manuscript we describe our GWAS , with an emphasis on a significant peak for susceptibility to NIHL on chromosome 17 within a haplotype block containing NADPH oxidase-3 ( Nox3 ) . Our peak was detected after an 8 kHz tone burst stimulus . Nox3 mutants and heterozygotes were then tested to validate our GWAS . The mutants and heterozygotes demonstrated a greater susceptibility to NIHL specifically at 8 kHz both on measures of distortion product otoacoustic emissions ( DPOAE ) and on auditory brainstem response ( ABR ) . We demonstrate that this sensitivity resides within the synaptic ribbons of the cochlea in the mutant animals specifically at 8 kHz . Our work is the first GWAS for NIHL in mice and elucidates the power of our approach to identify tonotopic genetic susceptibility to NIHL .
Noise-induced hearing loss ( NIHL ) is a worldwide leading occupational health risk in industrialized countries and is the second most common form of sensorineural hearing impairment , after presbyacusis [1] . In the United States , roughly 10% of the total population is exposed daily to hazardous levels of noise in the workplace [2] . The most extreme workplace environment for NIHL is the Armed Forces . According to the Department of Veterans Affairs , hearing loss is the most common disability among U . S . troops in the Middle East . The financial impact of these disability claims on the VA is staggering and likely will continue to grow . According to the American Tinnitus Association ( http://www . ata . org/ ) , the number of disability claims from hearing injury is expected to increase by 18% per year with a total cost of $1 . 2 billion annually [3] . Risk could be reduced with a better understanding of the biological processes that modulate susceptibility to damaging noise . It is believed that NIHL is a complex disease resulting from the interaction between environmental and genetic factors and it is well recognized that people with similar exposures to noise show variation in the amount of hearing loss , indicative of a genetic component [4] . Twin studies estimate heritability for noise-induced hearing loss ( NIHL ) of approximately 36% [5] . The discovery of gene by environment interactions in human disease , such as susceptibility to NIHL , has many inherent difficulties , most notably , controlling for exposure . Although several candidate gene association studies for NIHL in humans have been conducted , each is underpowered , un-replicated , and accounts for only a fraction of the genetic risk . In addition , no heritability studies have been performed , since families , where all subjects are exposed to identical noise conditions , are almost impossible to collect . The genetic basis of NIHL has been clearly demonstrated in animals as different susceptibilities to noise have been seen in different inbred stains of mice [4] . Mouse strains ( C57BL/6J ) exhibiting age-related hearing loss ( AHL ) were shown to be more susceptible to noise than other strains [6] . Also , several knockout mice including SOD1-/- [7] , GPX1-/- [8] , PMCA2-/- [9] and CDH23+/- [10] were shown to be more sensitive to noise than their wild-type littermates . The mouse has been an essential animal model for studies in hearing loss , and advances in mouse genetics , including genome sequence and high density single-nucleotide polymorphism ( SNP ) maps , provide a suitable system for the study of a complex trait such as NIHL [6] . The identification of novel genes is crucial for the discovery of new pathways and gene networks that will improve our knowledge of basic hearing biology and identify new therapeutic targets with the potential to combat NIHL . Due to the limitations of human genome-wide association study ( GWAS ) and quantitative trait locus ( QTL ) analyses in mice , we have chosen to use a genome-wide association strategy incorporating the Hybrid Mouse Diversity Panel ( HMDP ) . The HMDP is a collection of classical inbred ( CI ) and recombinant inbred ( RI ) strains whose genomes have been sequenced and/or genotyped at high resolution [11] . Power calculations have demonstrated that this panel is superior to traditional linkage analysis and is capable of detecting loci responsible for 5% of the overall variance . Several studies have successfully mapped candidate loci for complex traits using this panel and we have recently published a meta-analysis for age-related hearing loss incorporating the HMDP [12] [13] [14] [15] . In this manuscript we describe , for the first time , an association analysis with correction for population structure in the mapping of several loci for susceptibility to NIHL in inbred strains of mice . After completing a preliminary screen of the HMDP , an intriguing locus appeared warranting further exploration . Herein , we describe a genome-wide significant peak on ( Chr . ) 17 within a haplotype block containing NADPH oxidase-3 ( Nox3 ) and provide evidence supporting its role in susceptibility to NIHL . Furthermore , we demonstrate frequency-specific genetic susceptibility within the mouse cochlea .
The Institutional Care and Use Committee ( IACUC ) at University of Southern California , Los Angeles , approved the animal protocol for the HMDP strains and the Nox3het mice ( IACUC 12033 ) . HMDP strains and C57BL/6JEiJ Nox3het ( Nox3het/Nox3het , Nox3het/+ and wild-type ) were anesthetized with an intraperitoneal injection of a mixture of ketamine ( 80 mg/kg body weight ) and xylazine ( 16 mg/kg body weight ) . A detailed description of the HMDP ( strain selection , statistical power and mapping resolution ) is provided in Bennett BJ , et al . 2010 . [11] . Approximately four female mice for each HMDP strain were purchased from the Jackson Laboratory ( Bar Harbor , ME ) . Only female mice were tested to avoid confounding effects of sex . Mice were 4 weeks of age , and to ensure adequate acclimatization to a common environment , mice were aged until 5 weeks . 5-week-old mice were selected to eliminate the potential effects of age-related hearing loss contributing to our phenotype . All mice were maintained on a chow diet until sacrifice . Common and recombinant inbred strains were previously genotyped by the Broad Institute ( www . mousehapmap . org ) . Of the 140 , 000 SNPs available , 108 , 064 were informative ( allele frequency ≥ 5% and less than 20% missing data ) and were used for the association analysis . Stainless-steel electrodes were placed subcutaneously at the vertex of the head and the right mastoid , with a ground electrode at the base of the tail . Body temperature was maintained and monitored . Artificial tear ointment was applied to the eyes . Each mouse was recovered on a heating pad at body temperature . Auditory signals were presented as tone pips with a rise and a fall time of 0 . 5 msec and a total duration of 5 msec at the frequencies 4 , 8 , 12 , 16 , 24 , and 32 kHz . Tone pips were delivered below threshold and then increased in 5 dB increments until goal of 100 dB . Signals were presented at a rate of 30/second . Responses were filtered with a 0 . 3 to 3 kHz pass-band ( x10 , 000 times ) . For each stimulus intensity 512 waveforms were averaged . Hearing threshold was determined by inspection of auditory brainstem response ( ABR ) waveforms and was defined as the minimum intensity at which wave 1 could be distinguished . Data was stored for offline analysis of peak-to-peak ( P1-N1 ) values for wave 1 amplitudes . Post-exposure thresholds were evaluated by the same method 2 weeks post-exposure . Distortion product otoacoustic emissions ( DPOAEs ) were analyzed as input/output ( I-O ) functions with 2f1- f2 ( primary measure ) . Primary tones were set at a ratio of f2/f1 = 1 . 2 with the f2 between 8 to 32 kHz ( f2 level set 10 dB less than the f1 level ) and L2 ranging from 20 to 70 dB . The noise floor was measured by averaging 6 spectral points ( above and below the 2f1- f2 ) . After both waveform and spectral averaging DPOAEs were extracted . Threshold was defined as the L2 level needed to produce a DPOAE of 0 dB SPL with a signal to noise ratio ( SNR ) ≥ 3 dB . 6 week-old mice were exposed for 2 hours to 10 kHz octave band noise ( OBN ) at 108 dB SPL using a method adapted from Kujawa and Liberman ( 2009 ) [16] . The OBN noise exposure was previously described [17] . For 2 hours , mice were placed in a circular ¼-inch wire-mesh exposure cage with four shaped compartments and were able to move about within the compartment . The cage was placed in a MAC-1 soundproof chamber designed by Industrial Acoustics ( IAC , Bronx , NY ) and the sound chamber was lined with soundproofing acoustical foam to minimize reflections . Noise recordings were played with a Fostex FT17H Tweeter Speaker built into the top of the sound chamber . Calibration of the damaging noise was done with a B&K sound level meter with a variation of 1 . 5 dB across the cage . A data acquisition board from National Instruments ( National Instruments Corporation , Austin , Texas ) was regulated by custom software ( used to generate the stimuli and to process the responses ) . Stimuli were provided by a custom acoustic system , made up of two miniature speakers , and sound pressure was measured by a condenser microphone . Testing involved the right ear only . All hearing tests were performed in a separate MAC-1 soundproof chamber to eliminate both environmental and electrical noise . For each HMDP strain , both cochleae from each 8-week-old mouse were removed . The inner ear was micro-dissected and the surrounding soft tissue and the vestibular labyrinth was removed . The dissected cochleae were then frozen in liquid nitrogen and then ground to powder . RNA was extracted and purified by placing cochlea samples in RNA lysis buffer ( Ambion ) . The sample was incubated overnight ( 4°C ) , centrifuged ( 12 , 000g for 5 minutes ) to pellet insoluble materials and RNA isolated ( following manufacturer’s recommendations ) . This procedure generates approximately 300 ng of total RNA per mouse . Illumina’s Mouse whole genome expression , BeadChips , was used for the gene expression measurements . Amplifications and hybridizations were performed according to Illumina’s protocol ( Southern California Genome Consortium microarray core laboratory at UCLA ) . RNA was reverse transcribed to cDNA using Ambion cDNA synthesis kit ( AMIL1791 ) and then converted to cRNA and labeled with biotin . Further , 800ng of biotinylated cRNA product was hybridized to prepare whole genome arrays and was incubated overnight ( 16–20 hrs ) at 55°C . Arrays were washed and then stained with Cy3 label . Excess stain was removed by washing and then arrays were scanned on an Illumina BeadScan confocal laser scanner . EMMA is a statistical test for association mapping correcting for genetic relatedness and population structure and consider the mean per strain and also individual measurement per mouse to increase the statistical power . We have previously demonstrated that p <0 . 05 genome-wide equivalent for GWA using EMMA in the HMDP is P = 4 . 1×10-6 ( −log10P = 5 . 39 ) [18] . An R package implementation of EMMA is available online at http://mouse . cs . ucla . edu/emma . RefSeq genes were downloaded from the UCSC genome browser ( https://genome . ucsc . edu/cgi-bin/hgGateway ) using the NCBI Build37 genome assembly to characterize genes located in each association . EMMA was used to calculate association ( P-values ) for the probes corresponding to the RefSeq genes . The confidence interval ( 95% ) for the distribution of distances between the most significant and the true causal SNPs , for simulated associations that explain 5% of the variance in the HMDP , is 2 . 6 Mb [11] . Only SNPs mapping to each associated region were used in this analysis . We selected SNPs that were variant in at least one of the HMDP classical inbred strains . Non-synonymous SNPs within each region were downloaded from the Mouse Phenome Database ( http://phenome . jax . org/ ) . The generation and initial characterization of Nox3het allele was previously described [19] . The Nox3het allele arose spontaneously ( endogenous retroviral insertion into intron 12 ) on the GL/Le strain , but has since been made congenic onto the C57BL/6JEiJ strain . To circumvent the probability of additional alleles from the donor strain this congenic region was backcrossed for more than 10 generations . Since the downless mutant allele is not present in this strain the congenic interval containing Nox3 is likely less than 5 centimorgans ( http://jaxmice . jax . org/strain/002557 . html ) . Nox3het ( known as the head-tilt or het mice ) carry autosomal recessive , spontaneous mutations that lead to otoconial absence with no apparent abnormalities in other organs . The otoconia deficit results in head-tilting behavior and absent vestibular-evoked potentials ( VsEPs ) but normal thresholds ABR [20] . Pre exposure ABR , DPOAE and VsEP in male and female mice ( 5 weeks old ) of varying Nox3het genotype ( Nox3het/Nox3het and Nox3het/+ ) and wild-type ( C57BL/6JEiJ strain ) was measured as described above . Pre-exposure threshold levels were obtained at 1 week prior to noise exposure and the animals were assessed for noise damage 2 weeks after exposure . The ABR permanent threshold shift ( PTS ) was defined as the difference between pre-exposure and post-exposure thresholds at each tested frequency . One-way ANOVA was used to test the significance and post hoc Tukey test for multiple comparisons . Mice were sacrificed less than 24 hours after the post exposure ABR . Cochleae were dissected from the surrounding tissues and openings were made into the coils by piercing the apex and rupturing both the oval and the round windows . The dissection was done in cold PBS . After dissection , cochleae were fixed in 4% paraformaldehyde for overnight at 4°C and then washed with PBS . Further dissection was done to expose the organ of Corti . For permeabilization and blocking , tissue was immersed for 1 hour in PBS containing 0 . 2% Triton X-100 ( Sigma Chemical ) and 16% normal goat serum ( SouthernBiotech ) . Samples were incubated overnight at room temperature with primary antibodies ( rabbit anti-myosin6 , 1:500 , Proteus Biosciences and purified mouse anti-CtBP2 , 1:500 , BD Biosciences ) for doubled-staining . Secondary antibody was then applied and tissue was incubated in the dark overnight ( Alexa 594 donkey anti-rabbit , 1:500 , Life technologies and Alexa Fluor-488 anti-mouse , 1:500 , Life technologies ) . After , samples were washed three times in PBS and mounted on glass slides using Fluoromount G ( SouthernBiotech ) . Microscopy was carried out with a laser confocal microscope ( Olympus IX81 ) with epifluorescence light ( Olympus Fluoview FV1000 ) . Outer hair cell loss ( % per 100μm ) was counted and plotted as cytocochleogram by relating distance of cochlear apex to the tonotopic map of mice of strain CBA [21] . Percentages indicate the normalized location of the inner and outer hair cells in the cochlea ( 0% , apical and 100% , basal end ) in 10% steps . Synaptic ribbon density was plotted for each correspondent ABR frequency ( 4 , 8 , 12 , 16 , 24 and 32 kHz ) against the same tonotopic map . Inner hair cells were analyzed in a row ( 50 μm ) for each frequency . CtBP2 immunofluorescence spots were counted in z-stacks and divided by the number of inner hair cells ( measured as the quantity of nuclei ) in the sample . Polymerase chain reaction ( PCR ) was performed for Nox3 using the following primers: Nox3-int12F , GTTCTGGAGCACCACCTTGT; Nox3-int12R CCCATAGGGAGCCAAGAAAT; and ERV-R , TGTCAAGCTGACTCCACCAG [19] . PCR products were separated on a 1 . 5% agarose gel containing 0 . 5 mg/ml ethidium bromide .
In an effort to identify genomic regions associated with NIHL susceptibility , we phenotyped 5-week old female mice ( n = 297 ) from 64 HMDP strains ( n = 4–5/strain ) for thresholds after noise exposure using Auditory Brainstem Response thresholds at specific ABR stimulus frequencies . The stimuli consisted of 4 , 8 , 12 , 16 , 24 and 32 kHz tone bursts . A wide range of ABR thresholds were observed across the HMDP with differences of 3 . 22-fold between the lowest and the highest strains for thresholds at 8 kHz post-noise exposure ( Fig . 1 ) . Frequencies of 4 , 12 , 16 , 24 and 32 kHz demonstrated differences of 1 . 55 , 3 . 25 , 3 . 57 , 2 . 74 and 3 . 75-fold , respectively . EMMA algorithm was applied to each phenotype separately to identify genetic associations for the six tone-burst stimuli [18] . Adjusted association p-values were calculated for 108 , 064 SNPs with minor allele frequency of > 5% ( p < 0 . 05 genome-wide equivalent for GWA using EMMA in the HMDP is p = 4 . 1 x 10-6 , -log10P = 5 . 39 ) . At this threshold , genome-wide significant associations on Chr . 2 ( rs27972902; p = 8 . 6x10-7 ) and Chr . 17 ( rs33652818; p = 2 . 3x10-6 ) were identified for the 8 kHz stimuli ( Table 1 , Fig . 2 ) . Additionally , a significant association signal on Chr . 15 ( rs32934144; p = 1 . 7x10-6 ) was identified for the 16 kHz tone burst and two significant regions on Chr . 3 ( rs30795209; p = 5 . 5x10-7 ) and Chr . 15 ( rs32278602; p = 5 . 9x10-7 ) were identified at 32 kHz . Within each association peak there were 4 ( Chr . 15 ) , 11 ( Chr . 3 ) , 10 ( Chr . 17 ) and 2 ( Chr . 2 ) unique RefSeq genes . We next identified genes within each of the five intervals possessing functional alterations . Genes were selected based upon their regulation by a local expression QTL ( eQTL ) in the HMDP or if they harbored a non-synonymous ( NS ) SNP that was predicted to have functional consequences . For the eQTL analysis , we generated gene expression microarray profiles using RNA isolated from cochleae in 64 HMDP strains ( n = 3 arrays per strain ) . EMMA was then used to perform an association analysis between all SNPs and array probes mapping within each region . A total of 18 , 138 genes were represented by at least one probe , after excluding probes that overlapped SNPs , present among the classical inbred strains used in the HMDP ( see Methods ) . Of these , 6 genes ( 4 within Chr . 3 association and 2 within Chr . 17 association ) were identified with at least one probe whose expression was regulated by a local eQTL ( Table 2 ) . However , the only probe whose expression was regulated by a significant local eQTL in the cochlea was located on Chr . 17 . We determined whether any of the 27 genes implicated in our preliminary GWAS had a defined role in the inner ear . The associations on Chr . 2 , 3 and 15 did not harbor known cochlear genes . Only NADPH oxidase 3 ( Nox3 ) on Chr . 17 had been implicated in inner ear biology with mutants lacking otoconia in the utricular and saccular maculae [22] and its high expression in the inner ear [23] . Of all genes at the chromosome 17 locus , one gene , Tfb1m , had a significant ( 1 . 08x10-6 ) eQTL ( Fig . 3 ) . Of note , Nox3 , the gene in which our peak GWAS SNP is located , does not have an eQTL in the cochlea; however , there was a clear demonstration [23] that Nox3 is highly expressed ( at least 50-fold higher than in any other tissues ) in specific portions of the inner ear . Based on these data and the location of our peak GWAS SNP ( rs33652818 ) , we focused on Nox3 as a plausible candidate gene for NIHL at the chromosome 17 locus . To directly test the hypothesis that Nox3 was associated with susceptibility to NIHL we characterized previously generated Nox3het mice for pre- and post-noise exposure ABR thresholds and PTS after 4 , 8 , 12 , 16 , 24 and 32 kHz tone-burst stimuli . Consistent with our original GWAS finding , this analysis revealed a statistically significant reduction in the PTS in wild-type mice ( C57BL/6JEiJ strain ) compared to Nox3het/+ and Nox3het/Nox3het at 8 kHz ( Fig . 4 ) . As a comparison , the effects of the peak SNP ( rs33652818 ) at the Nox3 locus on ABR at various frequencies is shown in Fig . 5 . Interestingly , there were significant differences as a function of genotype at both the 4 kHz and the 8kHz test frequencies , although the level of significance at 4 kHz ( p = 1 . 1x10-4 ) is only suggestive ( S1 Fig ) and does not reach genome-wide significance ( Table 1 ) . Thus , the significant and highly suggestive association of rs33652818 with ABR at 8 and 4 kHz , respectively , in the HMDP , as well as the frequency-specific phenotype exhibited by the Nox3het/Nox3het mice , suggests that Nox3 may be involved in NIHL at the lower end of the frequency spectrum . For a detailed analysis of the entire auditory pathway , we next evaluated outer hair cell ( OHC ) activity using DPOAE and the inner hair cell ( IHC ) and neuronal responses by ABR wave I peak-to-peak amplitudes . Despite the absence of a statistically significant difference in DPOAE thresholds ( Fig . 6A ) at 8 , 16 , 22 and 32 kHz , there was a pronounced difference at 8 kHz in the wave 1 ABR peak-to-peak amplitudes ( Fig . 6B ) . The DPOAE ( Fig . 7A ) suprathreshold amplitudes ( dB SPL ) and ABR wave 1 amplitudes ( μV ) ( Fig . 7B ) for the 8 kHz tone burst were compared at different stimulus intensities . Both analyses demonstrated statistically significantly less noise damage in the wild-type in comparison to the heterozygous and mutant mice . To confirm these electrophysiological findings , we collected cochleae from pre- and post-noise exposure Nox3het mice and wild-type . First , we assessed OHC loss throughout the entire cochlea by creating a cytocochleogram ( Fig . 8A ) of immunolabeled ( Fig . 8B ) whole-mount organs of Corti to correlate with the DPOAE findings . Subsequently , the IHC afferent synaptic density ( Fig . 9 ) was analyzed as a marker of the neuronal responses ( suprathreshold ABR wave 1 amplitude ) . Despite the absence of a statistical significance in OHC loss , the Nox3het/+ and Nox3het/Nox3het mice demonstrated a significantly reduced post-noise exposure density of synaptic ribbons ( at the 8kHz tonotopic location ) .
We have , for the first time , used association analysis with correction for population structure to map several loci for hearing traits in inbred strains of mice . Our results identify a number of novel loci for susceptibility to NIHL . Additionally , our study demonstrates frequency-specific genetic susceptibilities to noise within the cochlea and the power of our GWAS to detect frequency-specific loci that are precisely recapitulated in a mutant mouse model . Mouse GWAS has revolutionized the field of genetics and has lead to the discovery of hundreds of genes that are involved in complex traits [24] . Our successful mapping largely came from the initial observation that there was a clear strain variation at all post noise exposure hearing phenotypes , reiterating the contribution of genetic factors to NIHL susceptibility . This wide distribution of phenotypes and genotypes facilitated our high-resolution genetic mapping . We used a combined set of 64 classic inbred and recombinant inbred strains , a portion of the HMDP , as an extension of the classical inbred strain association . This increased the statistical power of the classical association studies by including a set of recombinant inbred strains in the mapping panel [25] . The HMDP provided significant statistical power and resolution to identify a locus for NIHL susceptibility that was precisely modeled in a mutant strain [26] . Although this panel is composed of 100 commercially available inbred strains , with roughly two-thirds of this panel we were able to map 5 loci , reflecting the power to detect loci with moderate effect . In addition to the power present in this resource , the resolution of this panel is , in some cases , two orders of magnitude better than that achieved with linkage analysis , as we have recently demonstrated in our mouse GWAS for age-related hearing loss [27] . In an unprecedented manner , this new paradigm was applied to the first high-resolution mapping of candidate genes for NIHL susceptibility . Our GWAS generated significant associations in at least five loci at three different post-noise exposure stimulus frequencies , corresponding to a total 27 candidate genes . All of these candidate genes require adequate characterization , but the first gene to be validated by a genetic mutant mouse model was Nox3 . Nox3 was selected for further investigation based upon its relatively restricted expression in the cochleo-vestibular epithelium and spiral ganglion neurons [23] . The Nox3 gene was described in 2000 based upon its sequence similarity to other Nox isoforms ( encodes an NADPH oxidase ) [28] . The overall structure of Nox3 is highly similar to that of Nox1 and Nox2 [29] and Nox3 shares 56% amino acid with Nox2 [30] . Encoded by Nox3 , the six-transmembrane NADPH-binding protein interacts with a two-transmembrane protein ( encoded by Cyba ) and a cytosolic protein ( encoded by Noxo1 ) . This activation releases a functional NADPH oxidase complex that is able to transporting electrons across membranes towards oxygen ( O2 ) generating superoxide ( O2•- ) and subsequent reactive oxygen species ( ROS ) [19] . First studies on the Nox3 function were published in 2004 and generated the definition of Nox3 as an NADPH oxidase of the inner ear [23][22] . Banfi , et al . , performed analysis of Nox3 distribution ( real time PCR and in situ hybridization ) and reported high Nox3 expression in the inner ear ( cochlear/vestibular sensory epithelia and the spiral ganglion ) . Following exposure to cisplatin , HEK293 cells transfected with Nox3 produced O2•- spontaneously and generated a dramatic increase in O2•- production [23] . Paffenholz et al . [22] reported that mutations of the het locus affect Nox3 and that these head tilt mice ( het ) have impaired otoconial formation in the utricle and saccule resulting in balance defects , such as the inability to detect linear acceleration or gravity . Based upon this finding we chose to pursue interrogation of Nox3 , a gene within our locus on Chr . 17 . Subsequent studies have established a role for the Nox3 gene as the primary source of ROS generation in the cochlea , especially induced by cisplatin ototoxicity [31] . The knockdown of Nox3 ( pretreatment with siRNA ) prevented cisplatin ototoxicity with preservation of hearing thresholds and hair cells . Also , it reduced the expression of Nox3 and biomarkers of damage ( TRPV1 and KIM-1 ) in cochlear tissues [32] . siRNA-mediated gene silencing of Nox3 alleviated cisplatin-induced hearing loss in rats and reduced apoptosis of the sensory hair cells in the cochlea [33] . Although there was no similar evidence regarding NIHL , this key role for Nox3 in the development of cisplatin ototoxicity confirming its role in regulatory mechanisms of cochlear damage encouraged us to validate this candidate gene for NIHL . The only study exploring NIHL and the NOX family ( including Nox3 ) was completed in rats [34] . This study did not indicate whether the Nox3 gene decreased or increased the susceptibility to noise , but instead it evaluated Nox3 expression levels after noise exposure . Some members of the NADPH oxidase family ( Nox1 and Duox2 ) were up-regulated in the rat cochlea after noise exposure , suggesting that these isoforms could be linked to cochlear injury . In contrast , the Nox3 isoform was down-regulated after exposure to 100 dB SPL and 110 dB SPL by seven and fivefold respectively , which could represent an endogenous protective mechanism against oxidative stress . This protective mechanism may have decreased the impact of the noise among wild-type rats by reducing the expression of Nox3 and decreasing the difference related to mutants . However , the in vivo data was based on the use of a non-specific Nox inhibitor that targeted multiple members of this enzyme without conclusively demonstrating that Nox3 plays a role in NIHL . Our study , by contrast , has used animal models with naturally occurring genetic variation and specific genetic perturbation of Nox3 to directly implicate this oxidative stress enzyme in hearing . According to our study , noise exposure might have an opposite effect to cisplatin on Nox3 expression , suggesting differential involvement of Nox3 on noise and cisplatin-induced cochlear damage . Based upon this literature we hypothesized that the absence or reduction of the Nox3 gene product , responsible for the production of ROS in the cochlea , would reduce susceptibility to noise and were startled by our findings . A review of the literature shows there are several key protective mechanisms attributed to the Nox family of genes . These mechanisms include: host defense and inflammation ( ROS-dependent killing , inactivation of microbial virulence factors , regulation of pH and ion concentration in the phagosome and anti-inflammatory activity ) , regulation of gene expression ( TNF-alpha , TGF-beta1 and angiotensin II ) , cellular redox potential , cellular signaling ( inhibition of phosphatases , activation of kinases , regulation of ion channels and Ca2+ signaling ) , oxygen sensing ( kidney , carotid body and lungs ) , biosynthesis , regulation of blood pressure , cell growth , angiogenesis , differentiation and senescence [30] . These protective mechanisms may very well play a role in the findings of susceptibility to NIHL in the wild-type animals . We were able to validate our frequency-specific GWAS findings in isolation by studying Nox3het mutant mice . After noise exposure there was a statistically significant difference between the wild-type mice in comparison to the homozygous mutants and the heterozygotes on several measures of auditory function specifically and solely after the 8 kHz exposure . Contrary to the initial expectations , the presence of the Nox3 gene was clearly protective against noise damage . Also we were able to demonstrate the genotypic effect of the peak SNP at the same GWAS phenotype at 8 kHz . We also show genotypic effect on 4 kHz , but this finding was only suggestive in GWAS and not confirmed in Nox3het mutants . We dissected this phenotype in detail physiologically by assessing OHC function using DPOAEs and IHC/auditory nerve function using ABR . Although there was no statistically significant difference in DPOAE thresholds amongst the genotypes , there was a marked difference in the amplitude of wave 1 of the ABR after suprathreshold stimulation with the 8 kHz tone burst . This suggested that the mechanism of hearing loss , in relation to Nox3 , resided in the spiral ganglion neurons and likely at 8 kHz along the cochlear place map . There are many genes differentially expressed along the tonotopic axis of the cochlea , and this has been shown for Nox3 [35] . It is likely that our frequency specific finding of variation in susceptibility to NIHL is the result of this tonotopic expression pattern . Considering that all of the results pointed to the area of 8 kHz , we initiated a thorough electrophysiological and histological dissection at this particular frequency . The evaluation of the DPOAEs and suprathreshold wave 1 ABR amplitudes was performed at multiple stimulus intensity levels . For each study , the wild-type were more resistant to NIHL at only at 8 kHz . We performed immunohistochemistry two weeks after the noise exposure . Although the difference in OHC loss was not significant , we demonstrated a significantly higher density of synaptic ribbons in wild-type mice . Thus , the electrophysiological findings were verified by the immunohistochemistry , demonstrating that the presence of Nox3 is protective at the neuronal level and that the sensory neural hearing loss after noise exposure occurred at this level of the peripheral auditory system . The absence of differences in outer hair cell count was also verified by its corresponding electrophysiological measure of DPOAE thresholds . However , through the evaluation of DPOAE suprathreshold amplitudes , we were able to observe a statistically significant higher amplitude in the wild-type mice . These three different measures of the integrity of the outer hair cells ( outer hair cell count , DPOAE thresholds and DPOAE suprathresholds amplitudes ) have different sensitivity profiles to demonstrate the impact of noise . Probably DPOAE suprathreshold amplitude is the most sensitive measurement , since there is greater signal-noise ratio . This metric indicates that there is significantly less impact on the activity of the outer hair cells in wild-type mice . Although Nox3 is associated with production of O2•- in the inner ear , the Nox family has several physiological and potentially protective mechanisms . Definitely , this protective role explains the fact that the absence of Nox3 increased susceptibility to NIHL in our mouse models . However , there is a lack of specific studies about the mechanisms of the Nox3 gene due to this very focal expression in the inner ear and functional data on Nox3 have been only gathered in overexpression systems [36] . Most evidence regarding these mechanisms is derived from other isoforms , like Nox2 , which is functionally similar to Nox3 [29] . Thus , due to the limited literature , we relied on the other isoforms to formulate hypotheses about the mechanisms of susceptibility to NIHL . Since ROS are commonly related to inflammation , an anti-inflammatory activity of NOX enzymes would seem illogical . However , over recent years there has been a striking number of publications pointing in the opposite direction . Most of the data about the anti-inflammatory activity of Nox enzymes comes from studies using mice deficient in the phagocyte NADPH oxidase Nox2 as demonstrated by a decreased capacity to degrade phagocytized material in Nox2-deficient cells leading to the accumulation of debris [37] . Also , this hyperinflammation might be due to a lack of ROS-dependent signaling in Nox2-deficient phagocytes and ROS-dependent attenuation of Ca2+ signaling contributing to enhanced inflammation . Lastly , impairment of oxidative inactivation of proinflammatory mediators leads to a prolongation of the inflammatory response [30] . Hyperinflammation in NADPH oxidase-deficient mice was demonstrated in mouse models of Helicobacter gastritis [38][39] , arthritis [40] , demyelinating disease [41] , and sunburn [42] . In experimental lung influenza infection , Nox2 deficient mice demonstrated larger inflammatory infiltrates [43] . Also , by studying endothelial dysfunction , the absence of Nox4 resulted in reduction of endothelial nitric oxide synthase expression , nitric oxide production , and heme oxygenase-1 expression , which was associated with apoptosis and inflammatory activation [44] . There is mounting evidence that NOX enzymes have a role in limiting the inflammatory response and we have shown this to be true in noise-induced cochlear damage . This anti-inflammatory activity of NOX enzymes is poorly understood in the cochlea . So far , as described in other isoforms , our initial hypothesis is that there are important protective mechanisms , such as an anti-inflammatory response resulting from noise exposure . This anti-inflammatory mechanism would be crucial to protect the cochlea against noise injury , overcoming its potential for damage caused by the release of ROS . In this manuscript we report the first functional validation of a gene of the auditory system arising from a GWAS . We demonstrate Nox3 is involved in susceptibility to NIHL and mice deficient in Nox3 are the most susceptible . This finding was specific to 8 kHz both physiologically ( ABR threshold and DPOAE/ABR suprathreshold measures ) and histologically at the level of the hair cell/auditory neuron synaptic ribbon . Our findings validate the power of the HMDP for detecting NIHL susceptibility genes and the tonotopic genetic susceptibilities present within the mouse cochlea . | Noise-induced hearing loss ( NIHL ) is the most common work-related disease in the world and the second cause of hearing loss . Although several candidate gene association studies for NIHL in humans have been conducted , each are underpowered , un-replicated , and account for only a fraction of the genetic risk . Buoyed by the prospects and successes of human association studies , several groups have proposed mouse genome-wide association studies . The environment can be carefully controlled , facilitating the study of complex traits like NIHL . In this manuscript , we describe , for the first time , an association analysis with correction for population structure for the mapping of several loci for susceptibility to NIHL in inbred strains of mice . We identify Nox3 as the associated gene for susceptibility to NIHL that the genetic susceptibility is frequency specific and that it occurs at the level of the cochlear synaptic ribbon . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [] | 2015 | Genome-Wide Association Study Identifies Nox3 as a Critical Gene for Susceptibility to Noise-Induced Hearing Loss |
During CNS development , the nuclear protein SATB2 is expressed in superficial cortical layers and determines projection neuron identity . In the adult CNS , SATB2 is expressed in pyramidal neurons of all cortical layers and is a regulator of synaptic plasticity and long-term memory . Common variation in SATB2 locus confers risk of schizophrenia , whereas rare , de novo structural and single nucleotide variants cause severe intellectual disability and absent or limited speech . To characterize differences in SATB2 molecular function in developing vs adult neocortex , we isolated SATB2 protein interactomes at the two ontogenetic stages and identified multiple novel SATB2 interactors . SATB2 interactomes are highly enriched for proteins that stabilize de novo chromatin loops . The comparison between the neonatal and adult SATB2 protein complexes indicates a developmental shift in SATB2 molecular function , from transcriptional repression towards organization of chromosomal superstructure . Accordingly , gene sets regulated by SATB2 in the neocortex of neonatal and adult mice show limited overlap . Genes encoding SATB2 protein interactors were grouped for gene set analysis of human GWAS data . Common variants associated with human cognitive ability are enriched within the genes encoding adult but not neonatal SATB2 interactors . Our data support a shift in the function of SATB2 in cortex over lifetime and indicate that regulation of spatial chromatin architecture by the SATB2 interactome contributes to cognitive function in the general population .
SATB2 is a highly conserved nuclear protein that binds to matrix attachment regions in DNA via a homeodomain and two CUT domains [1] . In the embryonic cortex , SATB2 is expressed in superficial cortical layers and determines callosal vs . subcortical projection neuron identity by repressing Ctip2 transcription [2–4] . During CNS maturation SATB2 expression shifts towards the deep cortical layers and in the adult brain SATB2 is expressed in pyramidal neurons of all layers of the cerebral cortex and in the CA1 area of the hippocampus [5] , indicating a function in cognition . By eliminating SATB2 selectively from the mouse forebrain after the 3rd postnatal week of life , we and others have shown that it is indeed required for stabilization of long-term potentiation and long-term memory in the adult CNS [6 , 7] . In humans patients , alterations of the SATB2 gene ( de novo structural or single nucleotide variants ) cause developmental delay , intellectual disability , limited to absent speech and behavioral issues [8–14] . Human SATB2 has also been identified as a risk locus for schizophrenia [15 , 16] . To which extent symptoms in SATB2-related human pathologies depend on developmental or adult functions of the protein remains to be established . Potential approaches to address differential function of a transcriptional regulator include defining its protein interactors and characterizing the transcriptional responses that depend on it . SATB2-driven gene expression programs in the neonatal mouse cortex have previously been identified [17] . At protein level , in vitro and in vivo assays have implicated interactions of SATB2 in the embryonic cortex with components of the NuRD complex , including histone deacetylase 1 ( HDAC1 ) and metastasis-associated protein 2 ( MTA2 ) , as well as with the SKI protein causing repression of the Ctip2 locus and active suppression of the subcortical projection neuron fate [3 , 18] . However , SATB2 multiprotein complexes have not been analyzed using unbiased proteomic approaches , either in the developing or the adult cortex . Also unknown are SATB2-driven changes in gene transcription in the adult cortex . Here , we combine proteomic and transcriptomic approaches to characterize and compare SATB2 interacting partners and SATB2 transcriptional activity in neonatal vs . adult mouse cortex . We show that SATB2 interacts with different protein networks at the two ontogenetic stages and regulates distinct gene expression programs linked to cell projection morphogenesis and brain development at the neonatal stage vs synapse , neurotransmitter transport , and calcium ion binding and signaling at the adult stage . By combining our unbiased proteomic findings with analyses of human genetic data , we demonstrate that the genes encoding adult but not those coding for neonatal SATB2 interactors are enriched in common variants associated with cognitive function in the general population .
To identify SATB2-containing protein complexes , we immunoprecipitated endogenous SATB2 from neonatal and adult mouse cortical tissue and subjected the precipitates to liquid chromatography / mass spectrometry ( MS ) analysis ( Fig 1A ) . Cortical lysate from Satb2 knock out mice served as a negative control at both developmental stages . Additionally , we filtered out all non-nuclear proteins ( 6 in the neonatal cortex and 16 in the adult cortex SATB2 proteomes ) by bioinformatics means . This analysis identified 40 proteins in the SATB2 immunoprecipitates from neonatal cortex ( S1 Table ) and 53 proteins in the SATB2 immunoprecipitates from adult cortex ( S2 Table ) . The unfiltered neonatal and adult cortex SATB2 interactomes are provided in S8 and S9 Tables . Of note , the filtration based upon nuclear localization had minimal impact on the composition of the identified SATB2 proteomes ( S1 Fig , S8 and S9 Tables ) , most likely due to the specificity already achieved by using a knock-out lysate as a negative control . Among the neonatal SATB2 interactors were the previously reported HDAC1 and MTA family members [18–20] . All 90 remaining proteins represent newly identified SATB2 binding partners ( Fig 1B ) . In line with the reported binding of SATB2 to nuclear matrix attachment regions of the DNA [19] , we identified nuclear matrix proteins among the novel SATB2 interactors [21] , including HNRNPs , casein kinase II ( CKII ) , nucleolin , centromere binding protein ( CNP ) and scaffold attachment factor B1 ( SAFB1 ) . To validate MS data , we performed independent direct and reverse immunoprecipitations ( IP ) from cortical lysates , followed by immunoblotting ( IB ) with specific antibodies . Heterogeneous nuclear ribonucleoprotein L ( HNRNPL ) , HNRNPL-like , HNRNPC , and DHX9 were chosen for validation from the group of RNA-binding proteins; CUX1 , SATB1 , and ZFN638 were selected as representatives of the transcription factor group; HDAC1 was included as a member of the NuRD complex [18 , 20]; Lamin A/C , BAF ( barrier-to-autointegration factor ) and LAP2—as representatives of the nuclear lamina . All tested novel SATB2-binding partners were highly enriched in precipitates from wild-type lysates compared to SATB2-deficient lysates ( Fig 2A ) . SATB2 was also readily immunoprecipitated by a CUX1-directed antibody in a reciprocal IP , but not by a control rabbit IgG ( Fig 2B ) . To reveal functional relationships among the identified SATB2-interacting proteins , we used the STRING database [22] . Besides the NuRD complex [3 , 18 , 20] , we identified several novel SATB2-containing protein complexes , such as nuclear lamina-associated protein complex ( lamins , the LEM domain proteins LEM2 , LEM3 , LAP2 , and BAF ) , nuclear pore complex , RNA-binding/processing proteins and nucleolar proteins ( Fig 1B ) . The interactions of SATB2 with proteins of the nuclear pore complex were specific for the adult cortex , whereas the nucleolar proteins ( NOP58 , NHP2L1 , DKC1 , and FBL ) were found exclusively in the neonatal interactome . The protein complexes containing RNA-binding/processing proteins , nuclear lamina-associated proteins , and HDAC1-associated proteins were shared between the neonatal and adult SATB2 interactomes , although individual protein components differed ( Fig 1B ) . To test if this is determined by differences in expression levels of the interacting proteins , we compared the mRNA expression of SATB2-binding partners in neonatal vs . adult cortex ( Fig 2C ) . We observed a decreased expression of almost all neonatal SATB2 binding proteins in the adult cortex compared to neonatal cortex , thus providing a potential explanation for the lack of detected interactions in the adult stage . Conversely , while some of the adult cortex SATB2 interactors ( such as DMD , FMN1 , LMNA , PTK2B , RANBP2 , TSC22D4 ) showed increased expression in the adult compared to neonatal cortex , the majority of them had lower expression levels in the adult cortex , indicating that changes in expression are insufficient to explain the differences in the composition of the two interactomes . Next , we used the ConsensusPathDB bioinformatics tool to test which experimentally verified mammalian protein complexes are overrepresented in the neonatal and adult cortex SATB2 proteomes . Consistent with the previously described function of SATB2 as transcriptional repressor , the most enriched protein complexes in the neonatal SATB2 interactome were the SNF2h-cohesin-NuRD complex and the MTA1-HDAC core complex ( S3 Table ) . In contrast , the most overrepresented protein complexes in the adult cortex SATB2 proteome were C-complex spliceosome , Toposome , ATAC B Complex , Lamin A/C/Lamin B1/Lamin B2 , Spliceosome , capped , methylated pre-mRNP: CBC complex , and RanGAP/RanBP1/RanBP2 ( S4 Table ) . In agreement with these observations , the gene ontology ( GO ) enrichment analysis by the Metascape tool also revealed different overrepresented biological pathways and GO terms in the neonatal vs adult SATB2 cortical interactomes ( Fig 3A and 3B , S1 Fig ) . We found enrichment of NURD Complex , RNA splicing in the neonatal SATB2 interactome , in contrast to toposome , chromatin binding , nuclear organization found in the adult SATB2 intractome . Taken together , these results indicate a shift in the functions of neonatal vs . adult cortex SATB2 interactome from transcriptional repression towards organization of chromosomal structure . Of note , SATB2-binding partners in both neonatal and adult cortex were significantly enriched in proteins found at stable de novo chromatin loops [23] ( Fig 2C ) ( adult SATB2 interactome , OR = 25 . 31 , p < 0 . 0001; neonatal SATB2 interactome , OR = 28 . 15 , p < 0 . 0001 ) . Among the shared proteins were members of the HNRNP machinery ( HNRNPL , HNRNPL-like , HNRNPC , HNRNPU-like 2 , HNRNPA2/B1 ) and the ATP-dependent RNA helicase A ( DHX9 ) known to be specifically recruited to long-term stable chromatin loops and required for stabilizing loop topology [23] . This finding is consistent with a function of SATB2 as a chromosomal scaffolding protein [24] and higher-order chromatin organizer [25] . Our proteomic analysis demonstrated differences in the composition of SATB2 protein complexes at neonatal vs adult developmental stage . To explore if these differences contribute to differential gene regulation by SATB2 , we compared the sets of genes that are influenced by SATB2 in the cortex at the two ontogenetic stages . We performed gene expression analysis of adult cortex of SATB2-deficient and control mice by RNA-seq and identified 1157 differentially expressed genes ( adjusted p value <0 . 05 ) . To allow for a correct comparison , we re-analyzed the RNA-seq data published by McKenna et al . [17] that describe transcriptome changes in P0 SATB2-mutant cortices . We applied identical bioinformatics procedures ( pre-processing steps and threshold for differential expression ) in the analysis of the two RNA-seq datasets . To compare the global gene expression signatures of SATB2 mutant vs wild-type cortex at neonatal and adult stage , we first applied a threshold-free approach using the rank-rank hypergeometric overlap ( RRHO ) analysis [26] . This algorithm ranks the entire gene lists ( without introducing any cutoffs ) according to a signed log10-transformed t-test P-value and steps through the two gene lists to calculate if the number of overlapping genes is significantly more or less than would be expected by chance . The RRHO heatmap ( visualizing the matrix of the hypergeometric P-values ) and the rank-rank scatter plot ( in which each gene is plotted by its rank based on the direction-signed , log10-transformed t-test P-values in each gene list ) demonstrated only a very weak overlap between the neonatal and adult cortex gene sets ( Spearman’s ρ rank correlation coefficient = 0 . 013 ) . The overlap is mostly in the genes up-regulated in the SATB2 mutant cortex ( upper right quadrant in the RRHO heatmap ) , thus supporting a conserved function of SATB2 as transcriptional repressor at both developmental stages . We next compared the two gene expression profiles using as a cutoff for differential expression adjusted p value < 0 . 05 ( Fig 4B ) . We identified only 359 commonly regulated genes between SATB2-deficient and control cortices at adult vs neonatal stage ( OR = 1 . 23 , P = 0 . 0079 ) . Notably , 135 , i . e . approximately one third of the commonly regulated genes , were regulated in the opposite direction in the adult vs neonatal SATB2 mutant cortex , suggesting that a potential shift in the composition of SATB2-containing transcriptional complexes directly or indirectly determines whether transcription of these loci is repressed or activated . We next assessed the biological pathways influenced by SATB2-dependent transcription . GO analysis demonstrated enrichment of processes critical for cell projection morphogenesis and brain development in the genes regulated by SATB2 in the neonatal cortex ( Fig 4C ) . By contrast , SATB2-dependent genes in the adult cortex were enriched in GO terms related to neuronal physiology and synapses including neurotransmitter transport , ion channel complex , calcium ion binding , neuroactive ligand-receptor interaction , calcium signaling pathway ( Fig 4D ) . The GO terms enriched in the two SATB2-deficient transcriptomes are consistent with a differential role of SATB2 as a cell fate and neuron projection determinant at neonatal stage vs regulator of synaptic plasticity/physiology at the adult stage . The human SATB2 locus is highly constrained , rare mutations in the gene cause intellectual disability , and the gene influences cognitive ability in the general population [27–29] . Our transcriptome data showed that SATB2-dependent gene expression programs in the adult mouse cortex were enriched in GO terms associated with synaptic transmission and plasticity that are considered to underlie cognitive functions . We therefore asked if the genes encoding SATB2-containing protein complexes share the characteristics of SATB2 and are associated with common variation in general cognitive ability . We first explored the expression of the human orthologs of mouse SATB2 interactors in human tissues using an expert-curated list of human-mouse homologous genes ( http://www . informatics . jax . org/downloads/reports/index . html ) and the available expression data in human tissues and brain cell types [30 , 31] . The human orthologs of mouse SATB2 interactors were found to be widely expressed in a broad spectrum of human tissues ( S2 Fig ) . They are likely to be available for SATB2 interactions in human pyramidal neurons since they are co-expressed with SATB2 in the human adult cortex and excitatory pyramidal neurons ( S3 Fig , Fig 5 ) . Next , we tested the human SATB2 interactor gene-sets ( adult and neonatal ) for enrichment of highly constrained genes and intellectual disability ( ID ) genes . In the adult SATB2 interactome , 37 of 53 genes ( 69 . 8% ) were loss-of-function ( LoF ) intolerant ( S5 Table ) . In the neonatal SATB2 interactome , 28 of 40 genes ( 70% ) were loss-of-function ( LoF ) intolerant ( S5 Table ) . This identifies a very significant enrichment of highly constrained genes ( adult , P = 5 . 05x10-25 , neonatal , P = 6 . 71x10-21 ) and indicates that both SATB2 interactomes are under strong negative selection . Of the 53 genes in the adult SATB2 interactome , 6 ( 11 . 3% ) are reported to be ID genes . In the neonatal SATB2 interactome , 6 of 40 genes ( 15% ) are ID genes , representing a significant enrichment ( P = 0 . 007 ) [32] . The respective ID genes and the intellectual disabilities reported in OMIM are listed in S5 Table . To study the contribution of SATB2 interactomes to variation in cognitive function within the general human population , we employed data from genome-wide association studies ( GWAS ) of cognitive ability ( CA ) based on 269 , 867 individuals [33] and educational attainment ( EA ) based on 328 , 917 individuals [34] . Using MAGMA [35] to perform gene-set analysis ( GSA ) of the GWAS of CA and EA datasets , we found that the adult SATB2 interactome was significantly enriched for genes associated with CA ( β = 0 . 337 , P = 0 . 012 ) ( S6 Table ) . We also observed of strong tendency for enrichment in the case of genes associated with EA ( β = 0 . 201 , P = 0 . 056 ) . In contrast , common genetic variation in the neonatal SATB2 protein complexes was not significantly associated with common variation in either CA ( β = -0 . 101 , P = 0 . 728 ) or EA ( β = 0 . 161 , P = 0 . 169 ) . Of note , 15 of the total of 91 genes encoding SATB2 interactors are reported to be contributing to CA and or EA based on single SNP and single gene analyses ( S5 Table ) . Brain-expressed genes are a major contributor to cognitive function [33] . It is possible that the enrichment detected here could be due to the SATB2 interactome representing a set of brain-expressed genes . However , the adult SATB2 interactome enrichments were robust to the inclusion in the analysis of both ‘brain-expressed’ ( n = 14 , 243 ) and ‘brain-elevated’ ( n = 1 , 424 ) gene-sets as covariates ( P = 0 . 013 and P = 0 . 010 , respectively; S7 Table ) . The SATB2 interactome is enriched for LoF intolerant genes and such genes are also enriched within genome-wide significant trait-associated loci [36] . To examine if the enrichments we detect for CA is a property of polygenic phenotypes in general , we obtained GWAS summary statistics for five phenotypes and we tested the adult SATB2 interactome for enrichment in each one . These were brain-related diseases ( Alzheimer’s disease and Stroke ) and non-brain-related diseases ( Ulcerative Colitis , Cardiovascular Disease , and Type II Diabetes ) . Notably , the SATB2 interactome was not enriched for any of the five phenotypes tested ( Fig 6; S6 Table ) . As SATB2 is a risk locus for schizophrenia [15 , 16] and genes regulated by SATB2 contribute to schizophrenia [37] , we also tested if the adult SATB2 interactome was enriched for genes associated with schizophrenia and other major neuropsychiatric disorders ( Autism Spectrum Disorder , Attention-deficit/hyperactivity disorder ( ADHD ) , Bipolar Disorder , and Major Depression Disorder ) but no significant enrichments were detected ( S7 Table ) .
Our comparative analyses of SATB2–dependent gene expression programs and SATB2 protein complexes in neonatal vs . adult mouse cortex provide insight into SATB2’s changing physiological brain functions . SATB2-dependent transcriptional responses in postmitotic forebrain neurons shift during brain maturation from regulation of cell fate and morphology to regulation of neurotransmission and plasticity . For SATB2-interacting protein complexes , we observe a complementary shift from transcriptional repression towards organization of higher-order chromatin structure . Analyses of the gene-sets derived from our proteomic experiments in human GWAS data also support a change in the function of SATB2 since only the genes encoding adult but not those encoding neonatal cortex SATB2 protein interactors contribute to human cognitive function . Regarding SATB2 interaction partners , our data show that in both neonatal and adult mouse cortex SATB2 associates with proteins that form functionally interrelated protein networks—some shared , some unique to a single ontogenetic stage . Notably , the observed overlap between the adult and neonatal cortical data sets was found to be surprisingly limited . Even in the case of shared/similar protein complexes , the individual protein components appeared to be exchanged . Of note is the novel and unexpected interaction between SATB2 and the nuclear lamina , represented by LEMD2 , LEMD3 , LAP2 and BAF1 in the neonatal cortex , and Lamin A/C , Lamin B1 , and Lamin B2 in the adult cortex . This finding provides a strong indication that tethering chromatin to nuclear lamina [38] is among the conserved molecular functions of SATB2 . Other SATB2 interactors seem to be unique to the adult cortex , e . g . nuclear pore complex proteins , RNA helicases of the DDX family , some members of HNRNP family , RNA Pol II subunits , transcriptional repressors such as MECP2 and BCLAF1 , the scaffold attachment factors SAFB and SAFA2 . Another key finding of our proteomic analysis is that both adult and neonatal SATB2 interactomes are enriched in components of the molecular machinery involved in the de novo formation of stable chromatin loops [23] , including RNA helicases of the DDX family and HNRNPs . As this type of chromatin-loop stabilizing machinery does not include classical regulators of chromatin architecture , such as cohesin and CCCTC-binding factor ( CTCF ) [39] , our result suggests a role of SATB2-containing protein complexes in CTCF-independent stabilization of long-range chromatin contacts in cortical neurons . Given that regulation of 3D chromosomal conformations is likely to be of high relevance for memory formation and early adult-onset psychiatric diseases such as autism and schizophrenia [24 , 40] , the datasets described here provide valuable information regarding candidate molecular regulators of spatial chromatin configuration linked to normal and impaired cognition . The composition of the protein complexes SATB2 might be determined by expression levels , posttranslational modifications , or by alternative splicing of SATB2 and/or its interactors . We find that many of the adult cortex SATB2 interactors are expressed at lower levels in the adult cortex compared to neonatal cortex . Hence , it appears unlikely that their expression levels play a decisive role for SATB2 interaction . Furthermore , our immunoblotting data revealed indistinguishable molecular mass of SATB2 protein bands in neonatal and adult tissue lysates arguing against different isoforms of SATB2 being expressed at the two ontogenetic stages . Of note , SATB2 expression pattern shifts during brain maturation from superficial to all cortical layers ( [3 , 6] and Allen Developing Mouse Brain Atlas , 2008 ) . Accordingly , SATB2 in adult cortex is expressed in different subpopulations of cortical pyramidal neurons including both upper layer and deep layer excitatory neurons . This change in expression pattern might influence the composition of the available interaction partners . As for the SATB2-dependent transcriptomes , we also discovered functional differences in the biological processes and pathways orchestrated by SATB2 as a transcriptional regulator at the two stages—from morphogenetic and differentiation processes in the neonatal cortex to mechanisms of neurotransmission and plasticity in the adult forebrain . Another difference is that the number of regulated genes is much higher at the neonatal stage compared to the adult . Although there is significant overlap in the gene sets affected by the loss of SATB2 in neonatal and adult brain , as could be expected , the number of commonly regulated gene is surprisingly small . Furthermore , we observed that a substantial number of SATB2-dependent genes were regulated in opposite directions between the two stages . Thus , SATB2 appears to have a modulatory function on the transcription of these genetic loci and by itself does not determine their repression or inhibition . Our GSA provides further support for the divergence in the function of the neuronal SATB2 protein complexes between the two stages . Genes associated with cognitive function are enriched only in the adult but not in the neonatal SATB2 cortical interactome . The enrichment of the adult SATB2 interactome for common variants associated with cognition strongly indicates a role in human intelligence as already demonstrated for SATB2 itself [29 , 33 , 41] . Previous GSA using expert-curated pathways and GO gene-sets have identified dendrite and synapse-related pathways as significantly associated with major neuropsychiatric disorders [42 , 43] and cognitive function [33 , 41] . A connection of the pathophysiology of these diseases to postsynaptic complexes and a contribution to their common genetic architecture have been suggested [44 , 45] . More and more transcription factors and nuclear regulators are also emerging as genome-wide significant genes in the recent meta-analyses of GWAS of psychiatric diseases and human cognition [29 , 33 , 41 , 43 , 46] , demonstrating the importance of the neuronal nucleus in addition to synapses for the pathophysiology of mental disorders . Yet , elucidating the contribution of specific nuclear processes to brain disorders by relying on curated canonical pathways and gene-sets has so far yielded unsatisfactory results . By contrast , the gene-sets in our experiments were derived from unbiased proteomic analyses , grouping genes solely based on their physical interaction with SATB2 . This approach enabled us to discover unexpected associations . Most of the neocortical SATB2 interactors identified in our study are widely expressed across human non-neuronal tissues and also across different adult brain regions ( S2 and S3 Figs ) . Although some of the corresponding genetic loci have been previously linked to neuropsychiatric diseases or cognition ( S5 Table ) , the observed contribution of this combination of genes to general human cognitive ability , but not to other tested phenotypes is an unexpected finding . Now defined as a functional group , they present candidates for further studies on the mechanisms underlying intellectual disability and variability in intelligence . Moreover , by identifying SATB2 interactors as largely overlapping with the chromatin loop proteome [23] , our results support the emerging concept that 3D chromatin architecture is a determinant of human cognitive ability .
Neonatal Satb2 conditional mutants were generated by crossing Satb2flx/flx mice [6] with Nes-Cre transgenic mice [47] on a C57BL/6 background . Satb2flx/flx::Camk2a-Cre mice have been described elsewhere [6] . The following primary antibodies were used: anti-SATB2 ( ab92446 , Abcam ) ; anti-SATB1 ( ab92307 , Abcam ) ; anti-HDAC1 ( 10E2 ) ( 5356P , Cell signaling ) ; anti-HNRNPL ( ab156682 , Abcam ) ; anti-HNRNPL-like ( 4783S , Cell signaling ) ; anti-HNRNPC1/C2 ( M022726 , BOSTER ) ; anti-ZNF638 ( orb215138 , Biorbyt ) ; anti-DHX9 ( PA5-19542 , Thermo Scientific ) ; anti-CUX1 ( M-222 ) ( sc-13024 , Santa Cruz ) ; anti-Lamin A/C ( N-18 ) ( sc-6215 , Santa Cruz ) ; anti-ERK2 ( C-14 ) ( sc-154 , Santa Cruz ) , anti-BANF1 ( PU38143 , a generous gift from T . Haraguchi [48]; anti-pan LAP2 ( Lap2 2–12 ) , a generous gift from R . Foisner [49] , anti-CHMP3 ( HPA015673 , Atlas Antibodies ) . Cortices were dissected from either neonatal or three month-old mice . Tissue was homogenized with a Dounce homogenizer in an IP lysis buffer ( 25 mM Tris/HCl pH 7 . 4 , 150 mM NaCl , 1% NP-40 , 1 mM EDTA , 5% glycerol; Pierce ) . The lysates were incubated for 10 min on ice with shaking , followed by a 15 min centrifugation at 13000 × g , 4°C . The supernatant was used in immunoprecipitation reactions using Dynabeads Protein G Immunoprecipitation Kit ( Thermo Fischer ) according to the manufacturer’s instructions . Briefly , 50 μl of protein G Dynabeads were covalently linked to 5 μg of anti-Satb2 antibody and incubated with cortical lysates overnight at 4°C . For MS sequencing , 1 mg of total protein was used as starting material . On the next day , the beads-antibody-protein complexes were washed 3 times with washing buffer , resuspended in 2 x Roti-Load sample buffer ( Roth ) for elution and incubated at 95°C for 5 min . The eluates were run on 6% SDS-PAGE gels . Proteomic analysis was performed at the Protein Microanalysis Core Facility of Medical University of Innsbruck . Silver/Comassie-stained protein gels were divided into three molecular-weight ranges , cut and subjected to in-gel digestion as published previously [50] . Protein digests were analyzed using an UltiMate 3000 nano-HPLC system ( Dionex ) coupled to an LTQ Orbitrap XL mass spectrometer ( Thermo Fischer ) equipped with a nanospray ionization source . A homemade fritless fused silica microcapillary column ( 75 μm i . d . × 280 μm o . d . ) packed with 10 cm of 3 μm reverse-phase C18 material ( Reprosil ) was used . The gradient ( solvent A: 0 . 1% formic acid; solvent B: 0 . 1% formic acid in 85% acetonitrile ) started at 4% B . The concentration of solvent B was increased linearly from 4% to 50% during 50 min and from 50% to 100% during 5 min . A flow-rate of 250 nl / min was applied . Protein identification was performed via Sequest , Proteome Discoverer ( Version 1 . 3 , Thermo Scientific ) and the NCBInr database ( Mus musculus ) accepting variable modifications carbamidomethyl ( C ) and oxidation ( M ) . Specific cleavage sites for trypsin ( KR ) were selected with two missed cleavage sites allowed . Peptide tolerance was ±10 p . p . m . and MS/MS tolerance was ±0 . 8 Da . The criteria for positive identification of peptides were Xcorr>2 . 3 for doubly charged ions , Xcorr>2 . 8 for triply charged ions , Xcorr>3 . 3 for four-fold and higher charged ions and a FDR of 0 . 01 . For validation of MS data , protein G Dynabeads were coated with 5 μg of anti-SATB2 antibody , the beads were mixed with 500 μg of cortical lysate and incubated overnight at 4°C . The antibody-protein complexes were eluted in 2 x Roti-Load sample buffer , separated by SDS-PAGE and immunoblotted with antibodies against the novel interacting partners . Immunobloting was performed as described previously [51] . Membranes were blocked with 5% milk powder in TBST ( 0 . 1% Tween 20 in TBS ) for 1 h and then incubated overnight at 4°C with the corresponding primary antibodies diluted in blocking solution . After incubation with HRP-coupled secondary antibodies , the blots were developed using ECL reagent ( GE Healthcare ) and imaged with a FUSION-FX7 chemiluminescence detection system ( Vilber Lourmat ) . RNA-seq analysis was carried out as previously described [6] . In brief , RNA was isolated from cortical tissue of 3 month-old Satb2flx/flx and Satb2 cKO mice using Trizol ( Thermo Fisher Scientific ) . Libraries were made according to Illumina standard protocols ( TruSeq , Illumina ) and sequenced as single-end reads on a HiSeq platform according to established procedures . RNA-seq reads were mapped to mouse reference genome ( mm10 ) with STAR aligner [52] . Read counts were obtained using featureCounts [53] and normalized using the normalization algorithms of DESeq2 [54] . Differential gene expression analysis was performed with SARTools package [55] . A threshold cutoff of adjusted ( Benjamini-Hochberg ) p-value <0 . 05 was applied . RNA-seq data from McKenna et al . [17] ( GSE68911 ) were re-analyzed using the same pipeline as for the adult cortex and the same threshold cutoff for differential expression was applied . The gene expression profiles of SATB2-deficient vs wild-type cortices from the two datasets ( neonatal vs adult stage ) were compared by means of a rank-rank hypergeometric overlap ( RRHO ) analysis [26] . RRHO heat maps and rank scatter plot that graphically visualize correlations between two expression profiles were generated at http://systems . crump . ucla . edu/rankrank/ . Protein-protein interaction networks were extracted from the STRING 10 . 5 database ( http://string-db . org/ ) and clustered by using k-means clustering method . “Experiments” , “Databases” , “Text-mining” , “Co-expression” , “Neighborhood” , “Gene Fusion” , and “Co-occurrence” were used as prediction methods with a medium confidence threshold ( 0 . 4 ) . Pathway and process enrichment analysis was carried out by using Metascape bioinformatics tool ( http://metascape . org ) [56] with the following ontology sources: KEGG Pathway , GO Biological Processes , GO Cellular Components , GO Molecular Functions and CORUM . All genes in the genome were used as the enrichment background . Terms with a p-value < 0 . 01 , a minimum count of 3 , and an enrichment factor > 1 . 5 ( the ratio between the observed counts and the counts expected by chance ) were collected and grouped into clusters based on their membership similarities . P-values were calculated based on the accumulative hypergeometric distribution , and q-values were calculated using the Benjamini-Hochberg procedure to account for multiple testing . Kappa scores were used as the similarity metric when performing hierarchical clustering on the enriched terms , and sub-trees with a similarity of > 0 . 3 were considered a cluster . The most statistically significant term within a cluster was chosen to represent the cluster . ConsensusPathDB ( http://cpdb . molgen . mpg . de/ ) was used to perform an overrepresentation analysis of the SATB2 interactomes using the complex-based sets ( i . e . sets of genes whose protein products are members of the same annotated protein complex ) . The p-values were corrected for multiple testing using the FDR method . GWAS summary statistics were sourced for general CA [33] , EA [34] , Autism Spectrum Disorder [57] , Attention-deficit/hyperactivity disorder ( ADHD ) [58] , Bipolar disorder [46] , Major Depression Disorder [43] , Schizophrenia [16] , Stroke [59] , Alzheimer’s disease [60] Ulcerative Colitis [61] , Type II Diabetes [62] , and Cardiovascular Disease [63] . A GSA is a statistical method for simultaneously analyzing multiple genetic markers in order to determine their joint effect . We performed GSA using MAGMA ( http://ctg . cncr . nl/software/magma ) [35] and summary statistics from various GWAS identified above . MAGMA was chosen because it corrects for gene size and gene density ( potential confounders ) and has significantly more power than other GSA tools [64] . An analysis involved three steps . First , in the annotation step we mapped SNPs with available GWAS results on to genes ( GRCh37/hg19 start-stop coordinates +/-20kb ) . Second , in the gene analysis step we computed gene P values for each GWAS dataset . This gene analysis is based on a multiple linear principal components regression model that accounts for linkage disequilibrium ( LD ) between SNPs . The European panel of the 1000 Genomes data was used as a reference panel for LD . Third , a competitive GSA based on the gene P values , also using a regression structure , was used to test if the genes in a gene-set were more strongly associated with either phenotype than other genes in the genome . Sets of ‘brain-expressed’ ( n = 14 , 243 genes ) and ‘brain-elevated’ , i . e . genes that show an elevated expression in brain compared to other tissue types ( n = 1 , 424 ) gene-sets were sourced from the Human Protein Atlas ( https://www . proteinatlas . org/humanproteome/brain ) and used as covariates in a conditional MAGMA GSA . Genes were categorized as LoF intolerant if their probability of being LoF intolerant ( pLI ) metric was ≥ 0 . 9 based on the analysis of exome data for 60 , 706 humans by the Exome Aggregate Consortium [36] . A list of primary ID genes ( n = 1 , 069 ) was sourced from the curated SysID database of ID genes ( http://sysid . cmbi . umcn . nl/ ) [32] . A list of proteins recruited to the stabilized chromatin loops was sourced from [23] . Enrichment analysis of these gene/protein lists with our gene-sets was performed using 2x2 contingency tables with genes restricted to those annotated as protein coding using a background set of 19 , 626 genes ( https://www . ncbi . nlm . nih . gov/ ) . | SATB2 is a homeodomain protein that recruits transcriptional and epigenetic regulators to its DNA binding sites . We have recently shown that deletion of Satb2 from the forebrain of adult mice leads to impaired long-term memory . Human patients with SATB2 haploinsufficiency suffer from severe neurological symptoms including cognitive deficits , developmental delay and absent/limited speech . Here , we tested the hypothesis that SATB2 molecular functions differ between developing and adult cortex . Our results provide experimental support for this concept by demonstrating that 1 ) SATB2 interacts with different protein networks at the two ontogenetic stages , with a switch from transcriptional repression towards organization of chromatin structure and 2 ) SATB2 determines differential transcriptional programs in neonatal vs adult cortex . To explore the contribution of SATB2 interactomes to human cognition , we tested the sets of genes encoding SATB2 interactors for enrichment using the largest available genome-wide association studies ( GWAS ) datasets for cognitive ability and neuropsychiatric diseases . We found that genes encoding SATB2 interactomes are highly constrained . Rare high impact mutations in these genes cause severe cognitive disorders whereas common low impact variants influence general cognitive ability in the population . Given that SATB2 is a known 3D-genome organizer protein , our data emphasize the role of long-range chromatin interactions in human cognition and present novel and unsuspected candidates for further studies on the underpinnings of intelligence and the mechanisms underlying intelligence differences . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Material",
"and",
"methods"
] | [
"genome-wide",
"association",
"studies",
"interaction",
"networks",
"dna-binding",
"proteins",
"neuroscience",
"age",
"groups",
"adults",
"cognition",
"genome",
"analysis",
"epigenetics",
"chromatin",
"genomics",
"chromosome",
"biology",
"proteins",
"gene",
"expression",
... | 2019 | Genes encoding SATB2-interacting proteins in adult cerebral cortex contribute to human cognitive ability |
Reported urban malaria cases are increasing in Latin America , however , evidence of such trend remains insufficient . Here , we propose an integrated approach that allows characterizing malaria transmission at the rural-to-urban interface by combining epidemiological , entomological , and parasite genotyping methods . A descriptive study that combines active ( ACD ) , passive ( PCD ) , and reactive ( RCD ) case detection was performed in urban and peri-urban neighborhoods of Quibdó , Colombia . Heads of households were interviewed and epidemiological surveys were conducted to assess malaria prevalence and identify potential risk factors . Sixteen primary cases , eight by ACD and eight by PCD were recruited for RCD . Using the RCD strategy , prevalence of 1% by microscopy ( 6/604 ) and 9% by quantitative polymerase chain reaction ( qPCR ) ( 52/604 ) were found . A total of 73 houses and 289 volunteers were screened leading to 41 secondary cases , all of them in peri-urban settings ( 14% prevalence ) . Most secondary cases were genetically distinct from primary cases indicating that there were independent occurrences . Plasmodium vivax was the predominant species ( 76 . 3% , 71/93 ) , most of them being asymptomatic ( 46/71 ) . Urban and peri-urban neighborhoods had significant sociodemographic differences . Twenty-four potential breeding sites were identified , all in peri-urban areas . The predominant vectors for 1 , 305 adults were Anopheles nuneztovari ( 56 , 2% ) and An . Darlingi ( 42 , 5% ) . One An . nuneztovari specimen was confirmed naturally infected with P . falciparum by ELISA . This study found no evidence supporting the existence of urban malaria transmission in Quibdó . RCD strategy was more efficient for identifying malaria cases than ACD alone in areas where malaria transmission is variable and unstable . Incorporating parasite genotyping allows discovering hidden patterns of malaria transmission that cannot be detected otherwise . We propose to use the term “focal case” for those primary cases that lead to discovery of secondary but genetically unrelated malaria cases indicating undetected malaria transmission .
Malaria remains a major public health problem that affects 106 countries worldwide mostly in tropical and subtropical regions where ~3 . 4 billion people are at risk of infection and death [1] . Although malaria is mainly transmitted in rural areas where there are suitable environments for Anopheles mosquitoes breeding sites , malaria transmission in urban areas of endemic countries has been increasingly reported over the last three decades [2 , 3] . Unfortunately , the factors driving urban and peri-urban malaria transmission remain poorly characterized . As urban malaria cases are likely to be found at a broader range of primary care/diagnostic facilities , including hospitals and private laboratories failing to report them to the central surveillance system [4–7] , urban malaria control by National Malaria Control Programs ( NMCP ) requires important administrative changes . Furthermore , there are no clear definitions of “urban” , “peri-urban” , and “rural” settings that properly describe the socioeconomic and ecological contexts where malaria transmission occurs . Thus , there is a need for a rigorous and systematic approach to characterize malaria transmission in the rural-to-urban interface that could provide solid information to assess disease risk in such contexts . Like other epidemiological settings , urban malaria transmission is influenced by population movements from rural to urban and peri-urban areas . This rural population influx into urban and peri-urban areas facilitates the introduction of malaria from places where the disease is of high prevalence such as those where illegal mining and logging are common[8] . Furthermore , these underserved populations practice subsistence farming and inhabit poor housing with limited access to health services; such social dynamics favor mosquito breeding in areas considered administratively urban [9] . Here , we propose an integrated approach that aims to characterize epidemiologic and entomologic drivers of “urban” and “peri-urban” malarias in settings that are commonly found in Latin America . Our approach was tested in an endemic area of Colombia . Despite the fact that malaria prevalence is decreasing in Colombia with a 75% reduction in the number of cases since 2000 [1] , the National Surveillance System ( SIVIGILA ) reported an accelerated increase in urban malaria cases from 5 . 9% in 2011 [9] to 30% in 2015 [10] . Although this increase may be explained by population displacement due to political unrest and illegal crops and mining , there is still the possibility of autochthonous urban transmission . Because only a few studies have focused on urban malaria transmission in Colombia , the growing number of reports on urban malaria cases generates concerns and demands to unequivocally confirm the extent of urban and peri-urban transmission and to establish the corresponding control strategies . Thus , our integrated approach was used to study patterns of malaria transmission in five neighborhoods of Quibdó , the capital of the department of Choco ( Colombia ) [9 , 11] . These areas report the greatest number of so called “urban” cases providing an ideal setting for this investigation .
The study protocol was reviewed and approved by the institutional review board of Caucaseco Scientific Research Center ( CECIV , Cali-Colombia ) before initiation . Written informed consent ( IC ) was obtained from each volunteer at enrolment . Parents or legal guardians were asked to consent for children ( <18-year-old ) to participate in the study , and children older than seven years were asked to sign an informed assent if they wanted to participate . Information obtained from the participants was managed on principles of confidentiality . Immediately after blood sample processing , malaria-positive volunteers were informed and assessed during administration of appropriate anti-malarial treatment at the corresponding point of care . Asymptomatic volunteers did not receive treatment in concordance with the Colombian Ministry of Health ( MOH ) malaria treatment guidelines . This study was conducted in Quibdó , which is currently the municipality of Colombia with the highest reported number of malaria cases [10 , 12] . It is in the Department of Chocó , in the northern area of the Pacific coast in the border with Panama . It has an area of 3 , 337 km2 between the jungle of Darien and the Atrato and San Juan river basins[11] . It consists mostly of a dense tropical rain forest with warm weather ( average temperature of 28°C ) , relative humidity of 90% and an annual rainfall of 8 , 000–6 , 000 mm . It has an estimated total population of ~500 , 000 habitants ( 2015 ) , with a geographical dispersion 5 . 4 times higher than the rest of the country [13] . Five sentinel sites ( SS ) were selected based on location , urbanization and history of malaria cases: La Yesquita , Silencio and Roma which are neighborhoods with urban characteristics located in the center of the city have paved streets , public services and no vegetation close to the houses . On the other hand , Casa Blanca and Cabí are classified as peri-urban , located in the North and South ends of the municipality , respectively , with unpaved streets , variable housing infrastructure , lack of a sewage system and abundant vegetation . We have considered an urban area as some groups of buildings and contiguous structures grouped in blocks , which are delimited by streets or avenues , with a number of essential services such as aqueducts , sewage , electrical energy , and hospitals and schools . Capital cities and the remaining municipal administrative headings are urban areas . In contrast , a rural area is characterized by the dispersed disposition of houses and agricultural holdings , and a lack of road structure and public services . A peri-urban area is one that combines characteristics both urban and rural , usually located in areas outside the city ( DANE ( 2005 ) [14] A total of 1mL of blood was collected by venipuncture from every subject , of which ~50 μL were used for thick blood smear ( TBS ) and the remainder was stored in tubes containing EDTA , refrigerated at 4°C and transported by airplane to the laboratory in Cali for later qPCR analysis and microsatellite ( STRs ) genotyping . Samples were handled as potential biohazards and all laboratory staff strictly followed bio-safety standardized procedures . We characterized the urban-to-rural malaria interface by integrating epidemiological and entomological approaches . To determine the prevalence of malaria , we first identified eight malaria positive volunteers among individuals seeking diagnosis at the Ismael Roldán hospital in Quibdó . These volunteers reported to live in urban or peri-urban sites of Quibdó and were further selected as sentinel sites ( SS ) for active case detection ( ACD ) and reactive case detection ( RCD ) . Then , a cross-sectional survey was performed in five SS from Quibdó , three of them considered urban neighborhoods and two peri-urban . A visit to their neighbourhoods of origin was performed to classify them as urban ( 2/8 ) or peri-urban ( 6/8 ) . For RCD , four houses , closest to the primary case were selected in each neighborhood , using the Vector Born Diseases Program ( ETV ) census . Eight primary cases identified by ACD were studied with the RCD strategy as well . Men , women and children above one year of age who lived in the household and were present at the moment of the visit were enrolled . A Household was defined as the place where people enrolled in the study lived , including family members , servants , tenants , and others . Those who agreed to participate answered a symptoms survey and donated a blood sample . Epidemiological questionnaires were answered by the head of the household . A malarial household was defined as one with at least one infected person . The sample size ( n ) was calculated for each SS using an estimated prevalence ( P ) of 2 . 6% with a confidence level of 95% , 2 . 5% error ( d ) according to the following equation n0 = ( zα2 ) , where α = 1 . 96; P ( 1—P ) /d2 . Then , it was adjusted according to the population of each neighborhood ( N ) considering the equation n = n0/ ( 1+ ( n0-1 ) /N ) [15] . Collection of adult specimens: Mosquitoes collection was performed using Human Landing Catches ( HLC ) [22] from 18:00 hours to 6:00 hours in the households of infected volunteers diagnosed by RCD . Collections were carried out simultaneously indoors and outdoors for each house for two consecutive nights . Data on relative humidity and temperature were recorded . Mosquitoes were kept in cups labeled with the date , neighborhood , house code , capture time , mosquitoes quantity and collector's name . Specimens were sacrificed with tri-ethylamine and subsequently individually packaged in 1 , 5mL vials with a perforated lid , and conserved in airtight bags with silica gel . Technicians in charge of mosquito catches signed an informed consent prior HLC . Adult and immature mosquito specimens were determined using dichotonous keys for Anopheles of Colombia [23] . Detection of natural infection with Plasmodium spp: After taxonomic identification , mosquitoes’ head and thorax belonging to the same species and capture hour were pooled . Samples were macerated following the MR4 protocol and insert specifications . Circumsporozoite protein ( CS ) from P . falciparum , P . vivax VK-210 and VK-247 variants were detected by Enzyme-Linked ImmunoSorbent Assay ( ELISA ) using the kit distributed by the Center for Disease Control and Prevention ( CDC , Atlanta , USA ) [24 , 25] Immature collection: We searched SS for open water bodies . The larval habitats search was performed around the households , within a 500-meters radius . Each potential breeding site was georeferenced and characteristics such as size , vegetation type , water type , water body type and water use were recorded . Sampling was carried out using a standard dipper ( 350 mL ) taking ten dips per square meter . Larvae were stored in vials with ethanol for preservation . Each larval container was labeled with date , code , larval number , neighborhood and collector's name . Data were recorded in REDCap ( v . 6 . 9 . 4 ) web application and analyzed using the software R for statistical analysis ( v3 . 3 . 0 ) for variables like age , gender , sociodemographic features , living conditions and malaria positive cases , by site and as a total . Non-parametric tests and Chi-square and Fisher’s exact test were performed to check for differences between categories and calculate associations . A level of statistical significance of 5% was used and 95% confidence intervals were calculated for proportions . The limitations of the study were: A memory bias occurred while conducting the surveys , as for some volunteers it was very difficult to remember exactly the activities carried out weeks before the survey . Volunteers’ displacement was questioned only for the last month . Follow up of some infected cases after diagnosis was difficult because some of them had to leave the city due to their informal jobs” .
A total of 717 volunteers ( 60 . 0% female ) were surveyed using ACD detection in 135 households visited , 58 . 5% of them located in the peri-urban areas and the rest of them in urban areas . Significant differences were found in sociodemographic variables between urban and peri-urban neighborhoods ( Table 1 ) . Median ages were significantly higher in urban neighborhoods . Overall the Afro-American ethnic group was predominant ( 77 . 7% ) , indigenous population was significantly higher in both areas peri-urban areas ( 21 . 4% ) ; especially in Cabí , where 91 . 6% of the population share this ethnic origin . Education level was significantly lower in peri-urban areas , with 25 . 1% of the population being illiterate . Most frequent occupations were merchant in urban areas and housewife in peri-urban areas . Housing conditions were also significantly different ( Table 1 ) . In the urban neighborhoods , most houses were made of brick , 66 . 1% had aqueducts and the majority of them had access to electricity ( 100% ) , garbage collection system ( 89 . 3% ) and sewage system ( 67 . 9% ) . In the peri-urban neighborhoods , the predominant housing material was wood ( 59 . 7% ) , and none of the houses had an aqueduct service or sewage system . Water supply was obtained from rain in 74 of 79 houses while the remaining obtained water directly from the river or had a well . Only three houses in Casa Blanca and one in Cabí had a garbage collection system . Eight primary cases were recruited in a hospital in Quibdó where they attended to seek diagnosis . Four of them were caused by P . vivax , three by P . falciparum and one was a mixed infection . Two cases came from urban sites and the other six from peri-urban neighborhoods . Most of the cases were in Afro-descendants , with complete to incomplete secondary education and with informal jobs . Five of the eight cases were men . The overall prevalence of malaria by ACD was 1% ( 6/604 ) using microscopy and 9% ( 52/604 ) by qPCR . A total of 44 cases ( 85% ) of the 52 detected by qPCR were due to P . vivax being the predominant species . Ninety-six percent of the detected infections were in peri-urban areas , presenting a significant difference with those originated in urban neighborhoods ( p<0 . 0001 ) . Cabí was the neighborhood with the highest prevalence ( 29% ±8 SE ) , followed by Casa Blanca ( 9% ±5 SE qPCR ) . In the urban neighborhoods , only two submicroscopic infections were diagnosed ( Fig 1 ) . Eight primary asymptomatic infections from ACD were selected for RCD , six for P . vivax and two for P . falciparum . During the RCD , 33 houses were studied , 18 in Cabí , 10 in Casa Blanca and five in La Yesquita , and a total of 113 volunteers were surveyed . Twenty-seven secondary cases ( three by TBS and 27 by qPCR ) were identified , all of them from peri-urban areas . The overall percentage of positives among the screened people was 2 . 7% by microscopy and 24% by qPCR . A total of 13 malarial houses were found . In addition , a second round of RCD was performed around the eight symptomatic cases recruited by PCD . A total of 175 volunteers among family members and neighbors were included . Fourteen secondary infections were detected ( eight by TBS and 14 by qPCR ) , representing an infection rate of 4 . 6% by microscopy and 8% by PCR . Twelve cases were caused by P . falciparum and the other two by P . vivax . Twelve of the 14 secondary cases were female , most of them Afro-descendant , four housewives , three students and one with another type of job . Seven were children under eleven years , 7/14 were older than 17 years , one of them 47 years old . One 10-year-old child was an asymptomatic positive case by PCR from an urban site , he did not have recent history of recent displacement outside Quibdó but had moved to other peri-urban neighborhood close to his home . A total of 10 malarial houses were detected in this survey . In total , using the RCD strategy we found 41 malaria cases , 27 by P . vivax ( 66% ) and 14 P . falciparum ( 34% ) and 22 malarial houses . No mixed infections were detected . A high number of asymptomatic volunteers were identified , i . e . 58 of the 93 cases diagnosed by qPCR did not report any symptoms at the time of blood examination and none reported to have had malaria symptoms for the last 15 days . Asymptomatic infection was more frequent with P . vivax ( 46/71 ) . Of the nine volunteers with infection diagnosed by microscopy , seven presented with symptoms . Thirty-five cases showed symptoms; the most common symptoms were fever ( 25/35 ) , headache ( 25/35 ) , chills ( 11/35 ) , muscles pain ( 9/35 ) , malaise ( 6/35 ) , and profuse sweating ( 3/35 ) . Out of 8 primary cases ( 6 P . vivax and 2 P . falciparum ) , 27 secondary cases were detected in Cabí: 21 multiple infections ( >1 allele in at least 1 microsatellite locus ) and only 6 single infections for the set of microsatellite loci used . These multiple infections included 11 ( 52 . 4% ) with 2 alleles in at least 1 locus , 6 ( 28 . 6% ) with at least 2 alleles in 2 loci , and 4 ( 19% ) with > 2 alleles in 3 or more loci ( Fig 2 ) [26] . None of the genotypes found in the primary cases matched those found in the secondary cases . Furthermore , two primary cases were P . falciparum but all the detected secondary cases were P . vivax . This pattern indicates that the secondary cases were not related to the primary cases that allowed their detection . Nevertheless , some genotypes were shared among the secondary cases . Specifically , related genotypes were found in two patients that inhabited the same house ( 4 ) , as well as patients that inhabited houses that were near ( 4 and 7 separated by ~43 mts , 4 and 5 separated by ~13 mts , and 204 and 207 separated by ~18 mts ) ( Fig 2 ) . Entomological studies were performed in 15 houses where at least one malaria case was detected . These houses were distributed in three SS , one urban ( La Yesquita ) and two peri-urban ( Casa Blanca and Cabí ) . Neither mosquitoes nor breeding sites were found in the urban setting . On the contrary , both peri-urban areas registered the presence of immature and adult Anopheles mosquitoes . Natural infection by P . falciparum and P . vivax ( VK-210 and VK-247 variants ) was analyzed in 1 , 305 adult Anopheles mosquitoes; An . nuneztovari ( n = 734 ) , An . darlingi ( n = 555 ) , An . triannulatus ( n = 15 ) and An . apicimacula ( n = 1 ) . One An . nuneztovari obtained in Cabí was found infected with P . falciparum ( 1/590 ) , corresponding to an infection rate of 0 . 17% in this locality . This infected mosquito was captured biting indoors between 22:00h and 23:00h in the Urada indigenous community . A total of 24 open water bodies were examined and georeferenced . From these , four were positive for Anopheles mosquito larvae in Casa Blanca and six in Cabí . Of the ten positive larval habitats , seven were excavation site type , two were stream and one was a puddle . A total of 100 larvae were collected in the breeding sites from these 40 late ( 3rd and 4th ) instar larvae were identified belonging to the two most abundant species . An . nuneztovari was found in excavation sites and An . darlingi in streams and puddles ( Table 2 ) .
No evidence for urban malaria transmission was found in Quibdó . The cases found in urban areas were imported from other cities or peri-urban neighborhoods with high prevalence . Thus , malaria transmission is mainly peri-urban , and autochthonous transmission occurs mainly in indigenous communities . The implementation of RCD with molecular diagnostics and genotyping , allow the detection of hidden malaria transmission clusters . This approach is suitable to better understand the efficacy of the malaria control programs interventions . | Malaria is a disease of rural areas in developing countries . Although a rise in urban malaria cases has been noted during the last decade , this trend could be an artifact due to lack of solid data . In order to better understand “urban” and “peri-urban” malaria , we developed a rigorous and systematic methodology that allows characterizing malaria risk in such settings . Our approach is based on cross-sectional studies using active and reactive case detection strategies , genotyping of parasite isolates in order to better understand transmission patterns , and the local assessment of the entomological factors that allow active transmission in urban and peri-urban neighborhoods . This approach was tested in Quibdó , Colombia . No evidence of malaria transmission in urban areas was found . However , we found solid evidence indicating transmission in peri-urban areas due to Plasmodium vivax ( 86% ) . This was supported by the identification of Anopheles mosquitoes and their breeding places . Our results show that reactive case detection is not only an effective strategy to identify cases in areas where transmission is variable and unstable , but also allows the detection of hidden transmission when combined with genotyping methods . Such patterns are undetected by traditional surveillance methods . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"parasite",
"groups",
"plasmodium",
"neighborhoods",
"tropical",
"diseases",
"social",
"sciences",
"parasitic",
"diseases",
"animals",
"parasitology",
"developmental",
"biology",
"apicomplexa",
"molecular",
"biology"... | 2017 | Characterizing the malaria rural-to-urban transmission interface: The importance of reactive case detection |
Primaquine is the only licensed antimalarial for the radical cure of Plasmodium vivax infections . Many countries , however , do not administer primaquine due to fear of hemolysis in those with glucose-6-phosphate dehydrogenase ( G6PD ) deficiency . In other settings , primaquine is given without G6PD testing , putting patients at risk of hemolysis . New rapid diagnostic tests ( RDTs ) offer the opportunity to screen for G6PD deficiency prior to treatment with primaquine . Here we assessed the cost-effectiveness of using G6PD RDTs on the Thailand-Myanmar border and provide the model as an online tool for use in other settings . Decision tree models for the management of P . vivax malaria evaluated the costs and disability-adjusted life-years ( DALYs ) associated with recurrences and primaquine-induced hemolysis from a health care provider perspective . Screening with G6PD RDTs before primaquine use was compared to ( 1 ) giving chloroquine alone and ( 2 ) giving primaquine without screening . Data were taken from a recent study on the impact of primaquine on P . vivax recurrences and a literature review . Compared to the use of chloroquine alone , the screening strategy had similar costs while averting 0 . 026 and 0 . 024 DALYs per primary infection in males and females respectively . Compared to primaquine administered without screening , the screening strategy provided modest cost savings while averting 0 . 011 and 0 . 004 DALYs in males and females respectively . The probabilistic sensitivity analyses resulted in a greater than 75% certainty that the screening strategy was cost-effective at a willingness to pay threshold of US$500 , which is well below the common benchmark of per capita gross domestic product for Myanmar . In this setting G6PD RDTs could avert DALYs by reducing recurrences and reducing hemolytic risk in G6PD deficient patients at low costs or cost savings . The model results are limited by the paucity of data available in the literature for some parameter values , including the mortality rates for both primaquine-induced hemolysis and P . vivax . The online model provides an opportunity to use different parameter estimates to examine the validity of these findings in other settings .
Plasmodium vivax is an important public health concern , particularly in Asia and South America , where it is now responsible for the majority of malaria cases . While traditionally regarded as a benign disease , P . vivax malaria has been associated with severe and fatal outcomes [1 , 2] . As countries move toward malaria elimination and the overall incidence of malaria declines , the proportion of cases that are due to P . vivax infections increases [3 , 4] . A single infection of P . vivax can lead to multiple relapses due to its ability to form dormant liver stage parasites called hypnozoites . These relapses are indistinguishable from new infections and repeated episodes can lead to a cumulative risk of anemia and malnutrition [5 , 6] . In short latency relapse settings , the majority of P . vivax cases are thought to be due to relapses [7] . Primaquine is the only drug currently licensed for the radical cure of P . vivax; however , it can cause severe hemolysis in individuals with glucose-6-phosphate dehydrogenase ( G6PD ) deficiency , a common genetic disorder [8] that is positively associated with P . vivax incidence [9] . The prevalence of G6PD deficiency varies from less than 1% to more than 30% , with a mean of 8% in countries where malaria is endemic; equivalent to 350 million people worldwide [10] . G6PD deficiency is largely asymptomatic until individuals are exposed to oxidative stress from an external source , including certain drugs , such as primaquine , and foods , most notably fava beans [9] . The degree of enzyme deficiency varies widely depending upon the genotypic variant which varies with geographical region . A recent review found only 14 documented deaths attributable to primaquine use [11]; however , fatalities may have gone unreported [12] . The WHO recommends that primaquine be used for the radical cure of P . vivax infected patients who can be tested for G6PD deficiency [13] . The gold standard for diagnosing G6PD deficiency is the spectrophotometric assay , a test that requires a laboratory setting and specialized staff [14 , 15] . The Fluorescent Spot Test ( FST ) , which is the most widely used assay for G6PD deficiency , is easier to perform but requires basic laboratory equipment , electricity and a cold chain , rendering it difficult to use in remote settings . Thus , routine testing for G6PD deficiency prior to prescribing primaquine generally is not part of antimalarial policy in most countries [16] . Recently , the CareStart G6PD ( Access Bio , Somerset , NJ , USA ) lateral flow rapid diagnostic test ( RDT ) has become available for point of care testing . This phenotypic test has high sensitivity for an enzyme activity cut off of 30% [17 , 18]; hence false negative results would rarely lead to a G6PD deficient individual receiving primaquine with an attendant risk of hemolysis . Unlike the other G6PD RDT by BinaxNOW ( Alere , Orlando , FL , USA ) , the CareStart RDT can be used in settings where the temperature is above 25°C , a common necessity in P . vivax endemic settings[19 , 20] . The availability of point of care G6PD tests is of clinical and public health importance so that P . vivax patients have safe access to primaquine treatment for the prevention of relapses and the resulting health complications [20] . Here we evaluate the cost-effectiveness of using G6PD RDTs on the Thailand-Myanmar border and present our model as an interactive web tool that can be adapted to other settings .
A cost-effectiveness analysis [21] using a health care provider perspective was conducted with decision tree models for P . vivax infections using R statistical software [22] over a 1 year time horizon . The model was parameterized for the north-western border of Thailand with Myanmar ( Tak Province ) , with the benefit of data on recurrences from a recent clinical trial at the Shoklo Malaria Research Unit ( SMRU ) , which provides free of charge care to migrants and refugees [23] ( S1 Appendix ) . In this population of migrants and refugees , the prevalence of G6PD deficiency was documented to be 9–18% [24] . The most common genetic variant was the Mahidol variant ( 88% ) with Chinese-4 , Viangchan , Açores , Seattle , and Mediterranean variants also present [24] . Low , unstable P . vivax transmission is seen in this area [3] with a frequent relapse pattern [25] . In recent years , the overall number of malaria cases has been decreasing while the prevalence of P . vivax in the population has remained relatively stable at 9% [3] . Routine practice along the border is to administer 14 days of supervised therapy with or without G6PD screening to patients able to attend the clinic; in practice this is a small proportion of the patients . The testing of G6PD status with CareStart G6PD RDT before administering primaquine ( “screening strategy” ) was compared to a strategy in which no G6PD test is performed and primaquine is not used at all ( “chloroquine strategy” ) . In addition , the screening strategy is compared with a strategy where primaquine is given to all patients without testing for G6PD deficiency ( “primaquine strategy” ) ( Table 1 ) . The chloroquine strategy and primaquine strategy were not directly compared to each other because ( 1 ) it is unlikely that in settings where the chloroquine strategy is used switching to the primaquine strategy would be a viable option due to the evident concerns about safety and ( 2 ) it is unlikely that settings where the primaquine strategy is used would consider changing to the chloroquine strategy due to its inability to achieve radical cure . Recurrences were recorded over a one year time period in patients who were treated with chloroquine alone , as compared with those who were treated with chloroquine plus 14 days of supervised primaquine for each P . vivax episode ( 0 . 5 milligrams ( mg ) /kilogram ( kg ) /day ) [23] . The relative risk of having at least one recurrence following primaquine treatment was 0 . 22 as compared to those receiving chloroquine alone . For those who had at least one recurrence , the mean number of recurrences was 3 . 54 in the chloroquine arm and 1 . 16 in the primaquine arm . The model applied the inclusion criteria of the clinical trial , which was restricted to patients who were six months and older , not pregnant and presenting with uncomplicated P . vivax malaria [23] . The mean age in the clinical trial was 21 years; this was used for the disability-adjusted life-year ( DALY ) calculations for years of life lost ( S1 Appendix ) . The analysis and results were completed separately for males ( Fig 1 ) and females ( Fig 2 ) to account for their differences in risks and outcomes . Firstly , as pregnant females would not be prescribed primaquine due to the unknown G6PD status of the fetus , the screening and primaquine strategies modeled the inclusion of a pregnancy test for all women of childbearing age who were unaware that they were pregnant . Those who were identified or known to be pregnant would be treated with chloroquine only and would not have a G6PD RDT . Secondly , since G6PD is an X-linked disorder , males who have deficiency are hemizygous while females can be either homozygous or heterozygous with a range of G6PD expression levels . Accordingly , G6PD deficiency was divided into two groups: severe ( <30% enzyme activity ) and intermediate ( 30–69% enzyme activity ) ( S2 Appendix ) . Generally , only females can have intermediate deficiency and the outcomes in this group were taken from heterozygotes . The G6PD RDT with a cut off of 30% activity does not detect heterozygous females with intermediate activity [16 , 18]; accordingly , some women with intermediate deficiency who are identified as G6PD normal with currently available RDTs could be at risk of severe hemolysis when prescribed hemolytic drugs . Table 2 shows the parameters used in the model . Both the screening and primaquine strategies include the cost of supervised therapy in order to reflect the additional costs required for the gains in effectiveness seen in the clinical trial . For the screening strategy , weekly supervised primaquine therapy for 8 weeks was given to those who tested G6PD abnormal [26] and the effectiveness was taken from the trial results for 14 day therapy ( S1 Appendix ) . For each recurrence , the cost and DALY value used were taken from clinical episodes , severe malaria episodes and episodes that resulted in death and weighted proportionally . The probability for severe P . vivax was taken from a meta analysis of clinical studies using those with severe anemia , but other symptoms due to severe P . vivax were not included [2] . The probability of having a hemolytic episode that requires a transfusion in individuals with severe and intermediate G6PD deficiency treated with primaquine was taken from a study of children treated with Dapsone in Africa [27] . While this population may be different in terms of age and G6PD variant from those being treated with primaquine on the northwestern border of Thailand with Myanmar , this was the best available data on transfusion risk . The probability of hemolysis requiring transfusion for severe G6PD deficiency was taken from the proportion of hemizygotes and homozygotes in the study while the probability for females with intermediate deficiency was taken from the proportion of heterozygotes . It was assumed that 10% of patients requiring a transfusion did not receive one; of those , 10% died as a result of not receiving a transfusion . It was assumed that the decision to give a transfusion was made by a physician and that the costs are included in the cost of transfusion . Costs of commodities and service delivery were taken from Myanmar and Thailand and supplemented by international sources when needed . Costs are reported in 2014 United States Dollars ( US$ ) . The cost of supervised therapy was taken from data on annual costs of a community health worker in Myanmar [32] , assuming one half-day of pay per observation . The cost of hospitalization for a blood transfusion was included for severe hemolytic episodes which did not lead to death . Table 1 describes the costs for recurrences . The DALY weights were taken from the 2010 Global Burden of Disease Study [38] . These weights were combined with life tables for Myanmar [36] and assumptions about the length of illness to calculate the DALY burden for each strategy . In instances where the screening strategy averted DALYs while costing more money the incremental cost-effectiveness ratios ( ICER ) was calculated: ICER = Costs– CostbDALYb−DALYs Where Cost is the total cost of the strategy and DALYs is the total DALYs of the corresponding strategy . While the gross domestic product per capita for Myanmar is approximately US$1200 [39] , it has been argued that a lower willingness to pay threshold may be appropriate lower income countries [40]; consequently , a threshold of US$500 was chosen to reflect the resource limitations of healthcare facilities serving migrant and refugee communities . A one-way sensitivity analysis was conducted to examine the impact of parameter values on the overall outcome . Low and high values were taken from 95% confidence intervals ( CIs ) when available . When not available , the point estimate was varied by 50% and given wider intervals when necessary to reflect the uncertainty ( Table 2 ) . Results that varied from the base case by more than US$0 . 05 or 0 . 0002 DALYs averted were reported . A probabilistic sensitivity analysis ( PSA ) was conducted to incorporate the uncertainty of all parameters over 1000 sampling iterations using the parameter ranges used in the one-way sensitivity analysis . Table 2 lists the distributions used in the PSA . The sum of squared differences was minimized from the specified ranges to produce the shape values for the beta and gamma distributions and random numbers were generated from these distributions . The mean number of recurrences for each iteration was calculated from 100 bootstrapped data points that were randomly sampled from the data set with replacement ( S1 Appendix ) . The PSA produced a mean estimate and 95% credible intervals ( CrIs ) for the costs , DALYs and incremental results . A key concern is adherence to primaquine regimens by the patients as well as compliance to guidelines by prescribers , which is collectively referred to as “adherence” here . In order to account for this , a two-way sensitivity analysis examined the interplay of costs and benefits depending on adherence to the primaquine strategy ( whether primaquine was administered to the patient and the full course taken ) and screening strategy ( whether a G6PD RDT plus primaquine was administered to the patient and the full course taken ) . This cohort analysis assumed that at 0% adherence all individuals have a relative risk and mean number of recurrences equivalent to receiving chloroquine only . The proportion of individuals in the population who are adherent increases steadily until 100% adherence , which assumes that recurrences are equivalent to the base case . Costs of supervised primaquine and G6PD screening were also varied accordingly . Assumptions about adherence in individuals with G6PD deficiency who receive 14 day primaquine remain the same as the base case analysis .
Costs and DALYs for each strategy are shown in Table 3 and the cohort results are in Table 4 . On the Thailand-Myanmar border , the screening strategy averted more DALYs than the chloroquine strategy: 0 . 026 for males and 0 . 024 for females . These gains were produced for similar costs . The base case ICERs were US$6 . 3 and US$11 . 7 per DALY averted for males and females respectively . Fig 3 shows the results of the one-way sensitivity analysis in males ( see S3 Appendix for all results ) . The screening strategy always averted more DALYs than the chloroquine strategy ( S3C and S3D Appendix ) . Costs for the screening strategy were highest when radical cure had a low impact on recurrences , when the costs of supervised therapy and the G6PD RDT were increased , and also when the cost of a recurrence was decreased ( S3A and S3B Appendix ) . The only assumptions that made the screening strategy cost over US$500 per DALY averted were lowering the number of recurrences after chloroquine to 1 ( US$3678 . 7 in males and US$3724 . 5 in females ) and assuming the same relative risk of having at least one recurrence in females ( US$1223 . 6 ) . The mean costs and DALYs and CrIs estimated by the PSA are shown in Table 4 . The screening strategy costing more than the chloroquine strategy with mean incremental cost of US$0 . 8 ( 95%CrI: –17 . 4 to 19 . 7 ) and 0 . 026 DALYs averted ( 95%CrI: 0 . 007 to 0 . 117 ) per male ( Fig 4A ) . At a willingness to pay threshold of US$500 , the screening strategy had an 81 . 2% probability of being cost-effective ( Fig 5A ) . The PSA resulted in a mean incremental cost of US$0 . 75 ( 95%CrI: –15 . 0 to 20 . 0 ) with 0 . 023 DALYs averted ( 95%CrI: 0 . 006 to 0 . 122 ) per female ( Fig 4C ) and a 77 . 6% probability of being cost-effective at a willingness to pay threshold of US$500 ( Fig 5C ) . The ICERs were US$31 . 3 per DALY averted for males and US$32 . 4 for females . Again , the screening strategy resulted in better health outcomes in the base case with 0 . 011 DALYs averted in males and 0 . 004 in females ( Table 3 ) . The health gains in females were more modest due to their overall lower probability of hemolysis requiring transfusion . In addition , the screening strategy produced cost savings of US$7 . 1 and US$2 . 2 per male and female initially treated , respectively ( Table 3 ) . The simulation output indicated that one death due to hemolysis would be expected for every 6682 males and 15 , 994 non-pregnant females treated using the primaquine strategy . This would be reduced to one death per 668 , 164 males and 201 , 198 non-pregnant females treated with the screening strategy ( Table 4 ) . The one-way sensitivity analysis showed that changes in the parameter values had a smaller impact on the costs when comparing screening and primaquine strategies , especially for females , and results that consistently averted DALYs ( S3E–S3H Appendix ) . The parameters related to mortality , the need for transfusion and the prevalence of G6PD deficiency having the highest impact on DALY results ( S3G and S3H Appendix ) . The screening strategy was cost saving with the exception of raising the G6PD RDT cost to US$10 . 0 , which caused an incremental cost for the screening strategy of US$235 . 9 and US$1602 . 1 for males and females , respectively ( S3E and S3F Appendix ) . The screening strategy remained cost saving even at low levels of G6PD deficiency ( 7% ) . The PSA showed a mean cost savings of US$7 . 3 ( 95%CrI: -15 . 4 to 3 . 4 ) and 0 . 012 DALYs averted ( 95%CrI: 0 . 001 to 0 . 113 ) in males ( Fig 4B ) , and a mean cost savings of US$2 . 2 ( 95%CrI: –6 . 2 to 6 . 7 ) and 0 . 004 DALYs averted ( 95%CrI: 0 . 000 to 0 . 029 ) in females ( Fig 4D ) . The screening strategy had a 97 . 7% probability of being cost-effective for males at a willingness to pay threshold of US$500 . 0 ( Fig 5B ) . For females , the probability was 91 . 1% ( Fig 5D ) . Table 3 shows the cost and DALY estimates from the PSA . The two-way analysis ( Fig 6 ) demonstrated that the screening strategy would be cost-effective in scenarios where it is used to maintain or increase the number of patients who are adherent to their primaquine regimens . The impact of switching to the screening strategy was slightly less in females due to the exclusion of pregnant women from primaquine treatment and the low sensitivity of the G6PD RDT in women with intermediate G6PD deficiency . Due to the extensive heterogeneity and parameter uncertainty around key parameter estimates , notably relapse patterns [25] , G6PD variants and prevalence [10] and costs , a web-based interface was built using the R-Shiny application so that the model could be adapted to other settings as need be . See website ( https://malaria . shinyapps . io/g6pd_screening/ ) .
Point of care G6PD RDTs offer the opportunity for the safe uptake of primaquine for the prevention of recurrences . Our findings suggest that on the Thailand-Myanmar border the use of G6PD RDTs to identify patients with G6PD deficiency before supervised primaquine is likely to provide significant health benefits ( equivalent to between 1 and 9 days of perfect health ) compared to giving chloroquine alone or giving 14 day primaquine without G6PD testing . Furthermore , the use of point of care G6PD RDTs will potentially save costs or , at most , increase them moderately . Primaquine is currently the only licensed hypnozonticidal drug , but healthcare professionals who treat P . vivax cases are often more concerned with avoiding the immediate risk of hemolysis than with protecting the patient from the risks associated with future relapses . In other settings , primaquine may be administered without G6PD testing , putting individuals with G6PD deficiency at risk of severe hemolysis , although the degree of risk will depend upon local G6PD variants and their prevalence . In the scenario presented , the use of G6PD RDTs will save costs while averting DALYs compared to a policy in which primaquine is administered without G6PD testing . While our results give a high probability of cost savings when switching from the primaquine strategy to the screening strategy , this should not deter radical cure without screening in settings where screening is unavailable as the primaquine strategy averted more DALYs than the chloroquine strategy . Our model is based on supervised primaquine therapy and hence our findings may not be applicable to other settings where unsupervised primaquine is the norm and adherence to a complete course of treatment and thus effectiveness may be low [41] . Our two-way analysis on adherence ( Fig 6 ) enabled comparison between settings with varying adherence and how this impacts upon cost effectiveness . The screening strategy averts more DALYs than the primaquine strategy , even at relatively high primaquine strategy adherence and low screening strategy adherence . Shorter drug courses , such as 7 day primaquine and tafenoquine , should contribute to higher adherence levels and reduced costs for the primaquine and screening strategies . Overall , the screening strategy was less cost-effective in women as compared to men . This reflects a greater proportion of women who are excluded from receiving primaquine due to pregnancy and the lower risk of severe hemolysis in females with intermediate G6PD deficiency . Since men represented 65% of patients in the trial that the recurrence data were derived from ( S1 Appendix ) , the overall cost-effectiveness estimates per person presenting with P . vivax malaria would likely be closer to the results for males . The cost-effectiveness of using G6PD RDTs is also dependent on the diagnostic accuracy of the test . Our model draws on studies conducted on the Thailand-Myanmar border , which demonstrated a high sensitivity in healthy volunteers . While other studies have shown similar results [16 , 17] in healthy volunteers , a recent study in Brazil found that the sensitivity of the CareStart G6PD RDT dropped to 50% in patients with malaria compared to 80% in those who did not [15] . A recent cost-effectiveness analysis of male patients with P . vivax malaria in Brazil used a low sensitivity for the CareStart G6PD RDT ( 46% ) but still found it to be more cost-effective than both the BinaxNOW test and routine care; where the analysis also involved the prescription of primaquine without having a G6PD test [42] . This study , however , used the endpoints ‘adequately diagnosed case’ and ‘hospitalization avoided’ instead of DALYs . The Brazilian population was given a 94% probability of hospitalization when primaquine was given to G6PD deficient men . Our model differs in that we assume a lower rate of hospitalization due to severe hemolysis . We also include results for both genders and report DALYs , enabling comparisons with interventions for other diseases . Our study has a number of limitations , mostly related to our model assumptions . The cost-effectiveness of the screening strategy would be increased if it included the onward transmission of P . vivax or the longer term impact of repeated episodes , such as anemia , malnutrition and all-cause mortality . This is particularly relevant in areas such as the Thailand-Myanmar border where the estimated proportion of recurrences due to relapses is estimated to be 78% . The cost-effectiveness may decrease if some individuals were not able to metabolize primaquine , if healthcare workers were not able to utilize G6PD RDTs or supervise primaquine regimens , if the prevalence of G6PD deficiency in those presenting with P . vivax was lower to that in the general population , if the diagnostic accuracy of the G6PD test were lower , if healthcare facilities providing care for hemolytic episodes were not accessible or if the operational costs of implementing a switch to the screening strategy were included . These parameter limitations are similar to those highlighted in a recent review of the costs and cost-effectiveness of P . vivax control and elimination [43] . Finally , our model is limited by the paucity of data available in the literature for some parameter values , including the mortality rate for those who have a primaquine-induced hemolytic episode requiring transfusion but do not receive them . Our assumptions of primaquine induced mortality were derived from previous risks of mortality in patients treated with Dapsone in Africa and equated to a population risk of 1 in 6 , 682 administrations to males and 1 in 15 , 994 administrations in females . These risks are significantly higher than the risks documented in a previous review [11] but the screening strategy averted more DALYs than the primaquine strategy at a lower level of primaquine-induced mortality , though this is likely due to the utilization of weekly primaquine by the screening strategy . Other variables , such as the prevalence of G6PD deficiency will vary greatly depending on the epidemiological setting . Whilst it would be beneficial to gather more robust parameter estimates on which to base informed policy decisions , this should be tempered by the feasibility of gathering such data and the potential benefits of implementing appropriate policies sooner , especially in the context of elimination . Although our model is relatively simple , it provides a useful starting point for policy makers to compare the risks and benefits of using G6PD RDTs to enable the safe and effective use of primaquine . To assist in this process we provide an online tool with which policy makers and healthcare providers can vary the assumptions made in the model in keeping with local scenarios and as additional data becomes available ( https://malaria . shinyapps . io/g6pd_screening/ ) . As the only licensed antimalarial for the radical cure of P . vivax infections , primaquine will be a critical tool for the elimination of all malaria [44] and for the health gains provided to patients . The currently available G6PD RDTs can identify G6PD deficient males , making the screening strategy an attractive option regardless of current practice . In situations where blood transfusions are not accessible , further information may be required on the prevalence of G6PD deficiency and associated risk of hemolysis in females with intermediate G6PD deficiency who test normal by current G6PD RDT methods [28] . Despite the initial cost , point of care RDTs avert DALYs by reducing recurrences while diminishing the hemolytic risk in G6PD deficient patients . | A single infection with Plasmodium vivax can cause multiple episodes of illness due to dormant liver parasites called hypnozoites . Primaquine is the only drug currently available to treat hypnozoites but is under-used because it can cause life-threatening red blood cell damage in people who have an inherited condition called glucose-6-phosphate dehydrogenase ( G6PD ) deficiency . In other locations , primaquine is given without testing for G6PD deficiency , putting patients at risk of potentially fatal hemolysis . New rapid diagnostic tests provide the opportunity to screen for G6PD deficiency prior to giving patients primaquine . Our study describes a cost-effectiveness analysis conducted using data gathered from the Thailand-Myanmar border . Our results show that screening for G6PD deficiency followed by primaquine treatment provided a few days of disability-free health per patient treated . This was achieved for similar costs as not giving primaquine to anyone or cost savings when compared to giving primaquine without screening . In addition to the health gains provided to patients , the safe use of primaquine will be a critical tool to eliminate malaria . We provide an interactive cost-effectiveness tool online that can be adapted to other locations to examine the potential costs and benefits of using rapid diagnostic tests for G6PD in different scenarios . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"medicine",
"and",
"health",
"sciences",
"parasite",
"groups",
"clinical",
"laboratory",
"sciences",
"cost-effectiveness",
"analysis",
"maternal",
"health",
"plasmodium",
"obstetrics",
"and",
"gynecology",
"economic",
"analysis",
"drugs",
"tropical",
"diseases",
"social",... | 2017 | Using G6PD tests to enable the safe treatment of Plasmodium vivax infections with primaquine on the Thailand-Myanmar border: A cost-effectiveness analysis |
Proteins of the nuclear envelope ( NE ) are associated with a range of inherited disorders , most commonly involving muscular dystrophy and cardiomyopathy , as exemplified by Emery-Dreifuss muscular dystrophy ( EDMD ) . EDMD is both genetically and phenotypically variable , and some evidence of modifier genes has been reported . Six genes have so far been linked to EDMD , four encoding proteins associated with the LINC complex that connects the nucleus to the cytoskeleton . However , 50% of patients have no identifiable mutations in these genes . Using a candidate approach , we have identified putative disease-causing variants in the SUN1 and SUN2 genes , also encoding LINC complex components , in patients with EDMD and related myopathies . Our data also suggest that SUN1 and SUN2 can act as disease modifier genes in individuals with co-segregating mutations in other EDMD genes . Five SUN1/SUN2 variants examined impaired rearward nuclear repositioning in fibroblasts , confirming defective LINC complex function in nuclear-cytoskeletal coupling . Furthermore , myotubes from a patient carrying compound heterozygous SUN1 mutations displayed gross defects in myonuclear organization . This was accompanied by loss of recruitment of centrosomal marker , pericentrin , to the NE and impaired microtubule nucleation at the NE , events that are required for correct myonuclear arrangement . These defects were recapitulated in C2C12 myotubes expressing exogenous SUN1 variants , demonstrating a direct link between SUN1 mutation and impairment of nuclear-microtubule coupling and myonuclear positioning . Our findings strongly support an important role for SUN1 and SUN2 in muscle disease pathogenesis and support the hypothesis that defects in the LINC complex contribute to disease pathology through disruption of nuclear-microtubule association , resulting in defective myonuclear positioning .
The nuclear envelope ( NE ) is composed of the nuclear membranes , nuclear lamina and nuclear pore complexes and encloses the chromatin in eukaryotic cells . Lamin intermediate filament proteins are the major structural components of the NE and polymerize to form a fibrous meshwork that underlies the nucleoplasmic face of the inner nuclear membrane . This nuclear lamina is attached to the inner nuclear membrane through interactions with multiple integral inner nuclear membrane ( INM ) proteins [1] . Together , these proteins form a structural network that plays a vital role in supporting the NE and maintaining nuclear integrity , whilst also contributing to chromatin organization and regulation of gene expression ( reviewed in [2] ) . Mutations in genes encoding NE proteins are associated with a range of tissue-restricted inherited disorders that can affect striated muscle , bone , fat or neurons and in some cases cause premature ageing syndromes [3] . Most strikingly , different mutations in one gene – the LMNA gene that encodes A-type nuclear lamins – can cause many diseases , which have collectively been termed laminopathies [4] . Diseases affecting striated muscle are the most common of the laminopathies and include autosomal dominant and recessive Emery-Dreifuss muscular dystrophy ( EDMD2 and EDMD3 , respectively; OMIM#181350 ) , limb-girdle muscular dystrophy ( LGMD ) type 2B and dilated cardiomyopathy and conduction system disease ( CMD ) type 1A [5]–[8] . These diseases share the common feature of cardiomyopathy , but EDMD and LGMD also involve progressive muscle wasting and weakness . In all cases , premature sudden death can result from cardiac arrhythmia and conduction defects . Striated muscle disease , in particular EDMD , can also be caused by mutations in genes encoding other NE proteins . An X-linked form of EDMD ( EDMD1; OMIM#310300 ) is caused by mutations in EMD , that encodes the integral INM protein emerin [9] . Together , mutations in LMNA and EMD account for around 40% of cases of EDMD [10] . Rare mutations in the genes encoding FHL1 , TMEM43 ( also named LUMA ) , nesprin-1 and nesprin-2 have also been reported [11]–[13] . Interestingly , A-type lamins , nesprins and emerin all interact with each other [14]–[16] , contributing to a network that connects the nuclear lamina to the cytoskeleton , termed the LINC ( Linker of Nucleoskeleton and Cytoskeleton ) complex [17] . Furthermore , interactions are often perturbed by muscle disease-causing mutations , indicating that this network of interactions plays an important role in muscle function [12] , [18] , [19] . The central components of the LINC complex in mammals are SUN and nesprin proteins that reside in the INM and outer nuclear membrane ( ONM ) , respectively . The conserved SUN and KASH domains of the respective proteins interact in the perinuclear space to form a bridge spanning the INM , perinuclear space and ONM that connects the nuclear lamina to the cytoskeleton . The nucleoplasmic N-termini of the SUN proteins , SUN1 and SUN2 , interact with the nuclear lamina , anchoring the LINC complex at the NE [20]–[22] . In turn , the cytoplasmic domains of the nesprins connect to the cytoskeleton . There are 4 nesprin isoforms encoded by different genes . Giant isoforms of nesprins-1 and -2 contain an N-terminal calponin homology domain responsible for actin binding [23] , [24] and linkage to the centrosome through microtubules and their motor proteins [25] . Nesprin-3 connects to the cytoplasmic intermediate filament network through interaction with plectin [26] , whilst nesprin-4 is specific to epithelial cells and connects the NE to microtubules via the kinesin-1 motor protein [27] . There are several proposed mechanisms to explain the tissue specificity of EDMD and other laminopathies , which centre around the “gene expression” and “structural” hypotheses [28] . Current evidence strongly supports the “structural hypothesis” , which suggests that muscle-associated laminopathies primarily result from weakening of the structural networks of the nuclear lamina and cytoskeleton and the LINC complex that connects these two networks [29] . Since myocytes are subject to recurrent mechanical strain from contractile forces , weakening of these structural networks renders the cells particularly susceptible to damage . However , the LINC complex is also vital for correct myonuclear positioning [30]–[33] and defects in this process are implicated in impaired muscle function [34] , [35] . Despite the genetic studies so far carried out , causative mutations have been identified in only approximately 50% of EDMD and related muscle disease cases [10] . It is therefore highly likely that mutations in additional genes contribute to the disease . Furthermore , there is significant heterogeneity in disease severity even within families carrying the same gene mutation [36]–[40] , which has led to the suggestion of modifier genes [41]–[43] . Given that SUN1 and SUN2 interact with at least four of the known muscle disease-associated NE proteins and that these interactions can be perturbed by disease-causing LMNA and EMD mutations [44] , we investigated whether the SUN1 and SUN2 genes may also be mutated in some individuals . Screening of the SUN1 and SUN2 genes in a large cohort of patients with EDMD and phenotypically related myopathies identified SUN1 and/or SUN2 variants in several patients . Presence of SUN1 or SUN2 variants correlated with increased disease severity in patients with EDMD carrying mutations in other genes , thus identifying SUN1 and SUN2 as modifiers of the EDMD disease phenotype . We further provide evidence that these mutations disrupt nuclear-cytoskeletal connection and nuclear positioning , supporting the hypothesis that muscular dystrophies arise from defective nuclear-cytoskeletal coupling .
We analyzed DNA from 175 unrelated patients with EDMD and related myopathies , who had previously undergone screening for mutations in the LMNA , EMD , SYNE1/SYNE2 alpha and beta ( encoding short isoforms of nesprin-1 and nesprin-2 , respectively ) and FHL1 genes and in whom no causative mutation had been found . These included both sporadic cases and index patients from familial cases . Furthermore , there have been several reports of modifiers of the phenotype of LMNA-linked muscle diseases [41]–[43] . We therefore also screened EDMD patients carrying identified LMNA , SYNE1/SYNE2 alpha and beta and EMD mutations to determine whether mutation of SUN1 or SUN2 may influence disease phenotype . Most individuals were of Caucasian origin , except where otherwise stated . The 23 exons of the SUN1 gene ( see Figure S1 ) and 19 exons of the SUN2 gene ( ENSG00000100242 , ENST00000405510 ) , including intron/exon boundaries and promoter regions were analyzed . DNA was amplified using PCR and analyzed by direct Sanger sequencing . In total , we found 34 single nucleotide polymorphisms within the coding regions of SUN1 and SUN2 , 18 of which were classified as rare , non-synonymous changes following analysis of their frequencies in sequenced genome databases ( Table S1 , Figure S3 ) . Three of these variants did not segregate with disease in the respective families ( Figure S2 ) . In nine unrelated families or sporadic cases , however , we identified 10 rare non-synonymous variants in SUN1 and SUN2 for which we have obtained evidence of pathogenic effects , as deduced from genetic , phenotypic and/or functional data ( Figure 1A ) . We identified 5 rare , non-synonymous SUN1 and/or SUN2 variants in 3 individuals who lacked mutations in other genes but had EDMD or related myopathy phenotypes ( Table 1 ) . Sporadic patient MD-11 carried a single SUN2 p . R620C sequence variation . We had no access to DNA from family members for this patient , but the high degree of evolutionary conservation of R620 is supportive of disease-association ( Figure S3 ) . Patient MD-1 carried compound heterozygous SUN1 p . G68D and p . G338S variants . For patient MD-1 we had access to DNA from family members and observed apparent recessive inheritance , with one mutation coming from each of the unaffected parents ( Figure 1B ) . SUN1 p . G338S was also present in the reference population at low frequency ( see Table S1 ) . These residues are located within the poorly conserved N-terminal domain of the protein and , in this context , are moderately conserved ( Figure S3 ) , but functional data presented below present compelling evidence of the involvement of both mutations in disease causation . Sporadic patient MD-12 carried heterozygous changes in both SUN1 and SUN2 , encoding SUN1 p . W377C and SUN2 p . E438D , respectively . E438 is conserved in mammals , whilst W377 is conserved across all species examined ( Figure S3 ) . Because patients with EDMD-like phenotypes exhibit variable disease severity that could be explained by mutations or polymorphisms in additional genes , we screened for SUN1 and SUN2 variants in patients with known mutations in causative genes . SUN1 or SUN2 variants were indeed present in some patients from families with LMNA or X-linked EMD mutations ( Table 2; Figure 1C ) . These sequence changes correlated with increased disease severity . In one example , a SUN1 p . A203V polymorphism in patient MD-3 co-segregated with a previously reported EMD p . L84Pfs*6 mutation in two brothers with unusually severe EDMD ( Figure 1C , Family 3 ) [39] . The EMD p . L84Pfs*6 mutation , which abolishes emerin expression , has been reported in an unrelated family , where the course of the disease was significantly milder EDMD with later age of onset and no loss of ambulation [45] . Another unrelated patient carrying EMD p . L84Pfs*6 was included in this study but no SUN1 or SUN2 variants were found and their phenotype was similar to that described by Manilal et al . [45] . In another case , the SUN1 p . G76A mutation , when combined with EMD p . A56Pfs*9 in a previously described Korean patient ( MD-4 ) , led to a very severe clinical picture with complete atrioventricular block requiring pace maker implantation at age 14 years [46] . Similarly , SUN1 p . W377C was detected in combination with LMNA p . R453W in patient MD-5 . This individual had severe disease and died early at the age of 34 years from heart failure . The patient's son , carrying LMNA p . R453W only , did not show clinical signs of contractures or muscular weakness at age 10 years . LMNA p . R453W is a common EDMD-associated LMNA mutation and is generally not associated with severe cardiac disease , suggesting that , in patient MD-5 , SUN1 p . W377C had a modifying effect to increase disease severity [47] , [48] . The same SUN1 p . W377C variant was detected in patient MD-12 , who had an EDMD-like phenotype but did not have mutations in EMD or LMNA but carried a concurrent SUN2 p . E438D variant , as described above . We detected SUN2 mutations in combination with LMNA mutations in two additional index cases . Patient MD-6 carried LMNA p . T528K and SUN2 p . A56P , whilst patient MD-7 carried LMNA p . R98P and SUN2 p . V378I ( Table 2; Figure 1C ) . In both cases , the LMNA mutation had arisen de novo , whilst the SUN2 mutation was inherited from an unaffected parent . Again , the disease expression in both index patients was more severe than is typical for EDMD [49] , [50] , with early onset at age 1 and 4 years , respectively , and early heart involvement including heart transplantation before age 20 years ( Table 2 ) . We also identified two SUN2 variants , encoding variants p . M50T and pV378I , in patient MD-2 who had hypertrophic cardiomyopathy and also carried a mutation in MYBPC3 ( p . G148R ) , which encodes a myosin binding protein . The same MYBPC3 mutation was previously reported in a Dutch family , where a severely affected index patient had compound heterozygous mutations in MYCBP3 but other family members carrying only p . G148R were either asymptomatic or developed cardiomyopathy late in life [51] . This MYBPC3 mutation was present in both patient MD-2 and his father ( Figure 1C; S . Waldmueller , personal communication ) . Patient MD-2 presented as a 6 month-old boy with hypertrophic cardiomyopathy and died at 16 years from heart failure . His father , who carries the SUN2 p . M50T variant but not p . V378I , is asymptomatic . Thus , the SUN2 p . V378I variant appears to have a dramatic effect on disease severity . Notably , this variant is present in the reference population at low frequency ( see Table S1 ) , suggesting that it may be a relatively common genetic modifier of inherited cardiomyopathy . Our genetic results suggest that mutations or polymorphisms in SUN1 and SUN2 may cause muscular dystrophy and act as modifiers of EDMD and cardiomyopathy . To obtain additional evidence that these variants play a role in pathophysiology , we examined the effects of several variants on a known function of the LINC complex , namely centrosome orientation and nuclear movement in migrating cells . SUN2 , along with the nesprin-2G isoform , assembles into transmembrane actin-associated nuclear ( TAN ) lines that couple actin cables to the nucleus to move it rearward and reorient the centrosome toward the leading edge in migrating NIH3T3 fibroblasts [52] , [53] . While SUN1 is not in TAN lines , it also functions in connecting the nucleus to the cytoskeleton via the LINC complex . We expressed three myc-SUN1 and three myc-SUN2 variants in NIH3T3 fibroblasts at the edge of a wounded monolayer by DNA microinjection and stimulated nuclear movement and centrosome reorientation with the serum factor , lysophosphatidic acid ( LPA ) . Upon stimulation , non-expressing NIH3T3 cells or NIH3T3 cells exogenously expressing wild-type ( WT ) SUN1 or SUN2 , as well as the variant SUN2 M50T , reoriented their centrosomes ( Figure 2A–B ) . Notably , SUN2 p . M50T did not appear to influence disease severity in family 2 . In contrast , cells expressing the putative disease-causing SUN1 variants , G68D , G338S or W377C , inhibited centrosome reorientation by blocking rearward positioning of the nucleus ( Figure 2A–B ) . Similarly , cells expressing SUN2 A56P or R620C failed to reorient their centrosome due to an inability to position their nuclei rearward of the cell centroid ( Figure 2A–B ) . All of the expressed SUN1 and SUN2 variants had a normal nuclear localization similar to the wild-type proteins ( Figure 2A ) . The centrosome orientation defect in cells expressing the SUN variants occurred due to defective rearward nuclear movement and not disruption of positioning of the centrosome at the cell centroid ( Figure 2C ) . Hence , five putative disease-causing or disease-modifying SUN variants blocked rearward nuclear movement in migrating NIH3T3 fibroblasts and it can be concluded that these mutants disrupt LINC complex function . Having established that some of the variants identified in our patient cohort disrupted LINC complex function in fibroblasts , we next wished to examine their role in muscle cells . We were able to obtain primary myoblasts from patient MD-1 , carrying compound heterozygous SUN1 p . G68D/p . G338S variants . To gain initial insight into the cellular effects of these mutations , we examined expression of LINC complex components and known SUN1 binding partners in the myoblasts by immunofluorescence microscopy . Since nesprin-1 is not significantly expressed in myoblasts [54] , [55] ( Figure S4 ) , we stained for nesprin-2 , lamin A/C , emerin , SUN1 and SUN2 . No obvious defects in localization of SUN1 , SUN2 or their interacting NE partners were observed , but expression of SUN1 and nesprin-2 at the NE was enhanced in the patient myoblasts ( Figure 3A–B ) . Quantification of fluorescence intensity suggested that their expression was increased approximately 2-fold . Since fluorescence intensity does not always provide an accurate reflection of total protein levels , we then examined total protein expression level by Western blot and found that SUN1 levels were elevated 8-fold in the patient versus control myoblasts ( Figure 3C , E ) . Consistent with the fact that SUN proteins form complexes with , and are responsible for anchoring of nesprins in the ONM , we also observed a 4-fold increase in expression of the intermediate-sized muscle-enriched isoforms of nesprin-2 [55] in the patient myoblasts ( Figure 3D–E ) . However , it remains possible that the bands observed are degradation products of nesprin-2 giant . In contrast , levels of SUN2 , lamin A/C and emerin were not significantly altered in the patient cells ( Figure 3C , F ) . To exclude the possibility that the observed changes were due to different levels of background differentiation in the control and patient cultures , we quantified expression of myogenin , an early marker of myogenic differentiation , and found no detectable level in either culture ( data not shown ) . To address the mechanism of SUN1 elevation in the patient myoblasts , we examined SUN1 mRNA levels by qPCR and found no significant increase in mRNA level compared to the control , indicating that the p . G68D/p . G338S SUN1 variants do not lead to increased mRNA levels ( Figure S4 ) . However , in analyzing mRNA levels of other LINC complex-associated proteins , we observed a statistically significant increase in expression of LMNA , SUN2 , SYNE1 and SYNE2 . In contrast , EMD ( encoding emerin ) expression was decreased in the patient relative to control myoblasts ( Figure S4 ) . One hypothesis to explain NE-associated muscle diseases is the “structural hypothesis” , which suggests that muscle damage occurs due to weakening of the protein interaction network supporting the NE [28] , [56] . To determine whether muscle disease-associated SUN1 alterations disrupt interactions with other LINC complex components , we performed immunoprecipitation with anti-SUN1 antibodies on protein extracts from control and patient MD-1 myoblasts and detected co-precipitating proteins . We observed a reproducible reduction in SUN1 interaction with emerin in the patient myoblasts , whilst interaction with lamin A/C was not obviously perturbed ( Figure 4A–B ) . Consistent with the enhanced recruitment of nesprin-2 to the NE , interaction of SUN1 with nesprin-2 was also maintained in the MD-1 myoblasts ( Figure 4C ) . Larger isoforms of nesprin-2 were enriched in the immunoprecipitate compared to the initial lysates , indicating a preferential interaction of SUN1 with these less abundant isoforms . Thus , the SUN1 p . G68D/p . G338S variants in patient MD-1 appear to specifically disrupt interaction with emerin . We confirmed defective interactions with emerin in HEK293 cells transiently expressing GFP-emerin and myc-SUN1 constructs . SUN1 was immunoprecipitated using anti-myc antibodies and co-precipitating GFP-emerin detected by Western blotting . We observed a significant reduction in interaction of emerin with both SUN1 G338S and W377C , whilst there was only a modest decrease in interaction with SUN1 G68D ( Figure 4D–E ) . This correlates with the close proximity of G338 and W377 to the emerin binding site on SUN1 ( see Figure 1A; [44] ) . Abnormalities in nuclear morphology and a honeycombed pattern of protein expression at the NE are commonly observed in EDMD fibroblasts obtained from patients with LMNA or EMD mutations [57]–[59] . We did not observe any obvious defects in nuclear morphology or protein localization in myoblasts from patient MD-1 ( Figure 3A ) or in cells transiently expressing a range of SUN1 or SUN2 mutants ( for example , see Figure 2 ) . However , we sought to determine whether SUN1 mutants have a modifying effect to increase the severity of nuclear defects when expressed in combination with mutations in LMNA or EMD . To achieve this , we expressed SUN1 W377C , found in combination with a LMNA R453W mutation in patient MD-5 , in fibroblasts obtained from an EDMD2 patient carrying LMNA R401C . Cells expressing SUN1 W377C had a 4-fold increase in the level of nuclear dysmorphology , in terms of increase in nuclear blebbing and formation of honeycomb structures ( Figure 4F , arrows ) , compared to those expressing WT SUN1 ( Figure 4F–G ) . Cells expressing WT SUN1 had a 2-fold reduction in nuclear abnormalities compared to untransfected cells , suggesting a protective effect of SUN1 over-expression . Exogenous expression of EDMD2-associated lamin A/C variants and disruption of endogenous lamin A/C , emerin or the LINC complex all affect myoblast differentiation [60]–[63] . Furthermore , studies in Drosophila indicate that the LINC complex is required for correct myonuclear spacing [34] . We therefore determined whether the myoblasts carrying SUN1 variants had defects in their ability to differentiate to form mature , multinuclear myotubes . In order to account for inherent differences in differentiation capacity between cell lines , we compared MD-1 cultures with a total of 5 control cultures independently established from different individuals . SUN1 staining was clearly evident at the nuclear envelope of MD-1 myotubes , as previously observed in myoblasts ( Figure 5A ) . However , we found that the MD-1 myoblasts exhibited an increased rate of differentiation , both in terms of the number of myotubes and the number of nuclei per myotube . There was a striking increase in the percentage of nuclei within myotubes ( defined as muscle-specific caveolin3-positive cells containing at least 3 nuclei ) in patient cultures ( 28±6% versus 58±9% in control and MD-1 cultures , respectively; mean ± sem ) and a 5-fold increase in myotubes possessing more than 10 nuclei ( Figure 5A–E ) . Strikingly , 45% of these myotubes displayed gross nuclear misalignment and clustering , with up to 30 nuclei per myotube in some instances ( Figure 5D , F ) . Together with our earlier observation of defective nuclear repositioning in fibroblasts expressing disease-associated variants and in double-mutant EDMD2 fibroblasts , these findings suggested that mutations in SUN1 and SUN2 disrupt connections with the cytoskeleton , thereby perturbing nuclear anchorage . Studies have previously shown that SUN1 and SUN2 become concentrated at the poles of the nucleus during human primary myoblast differentiation and that this polarization is linked to correct myonuclear spacing [32] . We therefore examined SUN1 and SUN2 polarization in patient MD-1 myotube nuclei and found that , whilst SUN1 polarization was normal ( Figure S5 ) , SUN2 failed to polarize in clustered nuclei ( Figure 5D , G ) . In keeping with earlier observations in myoblasts , we also found that nesprin-2 fluorescence intensity was significantly increased in all patient myotubes ( Figure S6 ) . Ultrastructural analysis showed that myonuclear clustering occurred within single MD-1 myotubes ( Figure 5H ) and further demonstrated that enlarged highly differentiated myotubes with misaligned nuclei were devoid of detectable sarcomeric structures that were visible in controls ( Figure 5H , arrowheads ) . During myotube formation , the microtubule network is reorganized into a parallel array along the longitudinal axis of the myotube and is nucleated from the nuclear surface , which becomes the primary microtubule organizing centre ( MTOC ) of the cell [64] . As part of this process , centrosomes undergo partial disassembly and centrosomal proteins , including γ-tubulin , pericentrin and PCM-1 , become concentrated at the nuclear periphery [65] , [66] . We hypothesized that SUN proteins contribute to centrosomal protein recruitment to the NE and that polarization of SUN proteins at the nuclear poles may promote linear nuclear organization in myotubes . We therefore investigated the recruitment of centrosomal proteins to the NE during myogenesis , using pericentrin as a marker . In control myotubes we found that pericentrin was recruited to the NE and there was a suggestion that it concentrated at the poles of the nuclei , in a similar manner to SUN2 ( Figure 6A ) . In contrast , pericentrin failed to accumulate to any significant degree at the nuclear surface in MD-1 myotubes and was instead found in cytoplasmic foci . These findings support our hypothesis that SUN proteins are involved in the recruitment of pericentrin to the NE in myotubes . We next investigated whether microtubule nucleation from the NE was disrupted in the patient myotubes by observing microtubule regrowth following nocodazole-induced depolymerization . In control cells , microtubules could be clearly observed emanating from around the nuclear surface after nocodazole wash-out for 30 minutes ( Figure S7 ) . In contrast , the microtubule network in patient cells was very disorganized and often did not appear to be attached to the NE , suggesting a defect in microtubule nucleation , anchoring or organization . To investigate this further , we performed a short 5-minute nocodazole wash-out to detect sites of microtubule nucleation . Microtubule asters regrowing from the nuclear envelope were observed in control myotubes ( Figure 6B , panel a ) and committed myoblasts ( Figure 6B , panel b ) . Microtubule nucleation correlated with sites of pericentrin concentration at the nuclear poles ( arrows in Figure 6B ) . In MD-1 myotubes and committed myoblasts , microtubules were seen to nucleate mainly from multiple sites in the cytoplasm , corresponding with the locations of cytoplasmic pericentrin foci ( Figure 6B , arrowheads ) . Counting fifty myotubes per sample , we could demonstrate that the mean number of microtubules nucleating from myotube nuclei was significantly reduced in MD-1 ( Figure 6D ) . These data suggested that centrosome attachment to the nucleus may also be disrupted in myoblasts from this patient . We therefore examined nuclear-centrosomal distance in MD-1 myoblasts and indeed observed a 2-fold increase in separation between the nucleus and centrosomes in the patient myoblasts compared to controls ( mean distance 4 . 34 µm versus 2 . 07 µm , respectively ) ( Figure 6C , E ) . To confirm that loss of pericentrin recruitment to the NE was a direct consequence of SUN1 mutation , we observed pericentrin localization in C2C12 myotubes following transient transfection with myc-SUN1 variants . Pericentrin was absent from the NE of myonuclei expressing SUN1 G68D , G338S and W377C variants , but exhibited clear nuclear rim staining in myonuclei expressing WT SUN1 ( Figure 7 A–B ) . Thus , all 3 mutants tested acted in a dominant manner in C2C12 cells to displace pericentrin from the NE . Finally , to directly link the SUN1 mutants to the nuclear clustering phenotype observed myotubes of patient MD-1 , we assessed the degree of nuclear clustering in myotubes expressing WT SUN1 and the G68D , G338S and W377C variants . Whilst 22% of myotubes expressing WT SUN1 displayed myonuclear clustering , this value was increased by 2-fold in the cultures expressing each of the 3 SUN1 variants ( Figure 7C ) . These findings confirm that the SUN1 p . G68D and p . G338S mutations are the primary cause of the failure to recruit pericentrin to the NE and the defective nuclear positioning in myotubes from patient MD-1 and further indicate that this is likely to be common to muscle from patients carrying other SUN1 mutations , including p . W377C . In summary , our data demonstrate that muscle disease-associated alterations in SUN proteins result in loss of nuclear connectivity to the cytoskeleton . In myotubes , SUN1 mutations disrupt connections with centrosomal components and the microtubule network , in particular impairing microtubule organization and nucleation from the NE . This in turn is likely to lead to impaired myonuclear positioning in multinuclear myotubes , which we propose may be an important contributor to muscle dysfunction .
Our data add to an increasingly complex picture concerning the genetics of EDMD and related myopathies , where multiple genes , either alone or in combination , can cause or modify the disease phenotype . The SUN1 and SUN2 variants appear to be inherited in highly variable manners , with or without the presence of a mutation in a second gene . In 2 families ( families 1 and 2 ) , SUN1 or SUN2 variants were inherited from each of the unaffected parents of the index patients , strongly supporting an autosomal recessive mode of inheritance in those families . One sporadic case carried heterozygous mutations in both the SUN1 and SUN2 genes , suggesting that mutations in the 2 genes could have additive effects , as has been observed in Sun1/Sun2 knockout mice [30] . In other instances , the index case carried only one SUN1 or SUN2 variant and often represented a sporadic case , suggestive of either a dominant de novo mutation or presence of a concurrent mutation in another , as yet unidentified , gene . We also identified SUN1 or SUN2 variants in individuals from 4 families harbouring known LMNA or EMD mutations . In all cases , the SUN1/SUN2 mutation alone did not cause disease in other family members . However , disease severity was significantly increased in the individuals carrying both mutations compared to family members , or unrelated individuals , carrying only the LMNA or EMD mutation . Furthermore , their phenotype was on the severe end of the spectrum , as defined by Yates et al . [50] . These findings suggest that some SUN1/SUN2 variants act as modifiers to increase disease severity . There has been much speculation as to the existence of modifier genes in EDMD due to high variability in disease phenotype between affected individuals within families [37] , [38] , [41] , [47] and there is now some evidence to support this . In a family with X-linked EDMD caused by an EMD p . Y105X mutation , disease severity was increased in one individual due to a second mutation present in the LMNA gene [42] . Similarly , an individual with severe disease and carrying a LMNA p . R644C mutation was found to carry a second mutation in the gene encoding desmin [42] . In our cohort , for some cases ( such as patient MD-3 ) the increased severity was expressed as clinically more severe muscular dystrophy [39] . In other cases , the additional presence of a SUN1/SUN2 mutation was associated with more severe cardiac disease . For example , patient MD-5 from family 5 , carrying both LMNA p . R453W and SUN1 p . W377C mutations , developed cardiac disturbances at age 25 and died from heart failure at age 34 , which is much earlier than is typical for EDMD patients [50] . Furthermore , LMNA p . R453W is a relatively common mutation that has been reported in at least 15 individuals with EDMD and is usually associated with mild disease [8] , [48] , [67]–[70] . Thus , it is likely that the SUN1 mutation carried by patient MD-5 contributed to their increased disease severity . We also identified SUN1 p . W377C in combination with SUN2 p . E438D in a sporadic case , supporting the idea that a mutation in a second gene is required for disease causation in this instance . Further support for a modifying role for the p . W377C mutation came from the ability of this mutant to worsen nuclear dysmorphology when expressed in LMNA R401C patient fibroblasts . Individuals MD-6 and MD-7 ( carrying a SUN2 variant in addition to LMNA T528K and R99P mutations , respectively ) also had more severe disease than is typical for EDMD , with early onset at age 1 and 4 years , respectively , and unusually early requirement for heart transplantation . In each of these sporadic cases , the LMNA mutation arose de novo and so no comparisons can be made with family members , however , their clinical phenotype is consistent with the suggestion that the SUN2 variants contribute to increased disease severity . Whilst our genetic and cell-based data strongly support a modifying role for SUN mutations in some patients , studies involving larger patient cohorts will be necessary to prove this conclusively . For 8 of the rare , non-synonymous variants identified in our cohort , there was a lack of compelling evidence of their disease association . However , given the complex interplay between mutations in different genes , more investigation is required before entirely ruling out their involvement . In particular , SUN1 p . V846I was found in an isolated sporadic case and thus no co-segregation analysis could be performed to support its disease association . Yet this is a mutation of highly conserved residue ( Figure S3 ) that lies within the SUN domain that is involved in nesprin binding . Thus it will be important in the future to utilize functional studies to investigate the impact of such mutations on LINC complex interactions . Despite our findings increasing the number of known EDMD-associated genes to 8 , still almost 50% of patients in our cohort have no identified mutations in any of these genes . Most of these patients represent sporadic cases , with no family history of disease , making mutation screening difficult . Furthermore , since 6 of the known genes each account for only a small percentage of cases , it is likely that there are multiple genes remaining to be identified . Proteins associated with the LINC complex are clearly very strong candidates and there are several such proteins that should be examined as a priority , including Samp1 [71] , [72] . Through the various nesprin isoforms expressed at the ONM , the LINC complex mediates attachment to all three cytoskeletal filament networks [17] , [73] . In this study , we have demonstrated , in several different systems , that muscular dystrophy-associated mutations in SUN1 or SUN2 impair nuclear coupling to the both actin and microtubule networks and disrupt nuclear movement and positioning . In mouse NIH3T3 fibroblasts , five out of the six SUN1 and SUN2 variants that we examined inhibited rearward movement of the nucleus , which has previously been shown to be achieved through LINC complex attachment to actin cables closely associated with the nuclear surface [53] . Our data strongly indicate that at least five of the variants identified in our patient cohort have a negative functional impact upon nuclear-cytoskeletal connection via the LINC complex , which is likely to be a major contributor to muscle disease pathophysiology . One of the variants examined , SUN2 p . M50T , did not impair rearward nuclear movement , suggesting that this variant is not disease-causing and this is entirely possible given the complex genetics in the individual carrying this mutation ( patient MD-2 ) . In agreement with our findings , EDMD-associated lamin A variants were recently shown to cause a similar defect in nuclear movement in NIH3T3 cells [52] and disrupted nuclear-cytoskeletal coupling [74] . We also observed defects in nuclear positioning in differentiating myotubes derived from patient MD-1 , carrying compound heterozygous SUN1 p . G68D/p . G338S mutations , which is consistent with recent findings that proper SUN1 and SUN2 recruitment to the NE is required for myonuclear spacing [32] . In strong support for a direct role of SUN proteins , nuclear positioning in skeletal muscle is disrupted in double Sun1/Sun2 knockout mice or in mice with targeted disruptions of nesprin-1/nesprin-2 and this leads to clustering similar to that observed in our patient myotubes [30] , [31] , [75] . Thus , the phenotype we observed in patient MD-1 is consistent with a defect in the LINC complex . Several recent studies have shown that myonuclear position is controlled by nuclear attachment to the microtubule network and that this is mediated by the LINC complex [34] , [35] , [76]–[78] . At the onset of myoblast differentiation , proteins involved in microtubule nucleation redistribute from the centrosome to the NE . Our observations of impaired pericentrin recruitment and microtubule nucleation/organization at the NE in the patient myotubes therefore support a model whereby mutations in SUN proteins impair nuclear-microtubule connection and prevent correct positioning of myonuclei ( Figure 8 ) . It is also well established in other systems that unanchored nuclei float freely in the cytoplasm and tend to clump together , as observed in MD-1 myotubes [34] , [77] , [79] . It is currently not clear how the mutant SUN proteins , which are located at the INM , mediate disruption of microtubule attachment to the NE . Studies have indicated that nesprins are important for microtubule association with the NE , through their interaction with microtubule motor proteins [77] , [80] . However , we did not obtain any evidence that the central SUN1-nesprin-2 LINC complex interaction was perturbed in MD-1 myoblasts , suggesting that the defect may lie elsewhere . Instead , SUN1 interaction with emerin was disrupted and , consistent with this , both emerin mRNA and protein levels were reduced in myoblasts from patient MD-1 . Impairment of SUN1/SUN2 interaction with emerin has also been observed in cases of EDMD1 due to mutations in emerin itself [44] . Furthermore , EDMD has been associated with defects in emerin interaction with lamin A/C and nesprins [18] , [19] . Interestingly , emerin has been shown to partially localize at the ONM , where it may contribute to centrosomal attachment to the NE and , in agreement with our findings , others have observed increased centrosomal separation from the nucleus in EDMD1 cells [81] , [82] . Thus , dysregulation of emerin may play a role in disease causation . To date , most studies have focused on the role of the nuclear lamina and LINC complex in cellular resistance to mechanical strain in support of the “structural” hypothesis of laminopathy disease causation . There is now strong evidence to indicate that defects in these structural networks make a significant contribution to the pathophysiology of EDMD and related disorders [29] . Given our observations of defective interaction networks in SUN-mutated cells and uncoupling of nuclear-cytoskeletal connections , it is therefore likely that the variants we have discovered in SUN1 and SUN2 also impact upon cell mechanics . Nuclear dysmorphology is a common feature of laminopathy cells and , whilst the exact cause and effect of this phenomenon is not understood , it is likely to reflect changes in the organization of the nuclear lamina and its interactions with the nuclear envelope [83] , together with increased susceptibility to mechanical deformation . The exacerbation of nuclear dysmorphology , induced by SUN1 W377C expression in LMNA R401C fibroblasts , again highlights a role for SUN proteins in NE organization and integrity . Our findings in patient MD-1 myotubes further indicate that defects in nuclear positioning may play a significant role in disease pathogenesis , particularly since a link has also been made between loss of myonuclear anchoring and impaired muscle function [34] , [35] . Myonuclear positioning defects have been observed at the myotendinous junctions of LmnaH22P/H22P and Lmna knock-out mice , and in EDMD patients with mutations in LMNA [33] , [84] . However , to our knowledge , ours is the first observation of such a pronounced myonuclear mispositioning phenotype in humans . It will be important , in future studies , to demonstrate directly that uncoupling of the nucleus from the microtubule network through SUN mutation does lead to muscle disease in vivo with physiological expression levels of SUN mutants . In summary , our data clearly implicate defects in pericentrin recruitment , microtubule nucleation/organization and nuclear-cytoskeletal attachment in NE-associated muscular dystrophy pathogenesis and are in agreement with the bulk of results showing SUN1/SUN2 involvement in nuclear positioning and cell migration . It remains to be determined precisely how centrosomal components are recruited to the nuclear envelope in differentiating myotubes and how defects in this process result in misalignment of myonuclei in muscular dystrophy .
This study involved the use of human DNA samples and myoblasts derived from muscle biopsies . These were obtained following informed consent using protocols and consent forms approved by the Ethics Committee of Ernst-Moritz-Arndt University , Greifswald . EDMD patients for this study were selected based on the results of a routine diagnostic mutational analysis of EMD , LMNA , FHL1 , SYNE1 and SYNE2 . 175 pseudo-anonymized patients negative for mutations in these genes and 70 patients known to carry mutations in the genes encoding the LINC components emerin , lamin A/C and nesprin 1 or 2 alpha and beta were tested for mutations in SUN1 and SUN2 . The clinical features of these unrelated , predominantly Caucasian index cases were within the diagnostic criteria for EDMD [50] despite the variable clinical expression . Primer pairs for all the coding exons and flanking intronic sequences of SUN1 ( UNC84A , ENSG00000164828; see Fig . S1 ) and SUN2 ( UNC84B , ENSG00000100242 , ENST00000405510 ) were designed using Primer-Blast ( http://www . ncbi . nlm . nih . gov/tools/primer-blast/index . cgi; Table S2 ) . To standardize the sequencing reaction , all primers were tagged with an M13-tail ( forward: 5′-GTAAAACGACGGCCAGT-3′ reverse: 5′-CAGGAAACAGCTATGAC-3′ ) . Amplifications were performed in 25 µl volumes using Amplikon-Taq Polymerase ( Biomol ) under the following thermal conditions: initial denaturation at 94° for 5 min followed by 35 cycles of denaturation ( 94°C for 15 sec ) , annealing at the appropriate temperature for 15 sec ( see Table S2 ) and elongation ( 72°C for 1 min ) . A final elongation ( 72°C for 7 min ) preceded a 4°C cooling step Direct Sanger sequencing was used to analyse PCR products . Excess dNTPs and primers were removed using ExoSAP-IT ( Affymetrix ) . Sequencing reactions were performed using ABI BigDye Terminator v3 . 1 Cycle Sequencing Kit with addition of 5% DMSO to the reaction mix . M13-oligonucleotides were used as sequencing primers . The reactions were analysed on a 3130xl GA DNA Sequencer ( Applied Biosystems ) according to the manufacturer's instructions . All DNA variations identified were validated using a second independent DNA sample . Unique and rare sequence variations were tested for their frequency in 400 alleles of a Caucasian reference population . Additionally , sequence variations found in a patient of Turkish origin were tested in 138 alleles of a Turkish reference population . Co-segregation of DNA variations with the disease was analysed in patient families if available . For estimating the frequency of DNA variations found , restriction digestion and high resolution melting ( HRM ) were performed using patient DNA as positive control . Restriction enzymes cutting specifically at the DNA variation were selected using NEB-cutter ( http://tools . neb . com/NEBcutter2/ ) . HRM products amplified with LightCycler 480 High Resolution Melting Master ( Roche ) were analysed on a LightCycler 480 II ( Roche ) according to the manufacturer's instructions . Samples showing abnormal signals were examined by restriction endonuclease digestion or direct sequencing . The frequency of changes found in patients of different origin was estimated from online accessible genome sequencing data ( Table S1 ) . RNA was extracted from patient MD-1 and control myoblasts using TRIzol ( Invitrogen ) according to the manufacturer's instructions . Real-time PCR was performed using a RealTime ready custom panel and LightCycler 480 Probes Master ( Roche ) , with primers as described in Supplementary Material Table S3 , and evaluated on a LightCycler 480 II ( Roche ) , according to the manufacturer's instructions . Values for each gene were normalised to both actin and GAPDH . For plasmid constructs , the pCMVTag3B vector ( Stratagene ) was used to fuse a myc tag to the N-terminus of SUN1 . The 916 amino acid version of the SUN1 cDNA , lacking the ATG start codon , was generated by PCR amplification in two stages . First , codons 2–362 were amplified using primers 5′-CACAGAATTCGATTTTTCTCGGCTTCACAT-3′ and 5′-CACAGTCGACCTATCCGATCCTGCGCAAGATCTGC-3′ with IMAGE clone 40148216 as template and inserted into pCMVTag3B via the EcoRI and SalI sites . This introduced a BglII site via a silent mutation at codon 356–358 . Codons 352–916 were then amplified using 5′-TTACTTCTTGCTGCAGATCTTGCGCAGGATCGG-3′ and 5′-GAGAGTCGACTCACTTGACAGGTTCGCCATG-3′ from an oligo dT-primed reverse transcription of U2OS cell mRNA and cloned into the BglII-SalI sites of the initial construct . EDMD-associated mutations were introduced using the QuikChange II site-directed mutagenesis kit ( Stratagene ) , according to the manufacturer's instructions . Anti-human SUN1 2383 and anti-human SUN2 2853 antibodies have been described previously [44] . Anti-SUN1 Atlas antibody ( HPA008346 ) was obtained from Sigma prestige antibodies . Anti-nesprin-2 ( N2N3 ) antibody was kind gift from Q . Zhang ( King's College London ) and has been described previously [16] . Anti-nesprin-2G has been reported previously [53] . Anti-nesprin-2 monoclonal antibody ( IQ562 ) was purchased from Immuquest . Monoclonal anti-emerin antibody was a kind gift from G . Morris ( Center for Inherited Neuromuscular Disease , Oswestry , UK ) . Anti-lamin A/C ( sc-6215 ) and GFP antibodies were purchased from Santa Cruz Biotechnologies . Anti- GAPDH ( MAM374 ) was obtained from Millipore . Anti-α-tubulin ( T9026 ) , anti-β-actin ( A5441 ) , anti-γ-tubulin ( T6557 ) , anti-myc and anti-desmin antibodies were purchased from Sigma . Anti-caveolin 3 monoclonal antibody ( 610420 ) was purchased from Transduction Laboratories and anti-desmin polyclonal antibody ( MONX10657 ) was purchased from Monosan . Anti-pericentrin polyclonal antibody ( Ab4448 ) was obtained from Abcam . Myoblasts from patient MD-1 and controls were routinely cultured in high-glucose DMEM supplemented with 20% foetal bovine serum plus antibiotics penicillin , streptomycin and amphotericin B , at 37°C and 5% CO2 , and were used between passages 3 and 7 . Myoblasts at confluence were allowed to differentiate into myotubes in the same culture medium for 8–15 days , replacing the medium every 5 days . HeLa cells were cultured in DMEM supplemented with 10% FBS and antibiotics . For emerin co-immunoprecipitation experiments , HEK293 cells were transfected with the appropriate pCMVTag3-SUN1 constructs together with GFP-emerin [85] using Fugene 6 ( Promega ) , according to the manufacturer's instructions . pCMVTag3-SUN1 constructs were transfected into C2C12 mouse myotubes using the Amaxa Nucleofector ( Lonza ) , according to the manufacturer's instructions . Cultures were fixed 24 hours after transfection and processed for immunofluorescence analysis . NIH3T3 fibroblasts were cultured in 10% calf serum in DMEM ( Gibco ) as previously described [86] . Following serum starvation for two days , confluent monolayers were “wounded” by removing a strip of cells and nuclei of cells at the edge of the wound were microinjected with the appropriate myc-tagged SUN1 or SUN2 DNA plasmids . After expression for 2 hr , cells were stimulated with 10 µM LPA for 2 hr , fixed in 4% paraformaldehyde , extracted with Triton X-100 and stained with antibodies to tyrosinated α-tubulin ( rat monoclonal antibody at 1/40 of culture supernatant ) , myc ( mouse monoclonal antibody from clone 9E10 , Roche ) and DAPI ( Sigma ) followed by appropriate secondary antibodies . Stained samples were observed with a Nikon TE300 microscope using a 40× Plan Apo N . A . = 1 . 0 or 60× Plan-Apo N . A . = 1 . 4 objective and filter cubes optimized for DAPI , fluorescein/GFP , and rhodamine . Images were acquired with CoolSNAP HQ camera ( Photometrics ) driven by Metamorph software ( MDS Analytical Technologies ) and further processed in Image J . Centrosomes were considered oriented if they were localized in the pie-shaped sector between the nuclear membrane the leading edge scored , as described [86] , [87] . Random orientation is ∼33% by this measure . Nuclear and centrosome position relative to the cell centroid were determined as described [88] . Data were plotted as % of the cell radius to normalize for differences in cell size . To prepare total cell extracts for immunoblotting , cells were scraped into cold 1×phosphate-buffered saline ( PBS ) , pelleted by centrifuging at 200×g for 5 min and then pellets were resuspended in lysis buffer ( 10 mM HEPES [pH 7 . 4] , 5 mM EDTA , 50 mM NaCl , 1% Triton X-100 , 0 . 1% SDS ) supplemented with 1 mM PMSF and protease inhibitor cocktail ( Roche ) and an equal volume of Laemmli buffer was then added . For human myoblast immunoprecipitations , cells were grown on 10 cm dishes and then immunoprecipitated as described previously ( Haque et al . , 2006 ) using 2 µg of SUN1 2383 antibody . 5% of the initial lysate was retained for immunoblot analysis . All samples were boiled in an equal volume of 2×Laemmli buffer , resolved on 6% or 7 . 5% or 10% polyacrylamide gels , followed by semidry transfer onto nitrocellulose membrane . Membranes were probed using the appropriate primary antibodies and dilutions: hSUN1 ATLAS ( 1∶400 ) , hSUN2 2853 ( 1∶500 ) , lamin A/C ( 1∶2000 ) , emerin ( 1∶1500 ) , nesprin-2 N2N3 ( 1∶1500 ) , α-tubulin ( 1∶ 10 , 000 ) , β-actin ( 1∶20 , 000 ) , GAPDH ( 1∶ 10 , 000 ) . Primary antibodies were detected using horseradish peroxidase-conjugated secondary antibodies ( Sigma ) , and visualization was performed using ECL reagents ( Geneflow ) . Myoblasts and myotubes grown on glass coverslips were fixed in methanol at −20°C and processed for indirect immunofluorescence microscopy as previously described ( Haque et al . , 2006 ) . For SUN1 staining , cells were instead fixed in 4% paraformaldehyde and permeabilized with 0 . 5% Triton X-100 at room temperature for 5 min . Cells were washed in PBS and incubated with antibodies diluted in PBS–3% bovine serum albumin , using hSUN1 2383 ( 1∶150 ) , hSUN2 2853 ( 1∶100 ) , lamin A/C ( 1∶400 ) , emerin ( 1∶500 ) , nesprin-2G ( 1∶300 ) , γ-tubulin ( 1∶500 ) pericentrin ( 1∶50 ) , caveolin 3 ( 1∶30 ) and desmin ( 1∶100 ) antibodies . Secondary antibodies were goat anti-rabbit AlexaFluor 488 , donkey anti-mouse AlexaFluor 594 and donkey anti-goat AlexaFluor 594 ( Molecular Probes Inc . ) . DNA was stained with 50 µg/ml 4′ , 6-diamidino-2-phenylindole ( DAPI; Sigma ) . Coverslips were mounted in 80% glycerol–3% n-propyl gallate ( in PBS ) or ProLong gold antifading reagent ( Invitrogen ) . Fluorescence microscopy was performed with a Nikon TE300 inverted microscope with an ORCA-R2 charge-couple device camera ( Hamamatsu ) and Volocity software ( PerkinElmer ) . Where required fluorescence microscopy was also performed with Leica TCS SP5 confocal laser scanning microscope and Leica LAS AF software . Images were processed with Adobe Photoshop ( Adobe Systems ) . Quantification of fluorescence intensity was performed using an Olympus Scan∧R microscope with a 20× objective . Approximately 1000 nuclei from 3 independent experiments were randomly selected by their DAPI signal , and the intensity of SUN1 , SUN2 , emerin , lamin A/C and nesprin-2 was measured within the DAPI-stained region . Myotubes ( at passage 2–3 ) from patient and age-matched controls were fixed in 2 . 5% glutaraldehyde-0 . 1 M cacodylate buffer pH 7 . 4 for 3 h at 4°C . After post-fixation with 1% osmium tetroxide ( OsO4 ) in cacodylate buffer for 2 h , samples were dehydrated in an ethanol series , infiltrated with propylene oxide and embedded in Epon resin . Ultrathin sections ( 60 nm thick ) were stained with uranyl acetate and lead citrate ( 10 min each ) and were observed at 0° tilt angle with a Geol Jem 1011 transmission electron microscope , operated at 100 kV . At least 30 myoblasts/myotubes per sample were observed . In all cases , statistical analysis was performed using a Student's t-test to compare differences in values obtained for patient/mutant versus control samples . | Emery-Dreifuss muscular dystrophy ( EDMD ) is an inherited disorder involving muscle wasting and weakness , accompanied by cardiac defects . The disease is variable in its severity and also in its genetic cause . So far , 6 genes have been linked to EDMD , most encoding proteins that form a structural network that supports the nucleus of the cell and connects it to structural elements of the cytoplasm . This network is particularly important in muscle cells , providing resistance to mechanical strain . Weakening of this network is thought to contribute to development of muscle disease in these patients . Despite rigorous screening , at least 50% of patients with EDMD have no detectable mutation in the 6 known genes . We therefore undertook screening and identified mutations in two additional genes that encode other components of the nuclear structural network , SUN1 and SUN2 . Our findings add to the genetic complexity of this disease since some individuals carry mutations in more than one gene . We also show that the mutations disrupt connections between the nucleus and the structural elements of cytoplasm , leading to mis-positioning and clustering of nuclei in muscle cells . This nuclear mis-positioning is likely to be another factor contributing to pathogenesis of EDMD . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"biology",
"and",
"life",
"sciences",
"medicine",
"and",
"health",
"sciences"
] | 2014 | Muscular Dystrophy-Associated SUN1 and SUN2 Variants Disrupt Nuclear-Cytoskeletal Connections and Myonuclear Organization |
Leishmania infantum is the causative agent of visceral and cutaneous leishmaniasis in the Mediterranean region , South America , and China . MON-1 L . infantum is the predominating zymodeme in all endemic regions , both in humans and dogs , the reservoir host . In order to answer important epidemiological questions it is essential to discriminate strains of MON-1 . We have used a set of 14 microsatellite markers to analyse 141 strains of L . infantum mainly from Spain , Portugal , and Greece of which 107 strains were typed by MLEE as MON-1 . The highly variable microsatellites have the potential to discriminate MON-1 strains from other L . infantum zymodemes and even within MON-1 strains . Model- and distance-based analysis detected a considerable amount of structure within European L . infantum . Two major monophyletic groups—MON-1 and non-MON-1—could be distinguished , with non-MON-1 being more polymorphic . Strains of MON-98 , 77 , and 108 were always part of the MON-1 group . Among MON-1 , three geographically determined and genetically differentiated populations could be identified: ( 1 ) Greece; ( 2 ) Spain islands–Majorca/Ibiza; ( 3 ) mainland Portugal/Spain . All four populations showed a predominantly clonal structure; however , there are indications of occasional recombination events and gene flow even between MON-1 and non-MON-1 . Sand fly vectors seem to play an important role in sustaining genetic diversity . No correlation was observed between Leishmania genotypes , host specificity , and clinical manifestation . In the case of relapse/re-infection , only re-infections by a strain with a different MLMT profile can be unequivocally identified , since not all strains have individual MLMT profiles . In the present study for the first time several key epidemiological questions could be addressed for the MON-1 zymodeme , because of the high discriminatory power of microsatellite markers , thus creating a basis for further epidemiological investigations .
Visceral leishmaniasis ( VL ) caused by Leishmania infantum ( synonym L . chagasi , [1] ) is a public-health problem in most countries bordering the Mediterranean , China and South America . Currently , the epidemiology of Mediterranean VL is changing . Increasing incidence [2] and a shift in the bulk of cases from children to adults [3] , [4] , related to the emergence of HIV , has been reported . Since 1985 up to 80% of the cases have occurred in immunocompromised adults [5] , [6] . Leishmania/HIV co-infections have become increasingly frequent in Southern Europe , particularly in Spain , France , and Italy , with 25–70% of adult cases being related to HIV infection and up to 9% of AIDS cases developing VL [7] , [8] . In addition to the classical vector-based disease transmission , an anthroponotic cycle has emerged among intravenous drug users where syringes replace the sand fly vector [9]–[14] . Recently , L . infantum parasites have been found to have spread northward in continental Italy perhaps due to climatic changes [15]–[17] . Dogs are the main reservoir hosts for L . infantum , being part of the domestic ( pet dogs ) and peridomestic ( stray dogs and wild canids ) transmission cycles . The prevalence of canine leishmaniasis is high in all European Mediterranean countries [18]–[20] . The gold standard method for typing Leishmania is still Multilocus Enzyme Electrophoresis ( MLEE , isoenzyme analysis ) . Most widely used is the Montpellier system ( MON ) which is based on the analysis of 15 enzymes [21] . Leishmania infantum is characterized by a broad enzymatic polymorphism . At present this species includes 31 zymodemes of which 30 have been found in humans [22] . Some of them were related to VL only ( e . g . MON-27 , 28 , 72 , 77 , 187 ) , others only to cutaneous leishmaniasis ( CL ) ( e . g . MON-11 , 29 , 33 , 78 , 111 ) , and few were isolated from both VL and CL cases ( e . g . MON-1 , 24 , 34 , 80 ) . In L . infantum/HIV co-infections the tropism of some zymodemes ( MON-24 , 29 , 33 , 78 ) has changed from CL to VL [23]–[25] . MON-1 is the most prevalent zymodeme; it occurs in more than 30 countries worldwide , and represents approximately 70% of all identified strains [26] . In Europe , it varies between 88% in Southern France [25] , over 50% in Italy [27] , 96 . 7% in Portugal [28] and 44–58% in Spain [9] , [29] , [30] . Up to 73% of HIV/VL co-infections in Europe are due to this zymodeme [24] . In immunocompetent patients MON-1 causes up to 90% of VL cases , but only 20% of CL cases [11] . MON-1 is also the prevalent zymodeme in dogs [19] , [25] , [29] whilst other zymodemes such as MON-98 , 108 , 253 , 77 , and 24 are found occasionally . In contrast , MON-1 has only been detected in 18% of the sand fly samples [29] . Epidemiological studies on VL caused by L . infantum require the use of techniques that are able to differentiate MON-1 strains . The first indications for heterogeneity among MON-1 strains were based on RAPD analyses [31]–[34] , analysis of three microsatellite markers [35] , PCR-RFLP of the intergenic cpb and intragenic gp63 regions [36] and RFLP analysis of minicircle kDNA [37] , [38] . Microsatellites are tandemly repeated stretches of short nucleotide motives of 1–6 bp ubiquitously distributed in eukaryotic genomes . They mutate at rates five to six orders of magnitude higher than the bulk of DNA . These highly polymorphic and co-dominant markers have been shown to be very useful for population studies [39] and have been applied for a number of species , among them quite recently L . tropica [40] and the L . donovani complex [41] . A comparison of different genotyping methods targeting Leishmania DNA regions with different molecular clocks [42] revealed that kDNA PCR-RFLP and multilocus microsatellite typing ( MLMT ) were the most powerful tools for MON-1 strain tracking . In the present study we performed MLMT using a set of 14 microsatellite markers for 141 strains of L . infantum of different zymodemes with strong sampling emphasis on MON-1 , mainly from Spain , Portugal and Greece in order to investigate the population structure and dynamics in the corresponding natural foci . We also attempted to correlate microsatellite patterns with host specificity and manifestation of the disease . Strains from recurrent infections were included to test whether MLMT was able to differentiate between relapse and re-infection .
Sources , designation , geographical origins , MLEE identification and clinical manifestation of the Leishmania infantum strains are listed in Table 1 . The 66 Spanish strains were collected from humans and dogs in four regions-Madrid , Ibiza , Majorca , and Catalonia . Forty four strains , from humans , dogs and sand fly vectors , were obtained from four Portuguese regions: Metropolitan region of Lisbon , Alentejo , Algarve , Alto Douro . The 16 human and canine strains from Greece are originating from two foci-Athens and Crete . The seven strains from France were collected in four of the five known endemic foci: Cévennes , Côte d'Azur , Provence , and Pyrénées-Orientales . Additional strains or the respective DNA samples were obtained from the following culture collections: Centre National de Référence des Leishmania , Montpellier , France; KIT ( Royal Tropical Institute ) , Amsterdam , Netherlands; WHO's Jerusalem Reference Centre for Leishmaniases , Hebrew University–Hadassah Medical School , Jerusalem , Israel; London School of Hygiene and Tropical Medicine , London , UK . The strain set included isolates from 12 relapse cases . Ten of them represented two episodes of infection ( original infection and relapse ) , one three and another one four episodes . All parasites were cultivated as described previously [37] , [43] , [44] . DNA was isolated using proteinase K-phenol/chloroform extraction [45] , suspended in TE-buffer or distilled water and stored at 4°C until use . A set of fourteen primer pairs were used for microsatellite amplification ( Table 2 ) as previously described [41] , [46] . Fragments containing single microsatellites were analyzed by either MetaPhor agarose gel electrophoresis , PAGE or capillary electrophoresis . Four percent MetaPhor agarose gels ( BioWhittaker Molecular Applications , USA ) were used basically for a pre-screening of the strains for polymorphisms [46] . For PAGE 6 to 15 μl of the PCR product were mixed with loading buffer and run under non-denaturating conditions on 12% polyacrylamide gels at 1 kV for 6 h . Gels were silver stained and dried [47] . For high throughput studies PCR products from amplified microsatellites were analyzed with the fragment analysis tool of the CEQ 8000 automated genetic analysis system of Beckman Coulter , USA [46] , using fluorescence-conjugated forward primers ( Proligo , France ) for microsatellite amplification . Population structure was investigated by the STRUCTURE software [48] , which applies a Bayesian model-based clustering approach . This algorithm identifies genetically distinct populations on the basis of allele frequencies . Genetic clusters are constructed from the genotypes identified , estimating for each strain the fraction of its genotype that belongs to each cluster . This clustering method proved superior to distance-based approaches for processing data sets of low variability like those presented by L . infantum MON-1 . The following parameters were used: burning period of 20 , 000 iterations , probability estimates based on 200 , 000 Markov Chain Monte Carlo iterations . The most appropriate number of populations was determined by comparing log-likelihoods for values of K between 1 and 16 . The log-likelihood values were compared in a diagram . At the plateau ( maximum ) of the derived Gaussian graph the value of K captures the major structure of the populations . In addition we calculated ΔK , which is based on the rate of change in the log probability of data between successive K values [49] . Ancestral source populations were identified by decreasing the number of K . Phylogenetic analysis was based on microsatellite genetic distances , calculated with the program MICROSAT [50] for the numbers of repeats within each locus using two different measures: DAS ( Dps ) , based on the proportion of shared alleles [51] and Chord-distance [52] . Both distances follow the infinite allele model ( IAM ) . Neighbor-joining trees ( NJ ) of both distance matrices were constructed in PAUP , version 4 . 0b8 [53] . Confidence intervals were calculated by bootstrapping ( 100 replications ) [54] using the program POPULATIONS 1 . 2 . 28 ( http://www . legs . cnrs-gif . fr/bioinfo/populations ) . For visualising the genetic substructure at population and individual level we applied a factorial correspondence analysis ( FCA ) implemented in the GENETIX software [55] . This test places the individuals according to the similarity of their allelic state in a three dimensional space . Microsatellite markers as well as populations were analyzed with respect to diversity of alleles ( A ) , expected ( gene diversity ) and observed heterozygosity ( He and Ho , respectively ) , and the inbreeding coefficient FIS applying GDA ( http://hydrodictyon . eeb . uconn . edu/people/plewis/software . php ) and GENEPOP 3 . 4 [56] . Genetic differentiation and gene flow [57] was assessed by F-statistics calculating the FST ( theta ) values ( IAM ) [58] with the corresponding p-values ( confidence test ) using the MSA software [59] . Migration rates ( gene flow ) Nm were calculated as Nm = 0 . 25 ( 1-FST ) /FST [60] . Indices of association ( IA ) were calculated to test each population for clonality and recombination using MULTILOCUS [61] .
A total of 106 different MLMT genotypes based on 14 microsatellite markers were identified for 141 strains of L . infantum mostly from Southern Europe . A Bayesian model-based clustering algorithm implemented in the software STRUCTURE was used to infer the population structure of European L . infantum based on these MLMT data . According to ΔK , the most probable number of populations for the complete data set of 141 strains was four: ( 1 ) MON-1 from Greece , Turkey , Israel , and Tunisia ( +1 strain from France ) ; ( 2 ) MON-1 from Majorca and Ibiza ( Spain ) ; ( 3 ) MON-1 from Portugal and mainland Spain; and ( 4 ) non-MON-1 strains from Spain , Portugal , Italy , Malta ( Figure 1 ) . Ancestral source populations were identified by successively increasing the number of populations ( K ) from 2 as indicated by the bars next to the tree in Figure 2 . The first and most important split at K = 2 divided non-MON-1 from MON-1 strains . The MON-1 cluster was subdivided at K = 3 into a group of Iberian strains , which included strains from France , and a group comprising strains from Greece , Turkey , Israel , and Tunisia . Finally , at K = 4 , the Iberian group split into mainland and Balearic Islands ( Table 1 ) . One hundred and fifteen strains representing 84 distinct genotypes formed the MON-1 group and 26 strains representing 22 genotypes the non-MON-1 group , respectively . The assignment of strains to these groups was not always compatible with zymodeme identification: three Spanish MON-1 strains ES1 ( I ) , ES3 ( I ) , ES6 ( I ) , clustered with non-MON-1 strains , whereas strains INF-32 ( MON-77 ) , INF-35 ( MON-108 ) and all Greek MON-98 strains grouped with the MON-1 strains . Populations as defined by STRUCTURE were used for all subsequent population genetics analyses . In addition to the model-based algorithm , we used a genetic distance-based approach to infer the population structure of European L . infantum ( Figure 2 ) . The same 4 major populations as defined by STRUCTURE were detected , as indicated by bars , however they were not supported by significant bootstrap values ( data not shown ) . This may be due to the extremely high similarity of these strains , a certain amount of homoplasy and the existence of strains with mixed or intermediate genotypes . Bootstrap values >50% were obtained only for nodes between some subgroups inside the major clusters . The identical results obtained by two different methods , however , strongly support the existence of these particular main populations . The MON-1 and non-MON-1 clusters were monophyletic . Zymodemes MON-77 , 108 and 98 were , again , members of the MON-1 cluster . There were two subclusters of non-MON-1 strains that were also observed in STRUCTURE at K = 5 ( data not shown ) : one comprises predominantly MON-34 , MON-27 and MON-80 , the other MON-11 , 29 , 183 , 188 , 198 , 199 , 228 . MON-24 is present in both subclusters . Long branches in the non-MON-1 group indicated a higher diversity among these strains . The number of microsatellite alleles ranged from 2–10 ( mean 4 . 6 ) for the MON-1 strains , and from 3–13 ( mean 5 . 6 ) for the non-MON-1 strains ( Table 2 ) . The most variable markers were Li 22–35 and Lm2TG , the least variable ones Li 46–67 and TubCA . The observed heterozygosity ( Ho ) varied between 0–0 . 08 for MON-1 strains , and 0–0 . 269 for non-MON-1 strains , indicating that most microsatellite loci were heterozygous in at least one strain , with a higher degree of heterozygosity in non-MON-1 strains . The expected heterozygosity ( He ) as a measure of genetic diversity was between 0 . 018–0 . 768 ( MON-1 ) and 0 . 274–0 . 876 ( non-MON-1 ) , and , in most of the markers , much higher than the mean Ho ( 0 . 029 MON-1 and 0 . 118 non-MON-1 ) . The mean of the inbreeding coefficients was 0 . 912 and 0 . 81 for MON-1 and non-MON-1 strains , respectively , pointing to a homozygotes predominance . The non-MON-1 strains were more diverse , with greater distances on the Neighbor-joining tree , higher numbers of allelic variants and higher values of diversity measures , in spite of the much smaller number of strains , perhaps reflecting the combination of distinct zymodemes in this group . Traces of gene flow were detected between the four populations using the Bayesian algorythm ( Figure 1A , 3 ) . Several strains could not be assigned to only one population by STRUCTURE as they showed mixed ancestry . Some strains were considered more likely to belong to one of the populations , but others had clear shared memberships for two populations ( e . g . ES9 ( III ) , ES10 ( III ) , PT2 ( I ) , and INF-32 ) ( Figure 3; Table 3 , 4 ) . ES9 ( III ) and ES10 ( III ) ( human isolate with its relapse ) had heterozygous combinations of alleles characteristic for non-MON-1 strains and MON-1 strains ( population 2 ) for 10 of the 14 markers . The sand fly isolate PT2 ( I ) had such MON-1 ( population 3 ) /non-MON-1 allele combinations in 9 markers . The remaining 5 homozygous markers were in general not discriminating between MON-1 and non-MON-1 strains . INF-32 presented also a mixture of patterns typical for non-MON-1 and MON-1 ( population 3 ) , being heterozygous in 7 markers . In addition there were several strains which also have alleles typical for both , the MON-1 and non-MON-1 group , however with MON-1 alleles clearly predominating , as the canine strain PT13 ( II ) , the sandfly strains PT7 ( I ) and the human strains PT8 ( I ) and PT8 ( II ) ( Tables 3 , 4 ) . The two sandfly isolates PT7 ( I ) and PT6 ( I ) as well as PT8 ( I ) and PT8 ( II ) from Alto Douro showed an unique allele for marker Li 22–35 ( 118 bp ) that differed significantly from the usual MON-1 alleles and rather resembled the range of non-MON-1 alleles ( Tables 4 , 5 ) . ES16 ( I ) is homozygous for all loci , however represents a combination of alleles typical for different populations , with dominating population 3 membership ( Figure 3 , Table 3 ) . All those strains ( PT13 ( II ) , PT8 ( I ) , PT8 ( II ) , PT7 ( I ) , ES16 ( I ) ) seem to exhibit mosaic genotypes . Three of the strains showing heterozygous MON-1 and non-MON-1 alleles ( PT2 ( I ) , ES9 ( III ) , ES10 ( III ) ) had in the NJ tree an intermediate position between the two respective clusters . Strain INF-32 , although being heterozygous for some microsatellite markers was here part of the MON-1 cluster , suggesting the predominance of MON-1 traits . The four populations recognized by model-based and distance-based analyses were also supported by F statistics . All FST values ( Table 6 ) were >0 . 25 and significant ( p = 0 . 0001 ) indicating strong genetic differentiation between the four populations . Populations 2 ( Spain Majorca and Ibiza ) and 3 ( Portugal and Spain mainland ) were the most closely related samples , an observation that is highly congruent with their geographic distribution and the maritime transport . Moreover , migration rates ( Nm ) estimates between all four L . infantum populations were low ( Table 6 ) . The graphical representation of factorial correspondence analysis ( FCA ) of the MLMT data ( Figure 4 ) clearly mirrors the population structure parameters . The split between MON-1 and non-MON-1 strains ( K = 2 ) and the higher genotypic diversity within non-MON-1 strains is apparent from Figure 4A . The three main MON-1 populations are clearly separated , when only MON-1 strains were analyzed ( Figure 4B ) . The mean number of alleles ( MNA ) ranged from 2 . 79–3 . 07 among the three MON-1 populations , and was 6 . 29 for the non-MON-1 population ( Table 7 ) . The proportion of polymorphic loci was 0 . 786 for all three MON-1 populations and 1 . 0 for the non-MON-1 population . The values of He differed between MON-1 ( 0 . 197–0 . 348 ) and non-MON-1 ( 0 . 621 ) . The highest degree of heterozygosity was observed for the non-MON-1 population , the lowest in population 3 ( MON-1 Portugal+Spain mainland ) . There were no indications for aneuploidy in the tested strains , since we never observed three or four peaks in the electrophoregrams , suggestive of tetra- or triploidy of the tested markers . The inbreeding coefficient FIS ranged from 0 . 770 to 0 . 933 among the MON-1 populations and was lowest for the non-MON-1 population ( 0 . 725 ) . This suggests a high degree of inbreeding in all four populations , but especially in the MON-1 populations 1 ( Greece , Turkey , Israel , Tunisia ) and 3 ( Portugal+Spain mainland ) . Multilocus linkage associations were highly significant for all four populations pointing to a predominantly clonal reproduction within these populations . Allele frequencies were different in the different populations and group-specific alleles could be recognized ( Table 5 ) , which were congruent with the population splits . The number of group specific alleles was significantly higher among non-MON-1 strains ( 46 ) than in the three MON-1 populations ( 4–7 ) . Within MON-1 , population 1 ( Greece , Turkey , Israel , Tunisia ) had the highest number of specific alleles and it appeared as most diverse population , whereas population 3 was the least diverse . The distribution of the Mediterranean MON-1 strains belonging to populations 1–3 in the respective countries and their studied endemic foci as well as the proportion of each population sampled in the respective geographical region are illustrated in Figure 1B . No correlation was found between MLMT genotypes and host specificity or clinical manifestation ( Figure 1A , 2 ) . Human and canine isolates were present in all MON-1 populations . Strains from Leishmania/HIV+ co-infected patients were represented in all four populations , whereas those few from CL cases were found only in populations 3 ( mainland Spain+Portugal ) and 4 ( non-MON-1 ) . The three strains isolated from Portuguese Phlebotomus had quite peculiar genotypes . Strain PT2 ( I ) from Algarve had alleles characteristic for both , MON-1 and non-MON-1 , as described above . The two sand fly strains from Alto Douro , PT6 ( I ) and PT7 ( I ) , had a unique Li 22–35 allele ( 118 bp ) , found only in strains from this endemic focus and differing significantly from the common MON-1 alleles ( 92 and 94 bp ) , which resembles rather the range of non-MON-1 alleles . PT7 ( I ) also presented both MON-1 and non-MON-1 alleles for some of the markers , albeit with MON-1 genotype dominating . Of the 12 relapse cases , 6 occurred in MON-1 , one in MON-1/MON-98 , two in MON-24 , another two in MON-34 and one in MON-27 . Ten pairs of isolates presented identical MLMT profiles . Identical genotypes were found for a case from Greece , where the original strain was identified as MON-98 and the relapse as MON-1 [62] . In the case with four episodes , strains from episodes 1 and 2 ( ES11 ( I ) +ES12 ( I ) ) were identical but distinct from strains from episodes 3 and 4 ( ES13 ( III ) +ES14 ( III ) ) , which were identical . The two pairs differed only in Lm2TG with an additional allele present in episodes 1 and 2 . The case with three episodes ( ES15 , 16 , 17 ( III ) ) had differences in two markers . The first episode isolate showed two alleles for Li 23–41 and LIST7031 , the second only in Li 23–41 and the third was homozygous for both markers .
The population structure of European L . infantum had , so far , been poorly understood , because most strains belong to a single zymodeme , MON-1 . The recent development of microsatellite markers , which discriminate within this zymodeme [46] , enabled for the first time to reliably address key epidemiological questions such as i ) the existence of geographical subpopulations and the extent of gene flow between them , ii ) the impact of zoonotic and anthroponotic transmission cycles , iii ) the role of reservoir hosts and vectors in sustaining genetic diversity , iv ) comparison of genetic diversity in immuno-competent and immuno-compromised hosts , v ) differential identification of re-infection or relapse in treated patients , especially the immuno-compromised , vi ) the role of mutation and recombination in creating genetic diversity . We detected considerable genetic structure within European strains of L . infantum using model and distance-based analysis methods . The main split between MON-1 and non-MON-1 strains observed in previous MLMT studies of however only few strains of L . infantum [41] , [46] has been confirmed . Both studies had suggested that both MON-1 and non-MON-1 strains each form a monophyletic group , and are independent as are the other populations of the L . donovani complex identified for L . donovani strains from distinct geographical regions [41] . Monophyly of the MON-1 group was also supported by recent studies based on RFLP analysis of intergenic regions of cpb and gp63 [36] , MLST [63] , [64] and by a multifactorial genetic analysis of RFLP , MLMT , MLST and sequence analysis of non-coding regions [65] , which however , all included only few MON-1 strains . The non-MON-1 population was more diverse , with more alleles , longer branches in the tree and a broader distribution of the strains in FCA . This group may include several subpopulations , perhaps consistent with different zymodeme groups and/or geography . The position of two strains–INF-48 ( Buck , Malta ) and INF-57 ( ISS800 , Italy , Sicily ) was odd and could not be resolved with confidence . In most studies using different genetic markers they showed the most distant position among the whole L . infantum cluster , at the basis of the non-MON-1 cluster right after the split from the L . donovani strains [36] , [41] , [63] , [65] . A recent MLST study [64] placed INF-48 and INF-57 close to the MON-1 cluster . According to STRUCTURE in the present study , which does not include L . donovani , both strains are members of the non-MON-1 group , INF-48 had , however , partial membership in MON-1 . In the distance tree INF-48 had a basal position within the MON-1 cluster , and INF-57 was a member of the non-MON-1 cluster . The phylogenetic relationships of these strains and the structure of the non-MON-1 group should be tested using a much larger set of strains including as many zymodemes as possible . The MON-1 cluster also included other zymodemes , which must be closely related: MON-108 , MON-77 and MON-98 . Indeed , these and MON-1 , together with four other zymodemes ( MON-253 , 27 , 105 , 72; not studied here ) , formed a distinct subcluster in MLEE trees , with MON-1 as the putative original zymodeme and the other zymodemes differing in the mobility of only one isoenzyme ( Malic enzyme-ME , glucose-6-phosphate dehydrogenase–G6PD , NADH diaphorese-DIA , purine nucleoside phosphorylase 1-NP1 , phosphoglucomutase-PGM , respectively ) [22] . Different enzymatic profiles might result from post-translational modifications , as sequencing of ME which is discriminating MON-1 and MON-98 did not reveal any nucleotide differences between these zymodemes [63] . All other 20 zymodemes , 11 of which were studied here and assigned to the non-MON-1 population by MLMT , differed in 2 to 4 enzymes . Interestingly , all zymodymes of the non-MON-1 population share the 104 phenotype for MDH ( Malate dehydrogenase ) , in contrast to zymodemes MON-1 , MON-98 , MON-77 and MON-108 forming our MON-1 population which share the 100 phenotype for MDH . Similarly , for NP1 , all members of the non-MON-1 population present the 130 or 140 phenotype , whereas strains of the MON-1 population all present phenotype 100 . Furthermore , all zymodemes belonging to the MON-1 population were also found in dogs , which is in favour of the proximity of these zymodemes . Only few zymodemes of the non-MON-1 population , MON-199 , MON-34 , and MON-11 , have been occasionally isolated from canines . MLMT showed that MON-1 strains are not genetically identical as previously suggested by MLEE and many different genetic markers , but represent rather families of related clones . The MON-1 group is indeed a more homogenous population than the non-MON-1 group , and characterized by a low level of heterozygosity . One could speculate that the predominance of MON-1 and the lack of diversity are due to a quite recent evolutionary history like a bottleneck followed by a rapid epidemic spread . However , its wider spread when compared to other zymodemes might be due to a better fitness and success in infecting the canine host . Using MLMT , we identified for the first time , geographic substructures within MON-1 strains . Three clusters emerged from this study: ( i ) Greece/Turkey/Israel/Tunisia , ( ii ) Spanish Baleares islands and ( iii ) mainland Portugal and Spain . This observation has important epidemiological implications , as it allows the estimation of migration rates and provides a lacking biogeographical perspective for control strategies . Microsatellite markers allowed even differentiation of MON-1 strains between endemic foci within the same country . In Spain , the majority of the strains from the mainland were genetically separated from those isolated on the two Mediterranean islands , which is most probably related to the existing geographical barrier . Some strains from the mainland presented a genotype typical for the islands and vice versa , which can be explained by the transfer of parasites due to travel of infected persons or dogs . Analysing RFLP of kDNA [37] concluded that L . infantum from Majorca constituted a clonal population , though no mainland strains were included in that study . Whether strains from Catalonia ( the three analysed herein grouped together in all distance trees ) differ from other mainland strains should be re-investigated with larger sample sizes . In Portugal , no subclusters correlating with the four regions were identified , which is congruent with a recent RFLP study on kDNA [38] . These authors explain the lack of focus-specific genotypes by the small size of the country and the frequent migration between foci . According to our data , Alto Douro seems to be the most divergent endemic focus in Portugal , followed by the Algarve . Seven strains from Alto Douro , three human ( PT8 ( I ) , PT8 ( II ) , PT10 ( I ) ) , two canine ( PT13 ( II ) , PT14 ( II ) ) and two Phlebotomus ariasi isolates ( PT6 ( I ) and PT7 ( I ) ) , presented unique alleles for 8 of the 14 markers used , in some cases identical to those found only in non-MON-1 strains , in other cases resembling the repeat range typical for non-MON-1 strains and in three cases alleles were found exclusively in Alto Douro . The human isolate PT8 ( II ) had heterozygous loci combining the specific Alto Douro alleles and the alleles typical for all other strains of the mainland ES/PT population . Some of the strains from this focus ( PT8 ( I ) , PT7 ( I ) , PT8 ( II ) , PT13 ( II ) ) showed for single loci alleles typical for non-MON-1 . The Greek MON-1 strains could not be further differentiated regardless of whether they came from the Athens area or the island of Crete . The strains from Turkey , Israel and the single strain from North Africa ( Tunisia ) grouped with those from Greece but seem to represent a distinct genotype . This needs , however to be confirmed on a larger sample set from those regions . Two important interrelated epidemiological questions are the impact of zoonotic and anthroponotic transmission cycles and the role of reservoir hosts and vectors in sustaining genetic diversity . In our study no general relationships between MLMT genotypes , host and reservoir were detected . Human and canine isolates were present in all of the three main MON-1 populations . Moreover , in all MON-1 groups cases of identical MLMT profiles were identified for canine and human ( immunocompetent and immunocompromised ) isolates , thus pointing to transmission of the same parasite between humans and dogs , the domestic host . Such groups of identical genotypes were found in each of three MON-1 populations . The single isolate from a fox , representing a sylvatic host , did not show any special position within the MON-1 strains . An interesting question is why only a single predominating zymodeme is found in dogs–MON-1 . We could , however observe a considerable amount of diversity among the canine MON-1 isolates , even in such a small territory as Ibiza , where most of them had different genotypes . A complete transmission cycle and the vector status of P . ariasi was confirmed by detecting nearly identical genotypes for human , canine and sand fly isolates from the Alto Douro focus . Such complete transmission cycles were also shown recently using kDNA PCR-RFLP [37] , [38] . Importantly , our data did not advocate for a correlation between MLMT profile and clinical manifestation . VL was present in all populations , whereas CL only in populations 3 ( mainland Spain+Portugal ) and 4 ( non-MON-1 ) . VL/HIV+ co-infections were represented in all four populations and no special cluster of VL/HIV+ strains was observed . This is in contrast to reports from Spain [37] , [66] and Portugal [38] , which found kDNA PCR-RFLP patterns , which were associated with the immune status ( HIV+ immuno-compromised and immuno-competent ) of the patients . This has been attributed to the existence of anthroponotic transmission cycles due to needle sharing among drug users . Whether these contradictory results are caused by the use of different DNA types , nuclear and kinetoplast minicircle DNA , remains to be elucidated . By increasing K in STRUCTURE analysis we observed in K = 5 a split of the non-MON-1 population , present also in the distance-based tree . Interestingly , all newly described zymodemes , found exclusively in HIV+/leishmaniasis co-infections [9] , [30] , [67] are members of the same subgroup . We also found that the majority of strains with these zymodemes were heterozygous suggesting a relationship between anthroponotic syringe transmission cycles and recombination events . Identical MLMT genotypes were found for strains isolated from different episodes in 10 out of the 12 relapse cases . For the two cases presenting different MLMT profiles in the respective episodes it is possible that the first infection was due to two strains , of which one has been eliminated during the first curse of treatment . Another possibility is re-infection with a new strain . Because microsatellite markers did not detect an individual MLMT fingerprint for each strain , re-infection can only be detected if strains from successive episodes have different genotypes . Otherwise , de-novo infection with an identical genotype circulating in the same focus cannot be excluded . RFLP analysis of minicircle kDNA has been successfully applied for monitoring VL outbreaks among intra-venous drug users and for differentiation between relapses and re-infections [13] , [14] . However , because of the multicopy nature of minicircle kDNA , the RFLP profiles are very complex , difficult to interpret and problematic for inter-laboratory comparisons . The stability of the pattern in the course of treatment is also still under question . Microsatellite marker stability during long term cultivation and animal passages has been confirmed ( data not shown ) . Nevertheless , MLMT and kDNA PCR-RFLP are the most discriminative methods available to date for strain fingerprinting [42] . In this study , some amount of gene flow was detected between and within all four populations . This became evident in the STRUCTURE plots where some strains could not be clearly assigned and had membership coefficients for more than one population and was attributed to the presence of alleles typical for more than one population either in homozygous or heterozygous combinations in single strains . Inbreeding coefficients and indices of association as a measure of multilocus linkage disequilibrium ( LD ) pointed to a predominantly clonal propagation particularly of all populations found for MON-1 , but also the non-MON-1 population , as previously suggested [68]–[72] . However , it does not rule out the possibility of occasional genetic recombination , as supported here by the occurrence of strains of potential mosaic and heterozygous genotypes , even between MON-1 and non-MON-1 populations . Two explanations are likely: ( 1 ) mixed infections , ( 2 ) hybrid strains/recombination . No indications for aneuploidy or polyploidy , asexual mechanisms leading to genetic polymorphism during reversion to the normal ploidy and having been described to occur in Leishmania [73] , were obtained in our study . By using highly discriminatory markers on 141 strains of L . infantum , among them 119 MON-1 strains , we tried to minimize the impact of statistical type II errors ( too few strains , poorly discriminating markers ) which probably have led to underestimation of recombination events in Leishmania in previous studies , as was suggested among others by Tibayrenc [69] , [74] . Significantly , all three sand fly isolates but only five of the many human isolates and a single canine isolate were of mixed ancestry . Whilst the Phlebotomus perniciosus isolate PT ( II ) 2 had for most of the markers alleles characteristic for both , MON-1 and non-MON-1 , in the second sand fly isolate PT7 ( I ) , the canine strain PT13 ( II ) and the human isolates PT8 ( II ) and PT8 ( I ) MON-1 type alleles dominated . Furthermore , the two sand fly isolates from Alto Douro PT7 ( I ) and PT6 ( I ) had a unique allele for marker Li 22–35 that differed significantly from the usual MON-1 alleles and rather resembled the range of non-MON-1 alleles . Leishmania hybrids have been reported in the literature , even in cloned parasites [75]–[79] . The mechanisms underlying hybrid formation are , however still unknown . The observation of a higher number of putative recombinant genotypes in the vector could suggest that recombination occurs in the vector . Alternatively , vectors could transmit strains that are infective , but normally do not cause disease and that are , therefore , not isolated from the vertebrate host . Preliminary data on sand fly isolates indicate that vectors might play a role in sustaining genetic diversity . Mixed infections as a basis for recombination events are also conceivable in HIV patients , especially when parasites are transmitted by needle sharing [79] . That most of the new L . infantum zymodemes , almost exclusively detected in HIV co-infected patients [9] , [24] , [30] , [37] , [67] , [80] , are heterozygous favours this hypothesis . Recently , a substantial number of heterozygous sites and mosaic genotypes were identified in strains of L . donovani and L . infantum [64] . These authors concluded that genetic exchange would be the most plausible explanation for their data and that the importance of recombination events in Leishmania has been underestimated so far , which is supported by the present MLMT data . In the present study , for the first time several epidemiological questions linked with the MON-1 zymodeme could be addressed using fast evolving microsatellites . These markers are hypervariable , genetically neutral and co-dominant and are therefore ideal to detect fine scale and recent genetic structure . Last but not least , the analyses could be performed directly from biological material with a high throughput . The results obtained in different laboratories are comparable and can be stored in a database . For a comprehensive understanding of L . infantum epidemiology strains from all endemic foci around the Mediterranean have to be studied , and a good sampling strategy has to be applied ( e . g . more sand fly isolates , other viscerotropic and dermotropic zymodemes , study of foci with significant enzymatic polymorphism etc . ) . The basis for such a study is the present work , which can now be completed by the data of all those mentioned strains . | Visceral leishmaniasis is caused by protozoan parasites of the genus Leishmania . This disease is a public health problem in countries bordering the Mediterranean , in China , and South America . Until now , isoenzyme analysis , a method with several advantages but also some limitations , is the gold standard for typing the causative agent L . infantum . We have developed a new method based on hypervariable DNA markers , the microsatellites . Its higher discriminatory power , genotype-based analysis , the possibility to use biological material instead of parasite cultures , and the fast analysis are the major improvements . We could demonstrate for the first time that there exist different geographically determined populations within the predominant zymodeme of L . infantum , which has important epidemiological implications . We also tested for relationships between genotype and clinical picture and/or host background . Leishmania is considered to reproduce mainly clonally; however , we found some indication for recombination in our study . Our work constitutes a solid basis for further population and epidemiological studies of L . infantum by completing the existing microsatellite database by analysing strains from other endemic foci . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"infectious",
"diseases/neglected",
"tropical",
"diseases",
"molecular",
"biology/molecular",
"evolution",
"public",
"health",
"and",
"epidemiology/epidemiology",
"public",
"health",
"and",
"epidemiology/infectious",
"diseases",
"microbiology/parasitology",
"infectious",
"disease... | 2008 | Differentiation and Gene Flow among European Populations of Leishmania infantum MON-1 |
Large-scale conformational changes in proteins involve barrier-crossing transitions on the complex free energy surfaces of high-dimensional space . Such rare events cannot be efficiently captured by conventional molecular dynamics simulations . Here we show that , by combining the on-the-fly string method and the multi-state Bennett acceptance ratio ( MBAR ) method , the free energy profile of a conformational transition pathway in Escherichia coli adenylate kinase can be characterized in a high-dimensional space . The minimum free energy paths of the conformational transitions in adenylate kinase were explored by the on-the-fly string method in 20-dimensional space spanned by the 20 largest-amplitude principal modes , and the free energy and various kinds of average physical quantities along the pathways were successfully evaluated by the MBAR method . The influence of ligand binding on the pathways was characterized in terms of rigid-body motions of the lid-shaped ATP-binding domain ( LID ) and the AMP-binding ( AMPbd ) domains . It was found that the LID domain was able to partially close without the ligand , while the closure of the AMPbd domain required the ligand binding . The transition state ensemble of the ligand bound form was identified as those structures characterized by highly specific binding of the ligand to the AMPbd domain , and was validated by unrestrained MD simulations . It was also found that complete closure of the LID domain required the dehydration of solvents around the P-loop . These findings suggest that the interplay of the two different types of domain motion is an essential feature in the conformational transition of the enzyme .
Biological functions of proteins are mediated by dynamical processes occurring on complex energy landscapes [1] . These processes frequently involve large conformational transitions between two or more metastable states , induced by an external perturbation such as ligand binding [2] . Time scales of the conformational transition are frequently of order microseconds to seconds . To characterize such slow events in molecular dynamics ( MD ) trajectories , the free energy profile or the potential of mean force ( PMF ) along a reaction coordinate must be identified . In particular , the identification of the transition state ensemble ( TSE ) enables the barrier-height to be evaluated , and the correct kinetics would be reproduced if there is only a single dominant barrier . However , for proteins with many degrees of freedom , finding an adequate reaction coordinate and identifying the TSE is a challenging task placing high demands on computational resources . The finite-temperature string method [3] , [4] , and the on-the-fly string method [5] find a minimum free energy path ( MFEP ) from a high-dimensional space . Given a set of collective variables describing a conformational change , the MFEP is defined as the maximum likelihood path along the collective variables . The MFEP is expected to lie on the center of reactive trajectories and contains only important transitional motions [4] . Furthermore , since the MFEP approximately orthogonally intersects the isocommittor surfaces ( the surfaces of constant committor probability in the original space ) [4] , the TSE can be identified as the intersection with the isocommittor surface with probability of committing to the product ( or the reactant ) = 1/2 . The methods and MFEP concepts have been applied to various molecular systems [4]–[6] including protein conformational changes [7]–[9] . With regard to high-dimensional systems like proteins , the quality of the MFEP ( whether it satisfies the above-mentioned properties ) is particularly sensitive to choice of the collective variables . The collective variables should be selected such that their degrees of freedom are few enough to ensure a smooth free energy surface; at the same time they should be sufficiently many to approximate the committor probability [4] , [9] . To resolve these contrary requirements , effective dimensional reduction is required . Large conformational transitions of proteins , frequently dominated by their domain motions , can be well approximated by a small number of large-amplitude principal modes [2] , [10] . This suggests that the use of the principal components may be the best choice for approximating the committor probability with the fewest number of variables for such large conformational transitions involving domain motions . A further advantage is the smoothness of the free energy landscape in the space of the large-amplitude principal components . If the curvature of the MFEP is large , the MFEP may provide a poor approximation to the isocommittor surface since the flux can occur between non-adjacent structures along the path [9] . The selection of the large-amplitude principal components as the collective variables would maintain the curvature of the MFEP sufficiently small . Here , we conducted preliminary MD simulations around the two terminal structures of the transition and performed a principal component analysis to obtain the principal components ( see Materials and Methods for details ) . Following selection of a suitable MFEP , determination of the PMF and characterization of the physical quantities along the MFEP are needed to understand an in-depth mechanism of the transition . Although the finite-temperature string method yields a rigorous estimate of the gradient of the PMF under a large coupling constant with the collective variables [3]–[5] ( see Materials and Methods ) , errors in the estimates of the gradients and in the tangential directions of the pathway tend to accumulate during the integration process . To accurately quantify the PMF and the averages of various physical quantities in a multi-dimensional space , we utilized another statistical method , the multi-state Bennett acceptance ratio ( MBAR ) method [11] , which provides optimal estimates of free energy and other average quantities along the MFEP . Here , we applied the above proposed methods to the conformational change in Escherichia coli adenylate kinase ( AK ) , the best-studied of enzymes exhibiting a large conformational transition [12]–[23] . AK is a ubiquitous monomeric enzyme that regulates cellular energy homeostasis by catalyzing the reversible phosphoryl transfer reaction: ATP+AMP↔2ADP . According to the analysis of the crystal structures by the domain motion analysis program DynDom [24] , AK is composed of three relatively rigid domains ( Figure 1 ) ; the central domain ( CORE: residues 1–29 , 68–117 , and 161–214 ) , an AMP-binding domain ( AMPbd: 30–67 ) , and a lid-shaped ATP-binding domain ( LID: 118–167 ) . Inspection of the crystal structures suggests that , upon ligand binding , the enzyme undergoes a transition from the inactive open form to the catalytically competent closed structure [25] ( Figure 1 ) . This transition is mediated by large-scale closure motions of the LID and AMPbd domains insulating the substrates from the water environment , while occluding some catalytically relevant water molecules . The ATP phosphates are bound to the enzyme through the P-loop ( residues 7–13 ) , a widely-distributed ATP-binding motif . The interplay between AK's dynamics and function has been the subject of several experimental studies . 15N NMR spin relaxation studies have revealed that the LID and AMPbd domains fluctuate on nanosecond timescales while the CORE domain undergoes picosecond fluctuations [12] , [13] . The motions of these hinge regions are highly correlated with enzyme activity [14] . In particular , the opening of the LID domain , responsible for product release , is thought to be the rate-limiting step of the catalytic turnover [14] . Recent single-molecule Förster resonance energy transfer ( FRET ) experiments have revealed that the closed and open conformations of AK exist in dynamic equilibrium even with no ligand present [15] , [16] , and that the ligand's presence merely changes the populations of open and closed conformations . This behavior is reminiscent of the population-shift mechanism [26] rather than the induced-fit model [27] , in which structural transitions occur only after ligand binding . The population-shift like behaviour of AK has also been supported by simulation studies [17]–[20] . Lou and Cukier [17] , Kubitzki and de Groot [18] , and Beckstein et al . [19] employed various enhanced sampling methods to simulate ligand-free AK transitions . Arora and Brooks [20] applied the nudged elastic band method in the pathway search for both ligand-free and ligand-bound forms . These studies showed that , while the ligand-free form samples conformations near the closed structure [17]–[20] , ligand binding is required to stabilize the closed structure [20] . Despite the success of these studies based on all-atom level models , atomistic details of the transition pathways , including the structures around the TSE , have not been fully captured yet . In this study , we successfully evaluated the MFEP for both ligand-free and ligand-bound forms of AK using the on-the-fly string method , and calculated the PMF and the averages of various physical quantities using the MBAR method . Our analysis elucidates an in-depth mechanism of the conformational transition of AK .
The MFEPs for apo and holo-AKs , and their PMFs , were obtained from the string method and the MBAR method , respectively ( see Videos S1 and S2 ) . The MFEPs were calculated using the same 20 principal components selected for the collective variables . The holo-AK calculations were undertaken with the bisubstrate analog inhibitor ( Ap5A ) as the bound ligand without imposing any restraint on the ligand . Figures 2A and 2B show the MBAR estimates of the PMFs along the images of the MFEP ( the converged string at 12 ns in Figures 2A and 2B ) for apo and holo-AK , respectively . Here , the images on the string are numbered from the open ( ; PDBid: 4ake [28] ) to the closed conformation ( ; PDBid: 1ake [29] ) . These terminal images were fixed during the simulations to enable sampling of the conformations around the crystal structures . In the figures , the convergence of the PMF in the string method process is clearly seen in both systems . Convergence was also confirmed by the error estimates ( Figure S1 ) , and by the root-mean-square displacement ( RMSD ) of the string from its initial path ( Figure S2 ) . The PMF along the MFEP reveals a broad potential well on the open-side conformations of apo-AK , suggesting that the open form of AK is highly flexible [20] . This broad well is divided into two regions , the fully open ( ) and partially closed states ( , encircled ) by a small PMF barrier . In holo-AK ( Figure 2B ) , the MFEP exhibits a single substantial free energy barrier ( ) between the open and closed states , which does not appear in the initial path . This barrier will be identified as the transition state below . It is shown in the PMF along the MFEP that the closed form ( tightly binding the ligand ) is much more stable than the open form with loose binding to the ligand ( large fluctuations of the ligand will be shown later ) . To characterize the MFEP in terms of the domain motions , the MFEP was projected onto a space defined by two distances from the CORE domain , the distance to the LID domain and the distance to the AMPbd domain ( the distance between the mass centers of atoms for the two domains; Figures 2C and 2D ) . The PMF was also projected onto this space . The comparison of the two figures shows that ligand binding changes the energy landscape of AK , suggestive that this is not a simple population-shift mechanism . In apo-AK , the motions of the LID and AMPbd domains are weakly correlated , reflecting the zipper-like interactions on the LID-AMPbd interface [19] . The MFEP clearly indicates that the fully closed conformation ( ) involves the closure of the LID domain followed by the closure of the AMPbd domain . The higher flexibility of the LID domain has been reported in previous studies [17] , [19] , [20] . In holo-AK , the pathway can be described by two successive scenarios , that is , the LID-first-closing followed by the AMPbd-first-closing . In the open state ( ) , the MFEP is similar to that of apo-AK , revealing that LID closure occurs first . In the closed state ( ) , however , the AMPbd closure precedes the LID closure . This series of the domain movements was also identified by the domain motion analysis program DynDom [24] ( Figure S3 ) . It is known in the string method that the convergence of the pathway is dependent on the initial path . In order to check whether the MFEP obtained here is dependent on the initial path or not , we conducted another set of the calculations for apo-AK by using a different initial path , which has an AMPbd-first-closing pathway , opposed to the LID-first-closing pathway shown above . If the LID and AMPbd domains move independently of each other , it is expected that LID-first-closing and AMPbd-first-closing pathways are equally stable . Despite this initial setup , however , our calculation again showed the convergence toward the LID-first-closing pathway ( see Figure S4 ) . As described above , this tendency of the pathways would be due to the reflection of the highly flexible nature of the LID domain . Furthermore , in order to check whether the samples around the MFEP are consistent with the experiments , we compared the PMF as a function of the distance between the Cα atoms of Lys145 and Ile52 with the results of the single-molecule FRET experiment by Kern et al . [16] ( see Figure S5 ) . The PMF was calculated by using the samples obtained by the umbrella samplings around the MFEP . In the figure , the stable regions of the PMF for holo-AK are highly skewed toward the closed form , and some population toward the partially closed form was also observed even for apo-AK , which is consistent with the histogram of the FRET efficiency [16] . To more clearly illustrate the energetics along the MFEP in terms of the domain motions , we separately plot the PMF as a function of the two inter-domain distances defined above ( Figures 3A and 3B ) . We observe that the PMF of apo-AK has a double-well profile for the LID-CORE distance ( indicated by the blue line in Figure 3A ) , whereas the PMF in terms of the AMPbd-CORE distance is characterized by a single-well ( Figure 3B ) . The single-molecule FRET experiments monitoring the distances between specific residue pairs involving the LID domain ( LID-CORE ( Ala127-Ala194 ) [15] and LID-AMPbd ( Lys145- Ile52 ) [16] ) revealed the presence of double-well profiles in the ligand-free form . On the other hand , an electron transfer experiment probing the distance between the AMPbd and CORE domains ( Ala55-Val169 ) [30] showed only that the distance between the two domains decreased upon ligand binding . Considering the PMF profiles in the context of these experimental results , we suggest that the partially closed state ( ) in apo-AK ( Figure 2A ) can be ascribed to the LID-CORE interactions but not to the AMPbd-CORE interactions . To elucidate the origin of the stability of the partially closed state , we monitored the root mean square fluctuations ( RMSF ) of the atoms along the MFEP ( see Materials and Methods for details ) . Figures 3C and 3D show the RMSF along the MFEP for apo and holo-AK , respectively . In apo-AK ( Figure 3C ) , large fluctuations occur in the partially closed state ( ) around the LID-CORE hinge regions ( residue 110–120 , and 130–140 ) and the P-loop ( residue 10–20 ) . It has been proposed , in the studies of AK using coarse-grained models , that “cracking” or local unfolding occurs due to localized strain energy , and that the strained regions reside in the LID-CORE hinge and in the P-loop [21] , [23] . Our simulation using the all-atom model confirmed the existence of “cracking” in the partially closed state , and provided an atomically detailed picture of this phenomenon . The average structures around the partially closed state revealed that , in the open state , a highly stable Asp118-Lys136 salt bridge is broken by the strain induced by closing motion around ( Figure S6A ) . This salt bridge has been previously proposed to stabilize the open state while imparting a high enthalpic penalty to the closed state [18] . Breakage of the salt bridge releases the local strain and the accompanying increases in fluctuation may provide compensatory entropy to stabilize the partially closed state . A similar partially closed state of the LID domain was also found by the work of Lou and Cukier [31] in which they performed all-atom MD simulation of apo-AK at high temperature ( 500 K ) condition . In holo-AK , both of the LID-CORE and AMPbd-CORE distances exhibit double-well profiles ( indicated by the red lines in Figures 3A and 3B ) , separating the closed from the open state . The breakage of the 118–136 salt bridge at around is not accompanied by “cracking” of the hinge region ( Figure 3D ) . Instead , the hinge region is stabilized by binding of ATP ribose to Arg119 and His134 ( Figure S6B ) , leading to a smooth closure of the LID domain . This suggests that one role of the salt bridge breakage is rearrangement of the molecular interactions to accommodate ATP-binding [32] . P-loop fluctuations are also suppressed in holo-AK ( Figure 3D ) . Consistent with our findings , reduced backbone flexibilities in the presence of Ap5A were reported in the above-mentioned NMR study [13] . The origin of the double-well profile in holo-AK was investigated via the ligand-protein interactions . The motion of the ligand along the MFEP was firstly analyzed by focusing on the AMP adenine dynamics , since the release of the AMP moiety from the AMP-binding pocket was observed in the open state . It is again emphasized that the ligand is completely free from any restraint during the simulations . PCA was performed for the three-dimensional Cartesian coordinates of the center of mass of AMP adenine , and the coordinates were projected onto the resultant 1st PC in Figure 4A . The AMP adenine is observed to move as much as 10 Å in the open state ( ; Figure 4B ) , while it is confined to a narrow region of width 1–2 Å ( the binding pocket ) in the closed state ( ) . Such a reduction of the accessible space of the AMP adenine might generate a drastic decrease in entropy or an increase in the PMF barrier of the open-to-closed transition . Furthermore , close inspection of the PMF surface reveals the existence of a misbinding event at ( Figure 4B ) , in which the AMP ribose misbinds to Asp84 in the CORE domain , and is prevented from entering the AMP-binding pocket . This event further increases the barrier-height of the transition . The MFEP revealed that AMP adenine enters the AMP-binding pocket around , as indicated by a rapid decrease in the accessible area ( Figure 4A ) . This event is well correlated with the position of the PMF barrier along the MFEP ( Figure 2B ) . This coincidence between the binding process and the domain closure suggests that the two processes are closely coupled . Before analyzing the situation in detail , however , it is necessary to assess whether the observed PMF barrier around ( Figure 2B ) corresponds to a TSE , because the PMF barrier is not necessarily a signature of dynamical bottleneck in high-dimensional systems [33] . TSE validation is usually performed with a committor test [4] , [7] , [9] , [33] . In principle , the committor test launches unbiased MD simulations from structures chosen randomly from the barrier region , and tests whether the resultant trajectories reach the product state with probability 1/2 . Here , since limited computational resources precluded execution of a full committor test , 40 unbiased MD simulations of 10 ns were initiated from each of , 33 or 34 , a total of 120 simulations or 1 . 2 , and the distributions of the final structures after 10 ns were monitored [9] . Figure 5A shows the binned distributions of the final structures assigned by index of the nearest MFEP image ( the blue bars ) . When the simulations were initiated from the image at ( ) , the distribution biases to the open form-side ( the closed form-side ) relative to the initial structures . On the other hand , when starting from the image at , the distribution is roughly symmetric around the initial structures . This result suggests that the TSE is located at . In other words , it was validated that the TSE was successfully captured in the MFEP , and at the same time , the collective variables were good enough to describe the transition . A close inspection of the structures around the PMF barrier supported our insufficient committer test and revealed the mechanism of the ligand-induced domain closure . Figure 5B shows the hydrogen bond ( H-bond ) patterns between the ligand and the protein observed in the average structures at ( before the TSE ) and ( after the TSE ) . At , Thr31:OG1 ( AMPbd ) forms an H-bond with N7 of AMP adenine , and Gly85:O ( CORE ) forms one with adenine N6 . These two H-bonds mediate the hinge bending of the AMPbd-CORE domains . In addition , the H-bond between Gly85:O and adenine N6 helps the enzyme to distinguish between AMP and GMP; GMP lacks an NH2 group in the corresponding position of AMP [34] . This means that the specificity of AMP-binding operates at an early stage of the ligand binding process . At , the AMPbd-CORE distance becomes smaller than that at , which allows the formation of 3 additional H-bonds with the ligand: Gln92:OE1 ( CORE ) and adenine N6 , Lys57:O ( AMPbd ) and the ribose O2 , and Arg88:NH1 ( CORE ) and O1 of AMP . The resulting rapid enthalpy decrease stabilizes the closed conformation . Gln92:OE1 is also important in establishing AMP specificity; GMP lacks the counterpart atom , adenine N6 . The strictly conserved Arg88 residue is known to be crucial for positioning AMP so as to suitably receive a phosphate group from ATP [35] . With regard to the AMPbd closure , our result suggests that Arg88 ( CORE ) , in conjunction with Lys57 ( AMPbd ) , works to block adenine release from the exit channel and to further compact the AMPbd-CORE domains . A remaining question is how closure of the LID domain follows that of the AMPbd domain . Unlike the AMP-binding pocket , the ATP-binding sites , including the P-loop , are surrounded by charged residues , which attract interfacial water molecules . Upon LID closure , most of these water molecules will be dehydrated from the enzyme , but some may remain occluded . To characterize the behaviors of these water molecules , the 3D distribution function of their oxygen and hydrogen constituents were calculated along the MFEP using the MBAR method ( see Materials and Methods ) . Figures 6A , 6B , and 6C display the isosurface representations of the 3D distribution functions around the P-loop at , 41 , and 42 , respectively . The surfaces show the areas in which the atoms are distributed four times as probably as in the bulk phase . At , the ATP phosphates are not yet bound to the P-loop because an occluded water molecule ( encircled ) is wedged between the phosphate and the P-loop , inhibiting binding of ATP and and bending of the side-chain of “invariant lysine” ( Lys13 ) , a residue that plays a critical role in orienting the phosphates to the proper catalytic position [36] . This occluded water molecule may correspond to that found in the crystal structure of apo-AK ( PDBid: 4ake ) ( Figure 6D , encircled ) . Figures 6B and 6C clearly demonstrate that , upon removal of this water molecule , the ATP phosphates begin binding to the P-loop . These observations were confirmed by plots of the PMF surface mapped onto a space defined by the LID-CORE distance versus the index of image ( Figure 6F ) , which shows that the PMF decreases discontinuously upon dehydration followed by LID domain closure . Interestingly , compared with the crystal structure ( PDBid: 1ake ) ( Figure 6E ) , the position of the ATP moiety is shifted to the AMP side by one monophosphate unit . This may be a consequence of early binding of the AMP moiety . At a later stage ( around ) , this mismatch was corrected to form the same binding mode as observed in the crystal structure . This reformation of the binding mode may be induced by the tight binding of ATP adenine to the LID-CORE domains , and will not occur in the real enzymatic system containing ATP and AMP instead of the bisubstrate analog inhibitor , Ap5A .
In this study , we have applied the on-the-fly string method [5] and the MBAR method [11] to the conformational change of an enzyme , adenylate kinase , and successfully obtained the MFEP ( Figures 2A and 2B ) . The MFEP yielded a coarse-grained description of the conformational transitions in the domain motion space ( Figures 2C and 2D ) . At the same time , the atomistic-level characterization of the physical events along the MFEP provided a structural basis for the ligand-binding and the domain motions ( Figures 3–6 ) . This kind of multiscale approach used here is expected to be useful generally for complex biomolecules since the full space sampling can be avoided in an efficient manner . We have shown that in the TSE of holo-AK , the conformational transition is coupled to highly specific binding of the AMP moiety . Our results have been validated by unbiased MD simulations . The mechanism of the AMPbd domain closure is consistent with that proposed by the induced-fit model ( Figure S7A ) , and follows a process similar to that of protein kinase A , previously investigated by a coarse-grained model [32]: ( i ) the insertion of the ligand into the binding cleft initially compacts the system; ( ii ) additional contacts between the ligand and non-hinge region further compact the system . The closure of the LID domain is more complicated ( Figure S7B ) . It was shown that apo-AK can exist in a partially closed state , stabilized by the “cracking” of the LID-CORE hinge and the P-loop , even with no ligand present . The cracking of the hinge region enables rearrangement of molecular interactions for ATP-binding , which induces a smooth bending of the hinge . Along with the LID closure , ATP is conveyed into the P-loop , with removal of an occluded water molecule . The closure of the LID domain follows the “population-shift followed by induced-fit” scenario discussed in Ref . [37] , in which a transient local minimum is shifted toward the closed conformation upon ligand binding . This two-step process of the LID domain closure is similar to the two-step mechanism reported in recent simulation studies of the Lysine- , Arginine- , Ornithine-binding ( LAO ) protein [38] and the maltose binding protein [39] . In holo-AK , AMPbd domain closure occurs early ( at ) , while the LID domain closes at later stages ( ) . An interesting question is whether an alternative pathway is possible in the presence of the real ligands ( ATP and AMP ) instead of Ap5A . Ap5A artificially restrains the distance between the ATP and AMP moieties . During the process with real ligands , the dynamics of the LID and AMPbd domains is expected to be less correlated . Nevertheless , for full closing of the LID domain , we conjecture that the AMPbd domain should be closed first , enabling the interactions on the LID-AMPbd interface to drive the dehydration around the P-loop . This suggests that full recognition of ATP by the LID-CORE domains occurs at a later stage of the conformational transition . This conjecture may be related to the lower specificity of E . coli AK for ATP compared with AMP [34] . Nonspecific AMP-binding to the LID domain has previously been suggested to explain the observed AMP-mediated inhibition of E . coli AK at high AMP concentrations [40] . A missing ingredient in the present study is the quantitative decomposition of the free energy in each event , such as the ligand binding and the interactions on the LID-AMPbd interface . For enhanced understanding of the conformational change , our methods could be complemented by the alchemical approach [41] . Varying the chemical compositions of the system during the conformational change would enable us to elucidate the effects of ligand binding , cracking , and dehydration in a more direct manner .
We prepared three systems from the following initial structures: ( i ) “apo-open system” , X-ray crystal structure of the open-form without ligand ( PDBid: 4ake [28] ) , ( ii ) “holo-closed system” , crystal structure of closed-form with Ap5A ( PDBid: 1ake [29] ) , ( iii ) “apo-closed system” , structure created by removing Ap5A from the holo-closed system . The protonation states of the titratable groups at pH 7 were assigned by PROPKA [42] , implemented in the PDB2PQR program package [43] , [44] . The apo-open and apo-closed systems yielded identical assignments , which were used also for the holo-closed system . These systems were solvated in a periodic boundary box of water molecules using the LEaP module of the AMBER Tools ( version 1 . 4 ) [45] . A padding distance of 12 Å from the protein surface was used for the apo-open system . For the apo-closed and holo-closed systems , a longer padding distance of 20 Å was used to avoid interactions with periodic images during the closed-to-open transition . Two Na+ ions were added to neutralize the closed-apo and open-apo systems , while seven Na+ ions were required to neutralize the closed-holo system . The systems were equilibrated under the NVT condition at 300 K by the following procedure: First , the positions of solvent molecules and hydrogen atoms of the protein ( and Ap5A ) were relaxed by 1 , 000 step minimization with restraint of non-hydrogen atoms . Under the same restraints , the system was gradually heated up to 300 K over 200 ps , followed by 200 ps MD simulation under the NVT condition at 300 K while gradually decreasing the restraint forces to zero , but keeping the restraints on atoms needed in the string method . The system was further equilibrated by 200 ps MD simulation under the NPT condition ( 1 atm and 300 K ) , adjusting the density of the water environment to an appropriate level . The ensemble was finally switched back to NVT , and subjected to additional 200 ps simulation at 300 K , maintaining the restraints . The equilibration process was conducted using the Sander module of Amber 10 [45] , with the AMBER FF03 force field [46] for the protein , and TIP3P for water molecules [47] . The parameters for Ap5A were generated by the Antechamber module of AMBER Tools ( version 1 . 4 ) [45] using the AM1-BCC charge model and the general AMBER Force Field ( GAFF ) [48] . Covalent bonds involving hydrogen atoms were constrained by the SHAKE algorithm [49] with 2 fs integration time step . Long-range electrostatic interactions were evaluated by the particle mesh Ewald method [50] with a real-space cutoff of 8 Å . The Langevin thermostat ( collision frequency 1 ps−1 ) was used for the temperature control . The production runs , including the targeted MD , the on-the-fly string method , the umbrella sampling , and the committor test , were performed with our class library code for multicopy and multiscale MD simulations ( which will soon be available ) [T . Terada et al . , unpublished] , using the same parameter set described above ( unless otherwise noted ) . Protein structures and the isosurfaces of solvent density were drawn with PyMOL ( Version 1 . 3 , Schrödinger , LLC ) . The calculations were performed using the RIKEN Integrated Cluster of Clusters ( RICC ) facility . It has been shown that normal modes or principal modes provide a suitable basis set for representing domain motions of proteins [2] , [10] . In particular , it has been argued that the conformational change in AK can be captured by a set of principal modes of apo-AK [31] , [51] . In this study , we have defined the collective variables for the on-the-fly string method using the principal components of apo-AK . The PCA was carried out in the following manner: After the equilibration process , 3 ns MD simulations were executed at 300 K without restraint for both apo-open and apo-closed systems . The obtained MD snapshots from both systems were combined in a single PCA [52] , removing the external contributions by iteratively superimposing them onto the average coordinates [53] , [54] . The PCA was then conducted for the Cartesian coordinates of the atoms . It was found that the first principal mode representing the largest-amplitude merely represents the difference between the open and closed conformations . The fluctuations in the two structures were expressed in the principal modes of smaller amplitudes . The cumulative contributions of these modes ( ignoring the first ) are shown in Figure S8 . As expected , the principal modes represent the collective motions of the LID and AMPbd domains ( Figure S9 ) . The first 20 principal components ( 82% cumulative contribution , ignoring that of the first ) were adopted as the collective variable of the string method . These components were sufficient to describe the motions of three domains in AK for which at least degrees of freedom are required in the rigid-body approximation . The additional eight degrees of freedom were included as a buffer for possible errors in the estimation of the principal modes . The sum of the canonical correlation coefficients between the two sets of the 20 principal components , one calculated using the samples of the first half ( 0–1 . 5 ns ) snapshots and the other using the last half ( 1 . 5–3 ns ) snapshots , was 11 . 8 ( ∼12 ) , suggesting that the subspace of the domain motions was converged in 3 ns simulation . The initial paths for the string method were generated using the targeted MD ( TMD ) simulations [55] for apo and holo-AK . Whereas a previous study had constrained the one-dimensional RMSD value from the starting structure to the target structure [55] , in this study , we imposed 20-dimensional harmonic restraints ( with spring constant 10 kcal/mol/Å2 ) along the linear interpolation between the open and closed crystal structures in the 20 principal component space [8] , [9] . Starting from the closed conformation , TMD simulations of about 1 ns for the apo-closed and holo-closed systems were conducted with the open conformation as target . By imposing this direction , from the closed conformation to the open conformation , unfavorable steric crashes can be avoided . During the TMD , the Eckart condition [56] was imposed on the atoms . The initial path for the string was obtained as 65 structures on the TMD trajectory by equally partitioning the trajectory from the open to closed conformation . These 65 structures were equilibrated by 400 ps restrained MD simulations . Throughout this stage , the AMP moiety of the Ap5A separated from the AMP-binding pocket in the open state of the holo-closed system . However , the ATP phosphates remained bound to the correct side chains of the conserved arginines ( Arg123 , Arg156 , and Arg167 ) , and the ATP ribose was bound to the backbone of Arg119 . The on-the-fly string method [5] , a variant of the original finite-temperature string method [3] , is a powerful tool for finding the MFEP from high-dimensional free energy surfaces . The MFEP is searched on the free energy surface associated with M collective variables ( or M ( = 20 ) principal components of atoms in this case ) , , with x being the Cartesian coordinates of the entire system . The following equations are solved simultaneously: ( 1 ) ( 2 ) Equation 1 describes the time evolution of the string in the M-dimensional collective variable space , at position s on the string and at time t in the simulation . Equation 2 is a standard MD simulation with restraint of the collective variable around . The parameter is the “friction” coefficient controlling the dynamics of , is the mass of atom , and is a spring constant . For a proper choice of and , the string is driven by a negative gradient on the free energy surface and is expected to converge to the MFEP [5] . Here , we chose a strong spring constant of ( kcal/mol/Å2 ) to reduce the statistical bias in the estimation of the free energy gradient while maintaining the numerical stability of the simulation . Also , ( kcal s/mol/Å ) was chosen to reduce the statistical fluctuations of by slowing down the dynamics of compared that of . The position s on the string was discretized by 65 images , ( ) , numbered from the open to the closed conformation . The terminal images were fixed to sample around the open and closed crystal structures . The term “constraint” in Equation 1 indicates that the distances between adjacent images , , are kept equal through all , or [4] , [5] . It is noted that the principal components are based on the unitary transformation from the coordinates of the atoms to . Thus , the metric tensor appearing in the original formulation [5] due to the curvilinear nature of the collective variables can be reduced from an matrix as a function of x to a constant diagonal form . At the same time , two sets of the Cartesian coordinates of the entire system , and , required for the statistical independence of the metric tensor in the original formulation [5] , can also be reduced to the single variable . A further advantage of the principal components approach is that large-amplitude principal components may capture a smooth free energy surface and thus avoid the trapping of string images in local minima . To accurately quantify the PMF and the averages of various physical quantities , the MBAR method [11] , [57] was employed . The standard weighted histogram analysis method [58] requires an extremely large storage space for the grid points in the 20-dimensional space . The MBAR method requires no grid points and hence naturally circumvents this problem . Other advantages of the MBAR method are that the estimator is asymptotically unbiased and yields minimal variance , and that the statistical error can be estimated under the large sample limit [11] . In the following , we briefly summarize the umbrella sampling and the MBAR method . The umbrella sampling was conducted around the image of the string obtained as the MFEP or at t = 12 ns . For comparison , umbrella sampling was also performed around the string images at t = 0 , 2 , 4 , 6 , 8 , and 10 ns , generating a total of ensembles . The window potential used for the sampling was the harmonic restraint imposed on the 20 principal components . Here , we chose a weak spring constant of kcal/mol/Å2 to obtain sufficient phase space overlaps . Following a 200 ps equilibration , umbrella sampling was performed for 9 ns around each image of the MFEP . For the strings other than the MFEP , simulation time was limited to 0 . 8 ns to reduce the simulation run-time . For uncorrelated samples with coordinates obtained from the umbrella system ( ) , the MBAR equation defines an estimator of the free energy of the umbrella system up to an additive constant [11]: ( 3 ) where ( in units of ) with u being the potential function of the system . This coupled nonlinear equation was solved by a Newton-Raphson solver [11] . The samples were subsampled prior to calculation , based on the autocorrelation function of potential energy [59] . Calculations were performed in MATLAB ( The MathWorks , Inc . ) , confirming that the calculation was compatible with the Python implementation of the MBAR method ( https://simtk . org/home/pymbar ) [11] . Having obtained the free energy estimations , the equilibrium expectation of a mechanical observable under the unrestrained ensemble can be computed as a ratio of partition functions , and : ( 4 ) ( 5 ) where the double index of the coordinates , , is aggregated into a single index n ( ) , since the explicit notation of the umbrella windows ( from which samples were taken ) is not necessary in the remaining calculations . Equations 4 and 5 yield an estimator of the expectation : ( 6 ) where is the weight of the sample for the unrestrained equilibrium . The probability is now assigned to an arbitrary region , designated cell B , by ( 7 ) where is the indicator function which takes the value 1 if the system is in the cell B and 0 otherwise . Knowing , the PMF of cell B is given up to an additive constant in units of : ( 8 ) where is the relative volume of cell B , necessary to correct for non-uniform cell size . To evaluate the PMF around the image on the string in the 20-dimensional principal component space ( Figures 2A and 2B ) , the cells were assigned to each image of the string using the Voronoi tessellation [9] , [60] , [61]: ( 9 ) where is the Euclidean distance in the 20-dimensional space . It is noted that metric tensor is not required in this definition since our collective variables are defined with linear coordinates [9] , [60] , [61] ) . During the calculation , since the explicit determination of the cell boundaries in 20-dimensional space is not feasible , the samples were assigned by finding their nearest images . Samples located far from any image were excluded from the calculation as outliers by imposing a cutoff condition on Equation 9 , i . e . , samples for which 6 Å for all were eliminated . Increasing the cutoff distance to 7 Å did not affect the results . The relative volume in Equation 8 was approximated by the variance of the samples in cell , or the Voronoi cell with the cutoff condition was approximated by an ellipsoid . It was confirmed that the volume correction didn't give qualitative difference in the PMF due to the cutoff condition ( see Figure S10 ) . If the cells are spherical , a 10% error in estimating the length of the axis for each dimension roughly leads to a ( ) error , under the assumption that the errors are statistically independent of each other . As another way to estimate the PMF in a high-dimensional space without the Voronoi tessellation , one could perform the kernel density estimations using analytically tractable kernel functions , e . g . , hyper-spheres ( cutoff conditions ) , or Gaussian kernels . Since the volume of each kernel can be analytically calculated , this approach is free from the problem of the estimation errors in the volumes . In the strings before attaining the MFEP ( t = 0–10 ns ) , the distribution formed by the umbrella sampling tended to be biased to the minima of free energy surface , or to the MFEP . For the purpose of calculating accurate PMF for the non-converged strings ( Figures 2 and S1 ) , it was necessary that the Voronoi cells were defined by using all images generated from all strings . On the other hand , the average quantities along MFEP ( Figures 3–6 and S2 ) were able to be correctly evaluated by using only the samples generated from the MFEP . The RMSF values along the MFEP ( Figures 3C and 3D ) were evaluated with the average structure in cell calculated by ( 10 ) The average coordinates calculated by Equation 10 were also utilized in the various investigations ( Figures 5B , 6 , S3 , and S4 ) . The RMSF value in cell and atom i , , ( Figures 3C , and 3D ) was then evaluated by ( 11 ) where are the coordinates of atom i and sample n . The distribution function of oxygens and hydrogens of water molecules ( Figures 6A , 6B , and 6C ) was evaluated as follows: First , the instantaneous densities of water atoms at grid l in cell , , were calculated in 3D grids ( ) with grid size about 0 . 5 Å . The averaged density , , was then calculated by ( 12 ) In principle , the committor test launches unbiased MD simulations from structures chosen randomly from the barrier region and examines whether the resultant trajectories terminate in either the reactant or the product state with equal probability [7] , [9] , [33] . Here , since such a full committor test is not feasible due to the large system size , 40 unbiased MD simulations of 10 ns were started from , 33 or 34 , that is , 120 simulations or 1 . 2 in total , and the distributions of the final structures after 10 ns were monitored [9] . Ten initial coordinates were taken randomly from the snapshots of the umbrella sampling belonging to each of Voronoi cells of , 33 , or 34 . Forty unrestrained MD simulations were launched from each Voronoi cell using these ten coordinates , each assigned four sets of momenta generated from the Maxwell-Boltzmann distribution . The Voronoi cell index was assigned to the coordinates after 10 ns . A histogram of the indices is plotted in Figure 5A . | Conformational transitions of proteins have been postulated to play a central role in various protein functions such as catalysis , allosteric regulation , and signal transduction . Among these , the relation between enzymatic catalysis and dynamics has been particularly well-studied . The target molecule in this study , adenylate kinase from Escherichia coli , exists in an open state which allows binding of its substrates ( ATP and AMP ) , and a closed state in which catalytic reaction occurs . In this molecular simulation study , we have elucidated the atomic details of the conformational transition between the open and the closed states . A combined use of the path search method and the free energy calculation method enabled the transition pathways to be traced in atomic detail on micro- to millisecond time scales . Our simulations revealed that two ligand molecules , AMP and ATP , play a distinctive role in the transition scenario . The specific binding of AMP into the hinge region occurs first and creates a bottleneck in the transition . ATP-binding , which requires the dehydration of an occluded water molecule , is completed at a later stage of the transition . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"computational",
"chemistry",
"molecular",
"dynamics",
"biophysic",
"al",
"simulations",
"chemistry",
"biology",
"computational",
"biology"
] | 2012 | Minimum Free Energy Path of Ligand-Induced Transition in Adenylate Kinase |
The insula , particularly its posterior portion , is often regarded as a primary cortex for pain . However , this interpretation is largely based on reverse inference , and a specific involvement of the insula in pain has never been demonstrated . Taking advantage of the high spatiotemporal resolution of direct intracerebral recordings , we investigated whether the human insula exhibits local field potentials ( LFPs ) specific for pain . Forty-seven insular sites were investigated . Participants received brief stimuli belonging to four different modalities ( nociceptive , vibrotactile , auditory , and visual ) . Both nociceptive stimuli and non-nociceptive vibrotactile , auditory , and visual stimuli elicited consistent LFPs in the posterior and anterior insula , with matching spatial distributions . Furthermore , a blind source separation procedure showed that nociceptive LFPs are largely explained by multimodal neural activity also contributing to non-nociceptive LFPs . By revealing that LFPs elicited by nociceptive stimuli reflect activity unrelated to nociception and pain , our results confute the widespread assumption that these brain responses are a signature for pain perception and its modulation .
The human insula , in particular the region encompassing the dorsal posterior insula and the adjacent parietal operculum , is generally believed to play a specific role in the perception of pain . There are several reasons behind this belief . First , the insula is an important cortical target for nociceptive inputs ascending the spinothalamic tract [1] . Second , direct electrical stimulation of the human insula , as well as focal epileptic seizures in this region , may trigger an acute experience of pain [2–4] . Third , lesions of the insula may lead to a selective impairment of the ability to perceive nociceptive stimuli , as well as central pain [5] . Fourth , depth recordings in humans have shown that nociceptive stimuli elicit robust LFPs in this region , considered to reflect early stages of cortical processing specifically related to the perception of pain [6–9] . Fifth , electroencephalography ( EEG ) , positron emission tomography ( PET ) , and functional magnetic resonance imaging ( fMRI ) studies have shown consistently that the insula is activated by stimuli perceived as painful [10–16] . Finally , several studies have shown a significant correlation between the magnitude of the responses recorded in the insula and the intensity of perceived pain [15 , 17–20] . In particular , Segerdahl et al . [18] recently demonstrated a significant correlation between long-lasting changes in absolute cerebral blood flow ( CBF ) in the dorsal posterior insula and the intensity of perceived ongoing pain . All these observations provide support for a specific involvement of the insula in pain perception . Yet , this conclusion is challenged by several counterarguments or differing findings . Because they imply necessity and sufficiency , lesion studies and focal seizure cases could be expected to provide unequivocal evidence for a specific involvement of the insula in pain perception . However , the notion that pain constitutes a common ictal symptom associated with insular discharge comes from observations performed in only a few patients [3 , 4] . Furthermore , direct electrical stimulation of the insula in these patients appears to predominantly elicit nonpainful paresthesiae or warm sensations , especially when the stimulated area is not epileptogenic [2 , 21] . Finally , reports of insular lesions leading to impaired pain perception have been recently questioned by a study of 24 patients with stroke lesions involving the insula , in which no measurable change in pain thresholds could be objectified using quantitative sensory testing [22] . Most importantly , the assumption that the responses triggered in the insula by nociceptive stimuli are specific for pain is based on reverse inference , and the likelihood of this inference being correct depends on the exclusivity of the relationship between these responses and the experience of pain . In other words , to test whether these responses are specific for pain , one must not only demonstrate that stimuli perceived as painful elicit responses in the insula but also that these responses are elicited if and only if the stimulus is painful . Alongside the assumed pivotal role of the insula in pain perception , it is also widely acknowledged that the insula is involved in the processing of a range of non-nociceptive sensory inputs and that the insula contributes to a large number of cognitive , affective , interoceptive , and homeostatic functions , independently of sensory modality [23–30] . This is not surprising given the heterogeneous cytoarchitecture of the insula and its anatomical connections with a wide array of brain regions [31–36] . Therefore , at least part of the activity recorded in the insula while perceiving pain could reflect cognitive processes that are not specifically related to the pain experience , such as processes involved in orienting attention towards salient stimuli or processes involved in the production of homeostatic responses . The aim of the present study was to address this outstanding question , i . e . , to examine whether the insula exhibits responses specific for nociception and the perception of pain . For this purpose , we took advantage of the high spatiotemporal resolution of depth intracerebral EEG recordings performed in humans for the evaluation of refractory focal epilepsy [37] . Using a very straightforward experimental paradigm ( see Methods section ) , we compared the LFPs triggered by nociceptive stimuli eliciting a perception of pain to the LFPs triggered by non-nociceptive and nonpainful vibrotactile , auditory , and visual stimuli ( Fig 1 ) . We found that all four types of stimuli elicit highly similar LFPs in both the anterior and posterior portions of the insula . This indicates that , unlike previously thought , the greater part of the insular response to stimuli perceived as painful reflects multimodal activity that is entirely unspecific to pain .
Recordings were obtained from a total of 72 contacts ( 47 localized in the insula: 22 in the posterior insula , 25 in the anterior insula , and 25 at locations adjacent to the insula ) in six patients ( four patients with one electrode in the left insula , one patient with one electrode in the right insula , and one patient with electrodes in both the left and right insula ) . The anterior insula was identified as the region encompassing the short insular gyri ( anterior , middle , and posterior ) , the pole of the insula , and the transverse insular gyrus . The posterior insula was identified as the region composed of the anterior and posterior long insular gyri [38] . Although nociceptive stimuli elicited a clear burning/pricking sensation that was systematically qualified as painful , all stimuli were perceived as equally intense ( the average ratings of intensity of perception were not significantly different across sensory modalities; F = . 595; p = . 628 ) . In all patients , all four types of stimuli elicited clear LFPs at anterior and posterior insular contacts , appearing as large biphasic waves . The waveforms obtained at each insular contact of two representative subjects are shown in Fig 2 . The waveforms obtained in all the other participants are shown in S1 Fig . The latency and absolute amplitude of each of the two peaks were measured at each insular electrode contact and compared using a linear mixed models ( LMM ) analysis with “modality” ( nociceptive , vibrotactile , auditory , and visual ) , “contact location” ( anterior and posterior insular contacts ) and “side” ( stimuli delivered to the ipsilateral or contralateral side relative to the explored insular cortex ) as fixed factors and “subject” as a contextual variable . On average , the latencies of the first peak ( nociceptive: 184 ± 50 ms; vibrotactile: 113 ± 40 ms; auditory: 89 ± 23 ms; and visual: 140 ± 36 ms ) and of the second peak ( nociceptive: 296 ± 78 ms; vibrotactile: 205 ± 74 ms; auditory: 161 ± 31 ms; and visual: 216 ± 69 ms ) were significantly different across modalities ( main effect of “modality”; first peak: F = 125 . 25 , p < . 001; second peak: F = 95 . 86 , p < . 001 ) . Post-hoc comparisons showed that the average latency of the responses to nociceptive stimuli was significantly greater than the average latency of the responses to auditory ( first peak: p < . 001; second peak: p < . 001 ) , vibrotactile ( first peak: p < . 001; second peak: p < . 001 ) , and visual ( first peak: p < . 001; second peak: p < . 001 ) stimuli . These across-modality differences in latency can be explained by the difference in the time required for the sensory afferent volleys to reach the cortex [39 , 40] . In particular , the greater latency of the responses elicited by nociceptive stimulation as compared to vibrotactile stimulation ( latency difference of the first peak: 71 ± 90 ms; latency difference of the second peak: 91 ± 152 ms ) can be explained by the fact that small-diameter A-delta fibers conveying nociceptive input have a slower conduction velocity than large-diameter A-beta fibers conveying vibrotactile input . The latencies of the responses to stimuli delivered to the contralateral side ( first peak: 123 ± 47 ms; second peak: 204 ± 60 ms ) and ipsilateral side ( first peak: 139 ± 56 ms; second peak: 236 ± 97 ms ) relative to the explored insula were significantly different ( main effect of “side”; first peak: F = 21 . 16 , p < . 001; second peak: F = 33 . 21 , p < . 001 ) . Independently of the modality of the eliciting stimulus , the responses elicited by stimulation of the ipsilateral side were , on average , slightly delayed as compared to the responses elicited by stimulation of the contralateral side . This is compatible with previous reports also showing a small latency difference between insular LFPs elicited by nociceptive stimuli delivered to the ipsilateral versus contralateral hand [41] . In contrast , there was no significant effect of the factor “contact location” ( first peak: F = 0 . 32 , p = . 569; second peak: F = 0 . 64 , p = . 424 ) . The amplitudes of the first peak ( nociceptive: 19 ± 16 μV; vibrotactile: 13 ± 10 μV; auditory: 24 ± 15 μV; and visual: 11 ± 9 μV ) and the amplitudes of the second peak ( nociceptive: 31 ± 20 μV; vibrotactile: 32 ± 18 μV; auditory: 27 ± 19 μV; and visual: 24 ± 19 μV ) were significantly different across modalities ( main effect of “modality”: first peak: F = 27 . 49 , p < . 001; second peak: F = 5 . 34 , p = . 001 ) . Post-hoc comparisons showed that the amplitude of the first peak was significantly greater for the responses to auditory stimulation as compared to nociceptive ( p = . 010 ) , vibrotactile ( p < . 001 ) , and visual ( p < . 001 ) stimulation and that the amplitude of the second peak was significantly smaller for the responses to visual stimulation as compared to nociceptive ( p = . 009 ) and vibrotactile ( p = . 004 ) stimulation . For both peaks , there was no difference between the amplitude of the responses elicited by stimuli delivered to the ipsilateral and contralateral side ( first peak: F = 1 . 02 , p = . 312; second peak: F = 0 . 52 , p = . 473 ) . Furthermore , there was no difference between the amplitudes of the responses recorded from the anterior and posterior insula ( first peak: F = 0 . 13 , p = . 723; second peak: F = 0 . 60 , p = . 441 ) . The spatial distribution of the amplitudes of the LFPs elicited by the different types of stimuli modalities is shown in Fig 3 . Because the insula represents a relatively large area , and because it may contain spatially segregated subareas subtending different functions , it was crucial to determine whether the insular LFPs elicited by nociceptive stimulation and those elicited by non-nociceptive vibrotactile , auditory , and visual stimulation originate from spatially distinct or identical sources within the insula . For this purpose , linear current source density ( CSD ) plots were computed by numerical differentiation to approximate the second order spatial derivative of the LFPs recorded across the different , evenly spaced contacts of each insular electrode [42] . The obtained signals were then used to compute two-dimensional maps expressing the recorded signals as a function of time and electrode contact location and to identify all electrode contact locations showing inversions of polarity ( Fig 4 , upper panel ) . At the mesoscopic level of intracerebral EEG recordings , the electrical activity generated in a given area can be summarized as an equivalent current dipole , located close to the center of activity , and having an orientation that is orthogonal to the activated cortical surface . Contacts showing an inversion of polarity may thus be considered as located closest to a source of activity , respectively in front and behind the dipole source . In the vast majority of cases ( Fig 4 , lower panel ) , polarity reversals were observed at the same contacts for all four types of LFPs . This indicates that , at least at the level of intracerebral EEG recordings , the locations of the sources generating nociceptive LFPs in the insula can be considered as identical to the locations of the sources generating non-nociceptive vibrotactile , auditory , and visual elicited LFPs ( Fig 4 and S2 Fig ) . Because the insula may be involved in multiple aspects of sensory processing , nociceptive and non-nociceptive LFP waveforms could reflect a combination of modality-specific and multimodal activities ( i . e . , unimodal neural activity specifically related to the processing of input belonging to a given sensory modality and multimodal neural activity reflecting higher-order processes that are independent of sensory modality ) . To test this hypothesis , we used a blind source separation algorithm based on a probabilistic independent component analysis ( PICA ) to break down the LFP waveforms elicited by all four types of stimuli and recorded at the different insular contacts into a set of independent components ( ICs ) [43] . When applied to multichannel electrophysiological recordings , this algorithm separates the recorded signals into a linear combination of ICs , each having a fixed spatial projection onto the electrode contacts and a maximally independent time course . Assuming that modality-specific and multimodal responses have nonidentical spatial distributions across insular contacts , PICA can be expected to separate these responses into distinct ICs . The estimated number of independent sources contributing to the eight LFP waveforms ( four modalities x two sides of stimulation ) ranged , across insulae , between 2 and 6 ( 4 . 0 ± 1 . 5 ) . Multimodal ICs ( i . e . , ICs contributing to the responses elicited by all four types of stimuli ) were the main constituent of all LFPs , both when considering the responses elicited by stimuli to the contralateral side relative to the explored insula ( 3 . 0 ± 1 . 2 ICs; explaining 88% and 95% of the nociceptive LFP peaks , 98% and 93% of the vibrotactile LFP peaks , 95% and 95% of the auditory LFP peaks , and 74% and 78% of the visual LFP peaks; Fig 5 ) and when considering the responses elicited by stimulation of the ipsilateral side ( S3 Fig ) . Taken together , this indicates that nociceptive and non-nociceptive LFPs recorded from the insula predominantly reflect the same source of multimodal cortical activity . A smaller number of ICs appeared to contribute specifically to the LFPs elicited by somatosensory stimuli , regardless of whether the stimulus was nociceptive ( Fig 5 and S3 Fig ) . In addition , a small number of ICs contributed specifically to the LFPs elicited by auditory stimuli . Most importantly , not a single IC was categorized as nociceptive specific .
Our results clearly show that , in both the anterior and posterior insula , LFPs generated by transient nociceptive stimuli are unspecific for nociception and the perception of pain . Indeed , the large biphasic response elicited by nociceptive stimuli at insular contacts was indistinguishable from the large biphasic responses elicited by non-nociceptive vibrotactile , auditory , and visual stimuli , apart from the expected differences in response latencies , which are easily explained by variations in the time required for stimulus transduction , as well as variations in the time required for the afferent volleys to reach the cortex . These responses were recorded from the anterior , medial , and posterior short gyri and from the anterior and posterior long gyri . Although none of our subjects presented contact locations in the superior portion of the anterior long gyrus , LFPs recorded in this region in response to nociceptive stimulation were shown to be identical in morphology to the responses in the other portions of the posterior insula [44] . Not only do we show that all stimuli elicit consistent LFPs in the posterior and anterior insula , we also show that the LFPs elicited by nociceptive and non-nociceptive stimuli originate from the same regions within the posterior and anterior insula . This is demonstrated by the fact that polarity reversals occur at the same electrode contact locations and by the fact that the LFPs elicited by all four types of stimuli have matching spatial distributions across insular contacts . Finally , using a blind source separation algorithm , we show that the insular LFPs elicited by nociceptive stimuli can be largely explained by a source of activity also contributing to the LFPs elicited by non-nociceptive vibrotactile , auditory , and visual stimuli . This indicates that the recorded insular LFPs predominantly reflect a multimodal stage of sensory processing that is independent of nociception and the perception of pain . These findings urge a reinterpretation of the evidence supporting a specific involvement of the insula in pain perception . Ostrowsky and collaborators [21] showed that direct electrical stimulation of the posterior insula can elicit an unpleasant somatic experience , involving shock , burning , or pricking sensations . However , they also observed that stimulation of the insula is equally likely to elicit nonpainful somatic sensations , such as paresthesiae and warm sensations . Furthermore , although vivid painful experiences have been reported following direct electrical stimulation of the insula , these seem to occur mainly when stimulating an epileptogenic area [45] . Similarly , although pain can be associated with epileptic activity in the insula , it remains an uncommon manifestation of insular epilepsy , which has only been observed in a few cases [3 , 4] . Finally , although studies have shown that lesions of the insula can impair the ability to perceive pain [5] , there are also case reports of patients with extensive unilateral or bilateral insular damage showing little or no deficit in the ability to perceive pain , as indicated by the lack of changes in pain thresholds assessed using quantitative sensory testing [22 , 46] . At first glance , our results could appear to be in contradiction with the results of Frot et al . [47] , showing that nonpainful stimuli do not elicit consistent LFPs in the posterior insula . It must be highlighted that the nonpainful stimuli used by Frot et al . [47] were low pulses of radiant heat eliciting a mild sensation of warmth . In contrast , the nonpainful stimuli used in the present study elicited a sensation whose perceived intensity was similar to the perceived intensity of nociceptive stimulation . Hence , the finding that weak thermal stimuli do not elicit LFPs in the posterior insula but more intense vibrotactile , auditory , and visual stimuli elicit consistent LFPs in the posterior insula could be primarily related to differences in stimulus salience ( i . e . , the property of a stimulus to “stand out” relative to neighboring stimuli ) . Importantly , our finding that insular responses to transient sensory stimuli predominantly reflect multimodal activity is in agreement with several other studies suggesting a prominent role of the insula in cognition , attention , and human perception , independently of sensory modality [29 , 32 , 40 , 48–50] . The insula is a very heterogeneous region with a complex structural and functional connectivity . It is involved in a variety of functions , which are not limited to pain and nociception . Although it is often considered as a multidimensional integration site for pain [51] , the insula is multisensory in nature . The insula is considered to be part of a frontoparietal control network commonly activated during tasks that require controlled information processing [52 , 53] , as well as a core network [54–56] that is activated for the maintenance of focal attention . Furthermore , the insula has been related to the detection of salience [57] , possibly constituting a hub connecting sensory areas to other networks involved in the processing and integration of external and internal information [49] . Such a multimodal salience network would allow gaining a coherent representation of different salient conditions , including , but not limited to , pain-related experiences [40 , 58] . This leads us to hypothesize that insular LFPs predominantly reflect multimodal activity involved in detecting , orienting attention towards , and reacting to the occurrence of salient sensory events , regardless of the sensory pathways through which these events are conveyed [59–61] . Alternative interpretations should be considered . Besides being involved in a number of sensory and cognitive processes , the insula has also been associated with autonomic function , interoception , and emotions . Patients with damage in the parietal opercular insular region show an impaired ability to recognize facial expressions of emotions and to experience empathy [62] . Moreover , insular activation has been associated with the experience of disgust and fear [63 , 64] . Craig [65 , 66] described the dorsal posterior insula as an interoceptive system that would give rise to distinct feelings that originate from inside the body , including pain , itch , temperature , sensual touch , muscular and visceral sensations , vasomotor activity , hunger , and thirst . By providing a sense of one’s own physical status , these feelings would reflect needs of the body and underlie mood and affective states . Furthermore , the insula could play an important role in generating autonomic responses , such as those triggered by the occurrence of a salient sensory stimulus [60] or those related to the autonomic expression of emotions [65] . Interestingly , these interpretations could also account for the recent finding that CBF in the posterior insula correlates with the varying magnitude of long-lasting ongoing pain [18] . Finally , one should be cautious to not overinterpret our results . Although our findings clearly question the notion that insular LFPs reflect processes specifically involved in the perception of pain , they do not exclude a specific involvement of the insula in pain perception . Unlike single unit recordings , LFPs sample the activity of neurons at the population level . Indeed , it is thought that the main contribution to LFPs derives from synchronous postsynaptic activity occurring in the apical dendrites of pyramidal neurons located in the cortex surrounding the electrode contact [67] . Therefore , one cannot exclude the possibility that LFPs elicited by nociceptive and non-nociceptive stimuli might reflect the activity of distinct neurons intermingled within the same subregions of the insula . However , single unit recordings performed in the monkey insula suggest that the population of truly nociceptive-specific neurons is extremely sparse [68] . In conclusion , by showing that , in the insula , LFPs elicited by nociceptive stimuli are spatially indistinguishable from the LFPs elicited by non-nociceptive vibrotactile , auditory , and visual stimuli , our results confute the widespread assumption that these brain responses constitute a signature for pain perception and its modulation . Does this constitute a demonstration that the insula cannot be considered as a “primary cortex for pain ? ” Although it is important to acknowledge the fact that the function of primary sensory cortices is probably not restricted to the processing of sensory input belonging to its corresponding sensory modality and , instead , that primary sensory cortices subsume multisensory integration functions [69–71] , studies have shown that neurons sensitive to other modalities are rare within primary visual , auditory , and somatosensory areas . For this reason , and in striking contrast with our insular recordings , large-amplitude LFPs are recorded in primary sensory areas only if the eliciting stimulus activates afferents belonging to the corresponding sensory modality [72 , 73] . Therefore , although our results clearly do not exclude the existence of nociceptive-specific or pain-specific processes in the insula , they do highlight the lack of a spatially segregated parcel of the human insula that could be considered as a “primary cortex” for pain .
Six patients ( three females , mean age: 27 , range 19–43 y ) recruited at the Department of Neurology of the Saint Luc University Hospital ( Brussels , Belgium ) were included in the study . All participants suffered from focal epilepsy and , before functional surgery , were investigated using depth electrodes implanted in various brain regions suspected to be the origin of the seizures , including the anterior and posterior insula . The intracerebral EEG was recorded from a total of 72 sites . The localization of the insular electrodes for each patient can be seen in Fig 2 and S1 Fig . None of the patients presented ictal discharge onset in the insula , and low voltage fast activity was never present in this area during spontaneous seizures . The study was conducted at the patient bedside . Before the beginning of the experiment , the procedure was explained to the participant , who was exposed to a small number of test stimuli for familiarization . The experiment consisted of two sessions of four blocks each , one session per side of stimulation . In each block , the subject received stimuli belonging to one of four sensory modalities: nociceptive , vibrotactile , auditory , and visual . Each stimulation block consisted of 40 stimuli . The order of the stimulation blocks was randomized across participants . A blocked design was chosen to ensure that expecting the possible occurrence of a nociceptive stimulus would not affect the responses elicited by non-nociceptive stimuli [74] . The interstimulus interval ( ISI ) was large , variable , and self-paced by the experimenter ( 5–10 s ) . Participants were instructed to keep their gaze fixed on a black cross ( 3 x 3 cm ) placed in front of them on the edge of the bed , at a distance of ~2 m , 30° below eye level , for the whole duration of each block . To ensure that each stimulus was perceived and to maintain vigilance across time , participants were asked to press a button as soon as they felt the stimulation . Furthermore , participants provided a subjective intensity rating for each stimulus on a scale ranging from 0 to 10 ( 0 was defined as “undetected” and 10 was defined as “maximum intensity” ) . At the end of each block , the patients were asked to report whether they had perceived any of the stimuli as painful . Nociceptive somatosensory stimuli consisted of 50-ms pulses of radiant heat generated by a CO2 laser ( wavelength: 10 . 6 μm ) . The laser beam was transmitted via an optic fiber , and focusing lenses were used to set the diameter of the beam at the target site to 6 mm . The laser stimulator was equipped with a radiometer providing a continuous measure of the target skin temperature , which was used in a feedback loop to regulate laser power output . The power output of the laser was adjusted to raise the target skin temperature to 62 . 5°C in 10 ms and to maintain this temperature for 40 ms . To prevent nociceptor fatigue or sensitization , the laser beam was manually displaced after each stimulus [75] . Each laser stimulus elicited a clear painful pinprick sensation , previously shown to be related to the activation of Aδ fiber skin nociceptors [74] . Non-nociceptive somatosensory stimuli consisted in a 50-ms vibration at 250 Hz , delivered via a recoil-type vibrotactile transducer driven by a standard audio amplifier ( Haptuator , Tactile Labs , Canada ) and positioned on the palmar side of the index fingertip . Auditory stimuli were loud , lateralized sounds ( 0 . 5 left/right amplitude ratio ) delivered through earphones . The sounds consisted in a 50-ms tone at 800 Hz . Visual stimuli were 50-ms punctate flashes of light delivered by means of a light-emitting diode ( LED ) with a 12 lm luminous flux , a 5 . 10 cd luminous intensity , and a 120° visual angle ( GM5BW97333A , Sharp , Japan ) , placed on the hand dorsum . For each patient , a tailored implantation strategy was planned on the basis of the regions considered most likely to be ictal onset sites or propagation sites . The desired targets , including the insular cortex , were reached using commercially available depth electrodes ( AdTech , Racine , Wisconsin , United States; contact length: 2 . 4 mm; contact spacing: 5 mm ) , implanted using a frameless stereotactic technique through burr holes . The placement was guided by neuronavigation based on a 3D T1-weighted MRI sequence . A post-implantation 3D-T1 ( 3D-T1W ) MRI sequence was used to accurately identify single contact localizations . Intracerebral EEG recordings were performed using a Deltamed ( Paris , France ) acquisition system . Additional bipolar channels were used to record electromyographic ( EMG ) and electrocardiographic ( EKG ) activity . All signals were acquired at a 512 Hz sampling rate and analyzed offline using Letswave 6 [76] . The continuous recordings were referenced to the average of the two mastoid electrodes ( A1A2 ) , segmented into 1 . 5-s epochs ( −0 . 5 to 1 . 0 s relative to stimulus onset ) and band-pass filtered ( 0 . 3–40 Hz ) . After baseline subtraction ( reference interval: −0 . 5 to 0 s relative to stimulus onset ) , separate average waveforms were computed for each subject , stimulus type ( nociceptive somatosensory , non-nociceptive somatosensory , auditory , and visual ) , and side of stimulation . For two of the subjects , trials containing strong artifacts were corrected using an independent component analysis ( ICA ) algorithm [77] or removed after visual inspection . The latencies and amplitudes of the LFPs were compared using a LMM analysis as implemented in IBM SPSS Statistics 22 ( Armonk , New York: IBM ) with “modality” ( four levels: nociceptive , vibrotactile , auditory , and visual ) , “contact location” ( two levels: anterior and posterior insula ) and stimulation “side” ( two levels: stimuli delivered to the ipsilateral or contralateral side relative to the explored insula ) as fixed factors . Assuming that the responses recorded from the different contacts of a given subject are not independent , “subject” was used as a contextual variable grouping the insular contacts . Parameters were estimated using restricted maximum likelihood ( REML ) [78] . In all analyses , main effects were compared using the Bonferroni confidence interval adjustment . Linear CSD plots were computed by numerical differentiation to approximate the second order spatial derivative of the LFPs recorded across the different , evenly spaced contacts of the insular electrode [42] . The obtained signals were then used to compute two-dimensional maps expressing the recorded signals as a function of time and electrode contact location , using spline interpolation . The spatiotemporal maps were then used to identify visually all electrode contact locations showing polarity reversal , as well as to compare the spatial distribution of the LFPs elicited by nociceptive and non-nociceptive stimuli . A blind source separation algorithm was used to isolate the contribution of multimodal and modality-specific neural activities to the LFPs elicited by nociceptive and non-nociceptive vibrotactile , auditory , and visual stimuli . For each participant , the blind source separation was performed using runica [77 , 79] , an automated form of the Extended Infomax ICA algorithm [80] . When applied to multichannel recordings , this algorithm separates the recorded signal into a linear combination of ICs , each having a fixed spatial projection onto the electrode contacts and a maximally independent time course . When ICA is unconstrained , the total number of ICs equals the total number of channels . If the number of ICs is far greater than the actual number of independent sources , ICs containing spurious activity will appear because of overfitting . On the other hand , if the number of ICs is much smaller than the actual number of sources , information will be lost because of underfitting . For this purpose , ICA was constrained to an effective estimate of the intrinsic dimensionality of the original data ( PICA ) [81] . The estimate was obtained using a method based on maximum likelihoods and operating on the eigenvalues of a principle component analysis [43] . For each participant , the algorithm was applied to the eight average waveforms ( 4 types of stimuli x 2 sides ) obtained at all insular contacts ( 8–12 contacts ) . To estimate the contribution of each obtained IC to the LFPs elicited by the different types of stimuli , the time course of the amplitude of each IC ( μV ) was expressed as the standard deviation from the mean ( z-scores ) of the prestimulus intervals of all eight waveforms ( −0 . 5 to 0 s ) . If the poststimulus amplitude of an IC was greater than z = 1 . 5 , the IC was considered as reflecting stimulus-evoked activity . Each of these ICs was then classified according to its contribution to the eight LFP waveforms . For each IC and each side of stimulation , we computed the ratio between the z-score of a specific modality and the z-scores of the other three modalities [40 , 34] . If the ratio was ≥2 for one stimulus modality versus each of the other three modalities , the IC was classified as modality specific ( i . e . , nociceptive , non-nociceptive vibrotactile , auditory , or visual ) . If the computed ratio was ≥2 for both nociceptive and non-nociceptive somatosensory stimuli versus auditory and visual stimuli , the IC was classified as somatosensory specific . Finally , ICs that contributed to all four LFP waveforms were classified as multimodal . Crucially , the obtained results were not critically dependent on the number of dimensions used to constrain ICA or on the arbitrarily defined threshold of z ≥ 2 . In fact , all ICs were unambiguously multimodal or modality specific , and IC classifications obtained using other cut-off values ranging between 1 . 5 and 3 . 5 yielded identical results . The anterior insula was identified as the region encompassing the short insular gyri ( anterior , middle , and posterior ) , the pole of the insula , and the transverse insular gyrus . The posterior insula was identified as the region composed of the anterior and posterior long insular gyri [38] . Individual MRI were normalized to a standard echo-planar imaging ( EPI ) template in MNI space , using Statistical Parametric Mapping ( SPM8 , Wellcome Department of Imaging Neuroscience , London , United Kingdom ) . The anatomical location of each contact was identified on the 3D-T1W sequence with the help of multiplanar reformations , by a neuroradiologist ( MMS ) with 10 y of experience . | A widely accepted notion is that the insula , especially its posterior portion , plays a specific role in the perception of pain . This has led a number of researchers to consider activity recorded from this so-called “ouch zone” as an objective correlate of pain perception . We provide compelling evidence to the contrary . Using direct intracerebral recordings , we demonstrate that painful and nonpainful stimuli elicit very similar responses throughout the human insula . This observation argues against the notion that these responses reflect the brain activity through which pain emerges from nociception in the human brain . These findings have implications for basic theories , as well as for the development of diagnostic tests and the identification of therapeutic targets for the treatment of chronic pain . They question the use of these insular responses to assess the effects of pharmacological treatment or to assess pain in patients unable to communicate . Furthermore , they have legal implications , as they contradict the proposal that these responses could be used to determine unequivocally whether plaintiffs are truly experiencing the pain for which they are seeking redress . Finally , they undermine the rationale for neurosurgical procedures aiming at alleviating pain by targeting the posterior insula . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [] | 2016 | Nociceptive Local Field Potentials Recorded from the Human Insula Are Not Specific for Nociception |
Cytoplasmic flows are an ubiquitous feature of biological systems , in particular in large cells , such as oocytes and eggs in early animal development . Here we show that cytoplasmic flows in starfish oocytes , which can be imaged well with transmission light microscopy , are fully determined by the cortical dynamics during surface contraction waves . We first show that the dynamics of the oocyte surface is highly symmetric around the animal-vegetal axis . We then mathematically solve the Stokes equation for flows inside a deforming sphere using the measured surface displacements as boundary conditions . Our theoretical predictions agree very well with the intracellular flows quantified by particle image velocimetry , proving that during this stage the starfish cytoplasm behaves as a simple Newtonian fluid on the micrometer scale . We calculate the pressure field inside the oocyte and find that its gradient is too small as to explain polar body extrusion , in contrast to earlier suggestions . Myosin II inhibition by blebbistatin confirms this conclusion , because it diminishes cell shape changes and hydrodynamic flow , but does not abolish polar body formation .
The three most important cellular processes that drive animal development are cell division , cell shape changes and cell migration , which are all mediated by the cytoskeleton [1 , 2] . Strikingly , cytoskeletal regulation during early animal development often takes the form of mechanical-biochemical waves or pulses , for example in axolotl [3] , starfish [4] , Xenopus [5] , Drosophila [6–8] and mouse [9] . The biological function of such waves ranges from global coordination of cell division to localising cellular structures or molecular components [10–12] . In principle , it is expected that forces generated by the cytoskeleton lead to intracellular flows . Indeed cytoplasmic flows have been shown to have important functions in development , and have been proposed to mediate intracellular transport [13–15] , intracellular mixing [16] , force distribution [17 , 18] and pattern formation [19] . However , the relation between wave-like cell shape changes and cytoplasmic flows has rarely been investigated in quantitative detail in a developmental model system , mainly due to experimental technical limitations . Starfish oocytes are an ideal model system to investigate this relation because they are available in large numbers , they are large ( diameter 180 μm ) , transparent ( unlike most vertebrate eggs e . g . Xenopus or zebrafish ) and are not enclosed by a stiff shell ( like eggs of Drosophila or C . elegans ) . During early development , several stereotypical surface contraction waves ( SCWs ) run over the cell surface , leading to large surface deformations and intracellular flows [4 , 20 , 21] . However , despite a long tradition of studying them , it is unknown how exactly the surface deformations and the internal flows are related and what their biological functions are . In detail , it is not clear if the internal flows are a direct physical consequence of the cell shape changes or if the cytoplasmic flows are also driven by forces generated in the cytoplasm . For instance it has been shown that flows in Drosophila oocytes are driven by cytoplasmic microtubule networks [14] and that flows in C . elegans embryos are driven by cortical tension gradients along the AP-axis [22] . However , in terms of the physical properties , a key difference is that Drosophila and C . elegans develop in a stiff shell , which is typically not the case for deuterostome species that include vertebrates and also the echinoderm starfish . Such eggs without a shell , like starfish , have the possibility to drive large scale cytoplasmic flows by normal deformations of their soft cortex . Indeed , comparable cortical contractions have been observed in vertebrates like frog or mouse , clearly demonstrating the general importance of this process . For different species SCWs have been suggested to be involved in polar body extrusion , but the physical basis has not been tested quantitatively . This is particularly true for starfish , for which it has been suggested early on that the cytoplasmic flow is causal for the extrusion of the polar body following the SCW [4] , but quantitative evidence for this suggestion has been missing . Although it has later been demonstrated that polar body formation is not directly affected by forced changes in cytoplasmic flows [23] , the notion that the combination of hydrodynamic flow and local cortical weakening drives polar body extrusion is still present in the literature [24 , 25] . We recently showed that the polar body is generated even if the SCW is strongly reduced by myosin II inhibition [21] , but our earlier study did not address hydrodynamic flows . To study the physical relation between surface deformation and cytoplasmic flow in soft oocytes as well as its biological consequences for polar body extrusion , here we combine live cell imaging , quantitative image processing and biophysical modelling in starfish . During maturation of starfish oocytes there are two meiotic divisions , each leading to the formation of one polar body . Each is preceded by a surface contraction wave leading to surface deformations and internal hydrodynamic flow . In this study we focus on the most prominent first SCW that is associated with meiosis I . It takes about seven minutes to run over the surface of the oocyte of 180 μm diameter . Because the focus of our work is on hydrodynamic flow , we will use transmission light microscopy , which in starfish oocytes gives excellent contrast due to the presence of intracellular yolk particles and thus can be used to reconstruct hydrodynamic flow using particle image velocimetry ( PIV ) . We present a complete quantitative analysis of cell shape changes and hydrodynamic flows that demonstrate that the SCWs are the direct physical cause for the observed cytoplasmic flows . Thus no other mechanism is required to explain cytoplasmic flows during the SCW . Using a contraction model , we can not only study the large normal surface changes , but also the small tangential components . As an immediate consequence of our analytical treatment , we now can calculate the internal pressure field , that is not directly accessible experimentally [26] and has been obtained in similar systems before only in a numerical manner [17 , 22 , 27 , 28] . Our analytical calculations are ideal to quickly process large data sets and show that the cytoplasmic flows caused by the SCWs are not sufficient to explain polar body extrusion , in contrast to earlier suggestions [4] . We finally show that inhibition of myosin II activity by blebbistatin does strongly reduce both cell shape changes and hydrodynamic flow , but not polar body formation , which seems to be caused by myosin II-independent local polymerisation of actin . We start by imaging the full three-dimensional ( 3D ) shape of freely floating oocytes labelled with fluorescent dextran , in order to demonstrate that the SCW appears to be rotational symmetric around the animal-vegetal ( AV ) axis . Based on this symmetry , we then switch to two-dimensional ( 2D ) slices containing the AV-axis acquired with transmission light microscopy for oocytes in imaging chambers . This approach allows us to simultaneously quantify cell shape changes and cytoplasmic flows . Using a Fourier decomposition of the contour , we confirm that the SCW is highly symmetric in the 2D imaging plane regarding the AV-axis . We then use differential geometry for axisymmetric shapes to calculate previously inaccessible quantities such as mean curvature and Gaussian curvature as a function of time . In a next step , we develop an analytical hydrodynamic model that predicts the flows inside the oocyte as a result of the observed SCW . The mathematical model is based on the analytical solution of the Stokes equation and has the advantage that even videos with several thousand frames can now be easily analysed . Additionally , different contributions to the flows can be separated and compared to each other . We find excellent agreement between our predictions and the PIV-results . From this we can conclude that during the time of the SCW , the cytoplasm behaves as a simple Newtonian fluid on the scale of several micrometers . The model is general in the sense that it can be applied to any cell that is approximately spherical , as it is the case for the oocytes of most echinoderms and vertebrates , including mammals . Finally we use our finding to address the proposed biological function of cytoplasmic flows in starfish oocytes . We calculate the intracellular pressure field and find that it is not sufficient to account for polar body extrusion , in contrast to earlier suggestions . This is confirmed experimentally by blebbistatin experiments , which strongly reduce cell shape changes and hydrodynamic flow , but not polar body extrusion . By analysing the cytoplasmic flows we can furthermore assess the role of cytoplasmic flows for mixing of cytoplasmic components , including determinants that pattern embryonic development .
In order to image cell shape changes in 3D during the SCW , we microinjected a bright fluorescent dextran into starfish oocytes to mark the cell volume . We then used a Zeiss AiryscanFast microscope on freely floating cells to acquire one 3D image stack every 7 s over a 15 min time course including the SCW . Fig 1A and 1B show sequences of snapshots from a representative cell ( full videos provided as S1 and S2 Videos respectively ) . While Fig 1A shows the maximal intensity projections performed with the open source image processing package Fiji , Fig 1B is a rendering of the segmentation performed with the commercial image processing software Imaris ( for details , see the Methods section ) . In general , the cell is always close to spherical due to the large cortical tension of around 2 mN m−1 [21] . Close inspection shows that it takes the SCW roughly 7 min to run from the vegetal ( V ) to the animal ( A ) pole , and that the polar body starts to appear roughly 3-4 min after the wave has started . The wave itself is visible as a band of deformation running from the V- to the A-pole , confirming previous observations [21] . In Fig 1C and 1D we show the global quantities surface area and volume that can be extracted from the segmentation shown in Fig 1B . One sees that both are slightly reduced during the SCW , but then recover to pre-wave values . Interestingly , the area then shows an overshoot , which partially corresponds to the formation of the polar body , presumably because exocytosis and membrane flattening generate additional membrane area . Exocytosis of fluorescent dextran might also explain why the values for surface area and volume drop after the SCW . In Fig 1E we show the boundary contours from Fig 1A . While the shape changes are relatively small , it is apparent that they are symmetric around the AV-axis . In order to address the relation to hydrodynamic flow , we therefore turned to transmission light microscopy in 2D slices containing the AV-axis , that allowed simultaneous imaging of cell shape changes and cytoplasmic flows . We next recorded transmission light microscopy videos of oocytes in imaging chambers at the height of maximal area and selected oocytes with the AV-axis lying in this slice ( N = 24 ) . Because immature oocytes before meiosis I have their nucleus located at the A-pole , the AV-axis can be identified before the SCW starts . In agreement with Fig 1E , in 2D we always observed the stereotypic sequence of events shown schematically in Fig 2A . Naturally , the oocytes are surrounded by a jelly coat to protect them from the sea environment . By removing the jelly coat , one can observe even larger surface deformations , but here we show the results for one representative cell with the jelly layer kept , in order to reflect physiological conditions . A description of the cell surface in polar coordinates gives a radius function r ( θ , t ) ( Fig 2B ) . We defined polar coordinates such that the polar angle θ = 0 lies at the A-pole and the V-pole at θ = π . After a Fast Fourier Transform ( fft ) , a back transform with seven Fourier modes resulted in the smooth representation shown in Fig 2B . A comparison with the backtransform using even parts only ( symm fft ) revealed that the radius function is highly symmetric around the AV-axis ( Fig 2B ) . The AV-axis determined by symmetry considerations coincides very well with the axis connecting the centre of mass ( CM ) with the centre of the nucleus . In CM frame , during the full time course we first observed a local minimum on the AV-axis that then became a maximum and finally developed back to a minimum ( Fig 2C ) , whereas on the perpendicular axis towards the equator the opposite behaviour can be observed . The maximal radius changes are about 10% . The high symmetry around the AV-axis apparent in Fig 2B was in fact found for the whole time course of the SCW ( Fig 2D and S3 Video ) . Fig S1 Fig shows the surface tracking results for the complete data set ( N = 24 ) for the time point after the SCW , when maximal shape changes were observed . Considering the full data set demonstrates that the cell shown in Fig 2 is indeed representative and that the described shape changes are very stereotypical in starfish oocytes . As suggested by Fig 1 for the 3D data and by the symmetry in the 2D slices containing the AV-axis shown in Fig 2 , we assume rotational symmetry of the oocyte around the AV-axis to quantify the geometrical changes during the SCW on the basis of the 2D slices . Fig 3A shows a cartoon of the invaginated shape expected for a strong SCW passing the equator . Using differential geometry , one now can calculate the 3D curvature of the cell surface from the 2D data ( for details compare Methods section ) . For example , at the equator of an invaginated shape the two principal curvatures should have opposite signs . In the following we calculate all local geometrical properties of the 3D surface from the 2D data as a function of the polar angle θ ( Fig 3B ) . Fig 3C shows a kymograph of the normalised difference between the oocyte’s shape and a perfect sphere . We describe the surface by a radial coordinate r = a ( 1 + f ) , where a is the radius of a reference sphere and f describes the deviation from the spherical reference case . One sees that the beginning and end of the wave are clearly visible at times around 7 and 13 min , and that the wave velocity ( slope ) is well defined . Fig 3D shows the kymograph of the radial surface velocity in centre of mass frame of reference , which shows the same pattern . The deformation persists for several minutes after the SCW has passed ( Fig 3C and 3D , S3 Video ) . The mean curvature H and the Gaussian curvature K ( Fig 3E and 3F , respectively ) show similar patterns as do the deformation and velocity kymographs . Both curvatures change towards zero because as the wave passes by , the invaginated part at the equator acquires a saddle-shape ( a perfect saddle has H = 0 and negative K ) . The reduced curvatures show that the SCW can be described as a band of local flattening running from V-pole to A-pole , as concluded earlier from a purely two-dimensional analysis [21] . K1 and K2 shown in Fig 3G and 3H show the principal curvatures in φ-direction and θ-direction , respectively ( blue and yellow plane in Fig 3A ) . The wave is most pronounced in K2 , which corresponds to the curvature directly visible in the imaging plane ( Fig 3H ) . Fig S2 Fig uses the example of K2 to show that the 3D data presented in Fig 1 gives similar results when subjected to the procedures described here for the 2D data . However , the resolution and throughput is much better in the 2D case . After the geometric description of the SCW , we next asked how the cell shape changes relate to cytoplasmic flows . We therefore implemented an analytical model for the hydrodynamic flows inside a spherical shell and applied it to the problem of flows inside the deforming oocyte . With this model we are able to predict the flows of a Newtonian , incompressible fluid at low Reynolds number . The Stokes equation was solved analytically with no-slip boundary conditions on the boundaries extracted from the experiments ( Fig 4A ) . We note that no-slip boundary conditions are commonly applied when solving the Stokes equation of hydrodynamic flows in soft matter and biological systems [29–31] . Potential slip lengths at the boundary are expected to be in the nanometre range and therefore can be neglected in our context . Although volume and surface area are known to change during the time course of the SCW , this does not matter for our computational procedures because the Stokes equation does not have memory and can be solved anew for each time point . In principle the model can predict flows even if there is no rotational symmetry in the problem . Three examples are given in Fig 4B–4D , where rotational symmetry is assumed for all three setups . The effects of arbitrary tangential and radial surface movement can be calculated separately as shown here for a radial movement ( Fig 4B ) and a tangential movement ( Fig 4C ) . Due to linearity of the Stokes equation , different contributions to the flows can be added as it was done for rigid body movement in Fig 4D . In this case , radial and tangential surface movements were added to obtain the constant flow field for rigid body movement . Additionally , the model is able to predict the flow field inside a slightly deformed sphere by a perturbation ansatz as done for Fig 4D . Together with the flow field , this model gives the pressure field ( p from Fig 4A ) , that is developing inside the fluid . In the Methods section we give analytical expressions for these quantities , and additional information is provided in the S1 Appendix . In order to address the mechanical basis of the surface deformations , we used a contractile surface model that describes a band of increased surface tension moving over the oocyte cortex ( Fig 5A ) , following [21] and as confirmed now in 3D by Figs 1 and 3 . For each time step the shape of the surface is computed numerically by minimising the appropriate surface Hamiltonian . In our specific case , we used a Gaussian-shaped contraction band travelling with constant angular speed and constant width but changing amplitude ( see Fig S3 Fig ) . We adjusted the strength and width of the contraction band to visually reproduce the curvatures found in experiment ( Fig 5B and 5C ) . The angular width of the contraction band was determined to be 72° . Our contraction model also allowed us to predict the tangential surface movement . Here we used the model of a flowing viscous surface as previously applied to the cortex of C . elegans embryos [32] . The dynamics of the SCW are sufficiently slow ( timescale of minutes ) so that short time elastic contributions can be neglected . We combined the analytical solution for the tangential surface movement with the radial movement of the surface obtained before from modelling the shape changes . We then used the resulting tangential surface movement as an input for the hydrodynamic model in order to predict the internal flows that should result from a localised Gaussian contraction band . For four different positions these are shown in Fig 5D . These flows are rather robust to variation of parameters , such as width of the contraction band and the spatial decay length of the velocity . More details are given in the Methods section . We next compared our theoretical predictions with the cytoplasmic flow as measured by PIV [33 , 34] applied to the 2D transmission light microscopy data . Here we exploit the fact that in starfish , the naturally occurring yolk particles ( 1 μm to 2 μm in size ) follow the hydrodynamic flow and produce the contrast needed for the PIV-algorithm . Using this method , internal flows were quantified reliably during the full SCW with a spatial resolution of 10 μm ( Fig 6 , S4 Video , see Methods section for the PIV details ) . We find that during the SCW the flows at first drive the cytoplasm from vegetal to animal pole , then the flows reverse and drive the fluid back to the vegetal pole . The flows have curved streamlines and visually show a high degree of symmetry around the AV-axis ( Fig 6A–6D ) . Also the flows show a clear wave-like behaviour which is further confirmed by the kymographs in Fig 6E and 6F for the flows along the AV-axis and on the axis perpendicular to it , respectively ( colour code given as Fig 6G ) . With the hydrodynamic model introduced above the internal flows during the SCW can be predicted from the surface movement of the oocyte and compared with the measured flows . In the radial displacement model the radial surface movement obtained from experiment is used as sole input . This surface movement thus sets the boundary conditions for the solution of the Stokes equation for the cell interior . In Fig 7A a comparison between the flows predicted by this model and the experimental flows is shown . The radial surface movement is obtained by comparing the contour points of subsequent frames . We fitted the movement of the whole cell along the AV-axis to the experimental flows as the radial surface movement was calculated in the cell centre of mass frame . This model is able to explain the strength and the general shape of the flows during the SCW ( Fig 7A ) . The velocity of the internal flows is of comparable size to the radial surface velocity . The good correspondence between the predicted and the experimental flows justifies the assumption of modelling the cytoplasm as an incompressible Newtonian fluid on the 10 μm scale . In addition to the flows , the model predicts the internal pressure field which scales linearly with the fluid viscosity . The viscosity of the cytoplasm was determined to be about 5mPa s by comparing the diffusion coefficient of fluorescently labelled dextran in water and in the cytoplasm using fluorescence correlation spectroscopy ( FCS , compare Methods section ) . The pressure difference inside the oocyte is on the order of 200 μPa . A full video comparing experiment and this model can be found as S5 Video . Visual inspection of Fig 7A and S5 Video shows that the intracellular flow might have a stronger tangential component close to the cortical surface than predicted by the radial displacement model . Indeed experimentally often a strong tangential flow component can be observed ( S6 and S7 Videos ) . However , it is very difficult to quantitatively measure tangential surface movement experimentally . We therefore used the mechanical contraction model to predict possible tangential flows ( compare Fig 5 ) . Adding the effect of tangential surface movement to the predictions of the radial displacement model resulted in the tangential displacement model . We next fitted the strength and the position of a Gaussian contraction band to the experimental flows ( see the Methods section for further details ) . The result for one time step is shown in Fig 7B where the quality of the modelled flows further improves by taking tangential surface movement into account ( full comparison given as S8 Video ) . In Fig S4 Fig a detailed view for additional time steps for this method are shown . The maximum internal pressure difference inside the cell is again found to be around 200 μPa , but now the gradient is more clearly directed along the AV-axis . The addition of the tangential movement to the model roughly doubled the accuracy of the model during the SCW , as shown in Fig 7C and 7D . Here the plotted normalised residuals are defined as the root of the squared difference between modelled and measured flows normalised by the strength of the measured flows . The increase in agreement becomes apparent during the wave ( full time course in C , zoom in to the wave in D ) , when significant flows are measured . Before and after the wave , the normalised residuals fluctuate around unity by definition . In summary , we conclude that the flow pattern during the SCW in the cytoplasm is driven purely by the surface movement . This finding implies that the cytoplasm behaves like a Newtonian fluid during this stage of the development . This is surprising given the microscopic complexity of the cytoplasm and reflects that during SCWs , the cytoplasm of the starfish oocytes does not contain any relevant ( cytoskeletal ) structures on the scale of several micrometers . We also note that the model predicts a significant tangential component , which experimentally cannot be measured easily . We finally discuss the possible biological functions of the observed phenomena . Mixing of the cytoplasm has been proposed as a potential biological function of flows in various species [15 , 16 , 32] . To elucidate the role of the SCWs for intracellular mixing , we calculated trajectories of fictitious particles embedded into the induced flow fields . We observe that most simulated tracer particles return to their original position after the SCW ( Fig 8A ) . Only regions near the animal pole show significant displacement due to the fact that the cell is still highly deformed at the end of imaging as shown in the kymographs in Fig 3 . Therefore intracellular transport of particles on the scale of several micrometers can be ruled out as a potential function of the SCW . Additionally we do not observe any mixing in the trajectories due to the SCW as the streamlines have the same shape and move in parallel to each other . Having access to the flow field and the pressure field , we are also able to revisit the earlier suggestion that that the SCW drives extrusion of the polar body [4] . If the polar body was pushed out by intracellular pressure , as in blebbing [35 , 36] , large pressure differences between inside and outside would be required to be built up by the hydrodynamic flow . Before the SCW , the pressure difference between inside and outside of the oocyte can be estimated from the Laplace law , Δp = 2γ/R . The tension in the oocyte surface has been measured by micropipette suction experiments to be around 2 mNm−1 [21] . With a cell radius of R = 90 μm , this gives a pressure difference of 44 Pa . This value is similar to but somehow smaller than typical values for human cells , which have been reported to be of the order to several hundred Pa [35 , 37] , because they have smaller sizes but comparable levels of cortical tension . If hydrodynamic flow now led to polar body extrusion , one would expect at least a contribution of the same order of magnitude as 44 Pa . The internal pressure difference predicted by our hydrodynamic model however is only about 200 μPa , which is five orders of magnitude smaller . Therefore we conclude that the internal pressure difference due to the SCW is negligible compared to the total pressure difference needed for polar body extrusion . This is illustrated in Fig 8B and 8C . Bischof et al . showed that the SCW is significantly weakened when non-muscle myosin II contractility is inhibited by blebbistatin [21] . In agreement with our conclusion that cell shape changes drive hydrodynamic flow , we find experimentally that under blebbistatin , also hydrodynamic flows are weakened ( Fig 8D and S9 and S10 Videos ) . Yet these cells do form a polar body ( Fig 8E ) , although they are not able to pinch in the neck . By imaging fluorescently labelled actin , a clear accumulation of actin at the position where the polar body is formed becomes visible ( Fig 8F and S11 Video ) . This suggests that actin polymerisation is the underlying mechanism to generate polar body formation , similar to the mechanism of polar body extrusion in Xenopus [38] .
SCWs have been described in various species , including axolotl , frog and starfish , for more than 30 years [3] . Recently Bischof et al . have revealed the underlying molecular mechanism for starfish [21] , explaining the travelling band of local flattening by local activation of the Rho-module coupled to the cell cycle by a cdk1-gradient . However , this analysis did not consider the consequences for the cytoplasmic flow . Here we provided a full quantitative analysis of the hydrodynamics during the SCW and therefore for the first time arrive at a complete description of the SCW during starfish meiosis . We first provided a full 3D analysis of the cell shape changes during the SCW . Imaging oocytes labelled with fluorescent dextran revealed that they are roughly axisymmetric around the AV-axis , thus justifying a 2D approach in which we only image 2D slices containing the AV-axis with transmission light microscopy , which allowed us to simultaneously measure cell shape and hydrodynamic flow . The assumption of rotational symmetry around the AV-axis has been suggested before [25] and is also supported by the observation that the orientation of the SCW is fully determined by the position of the nucleus [21] . Here it is further supported by the observations that both cell shape changes and hydrodynamic flow are highly symmetric in the plane containing the AV-axis . Assuming axisymmetry and using differential geometry , we are able to calculate 3D curvatures from 2D data . Our quantification of cell shape changes in all directions confirms the results of Bischof et al . [21] , and in addition gives also mean and Gaussian curvature , which are important to fully evaluate the relevant surface energies [39] . By using the different curvatures , we also were able to estimate the angular width of the contraction band to be about 72° . In principle , one also can calculate surface area and volume for the rotational symmetric surface assumed here ( compare Methods section ) . For the cell shown in Fig 3 , we found that surface area and volume reduce by about 3% and 4% , respectively , during the SCW . Regarding our whole data set ( N = 24 , including the cases with the jelly removed ) , we found reductions by ( 8 ± 2 ) % and ( 12 ± 3 ) % , respectively ( Fig S5 Fig ) . Although the reduction during SCW and the following overshoot in area are in agreement with the 3D data from Fig 1 , the large magnitude of these changes and the overshoot in volume ( which here is calculated from the same radius data as the area ) are probably artefacts of the 2D method , where oocytes are confined in imaging chambers . However , this does not directly affect the main focus of our work , namely analysis of the hydrodynamic flow , for which surface area and volume are not essential . In the future , one might envision a more general modelling framework that also includes water flow through the cell surface . This however requires more precise measurements of surface area and volume , and a corresponding image processing pipeline , that was not the focus of the work reported here . In general , we anticipate that very good resolution was required to resolve these effects . If one assumes an outflow of ΔV = 104 μm3 in T = 2 min and a surface area of A = 105 μm2 ( compare Fig 1 ) , then the resulting movement of the surface would be v = ΔV/ ( AT ) = 10−3 μm/s , much smaller than the typical speed of surface movement around 0 . 2 μm/s . With the current resolution , this effect can thus be safely neglected . The main focus of this work is the hydrodynamic flow during SCWs . Hamaguchi and Hiramoto have given a pictorial description of the flows during the SCW in their seminal paper in 1978 [4] . Since then , methods to quantitatively measure flow have evolved tremendously . We measured the internal flows by PIV on transmission light microscopy recording , which gave us a quantitative description of the flows with high temporal and spatial resolution . We used the naturally occurring yolk particles as tracers of the internal flows which circumvents the use of artificial tracer particles . We found that the SCW and the flows resulting from it can both be described as direct physical consequences of a band of local contraction travelling from vegetal to animal pole ( S1 and S2 Videos ) . Our work not only provides a quantitative description of the internal hydrodynamics within the oocyte at high resolution , but also explains the origin of the induced flows . The model is based on an analytical solution of the Stokes equation in spherical geometry . The assumption of incompressibility and low Reynolds number ( Re ≈ 2 × 10−4 ) seems to be satisfied for the starfish cytoplasm . Additionally the cytoplasm is described to behave as a Newtonian fluid . The good agreement between measurement and the model shows that the cytoplasm behaves as a simple fluid regardless of its complex structure . Such behaviour is also known for Drosophila [14] . Even though the volume of the cell changes over time ( Fig 1 ) , the assumption of incompressibility can be made as explained in the Methods section . The only input of the model is the oocyte shape and its surface velocities which serve as boundary conditions . We assumed no-slip boundary conditions in order to couple the internal flows with the surface movement . In addition to the internal flows the model predicts the pressure field inside the oocyte . In a simple version our model has only one free parameter , which is the movement of the centre of mass along the AV-axis . This radial displacement model already explains most features of the flows . By extending the model to tangential flows due to local contraction of the cell surface , we could show that tangential surface flows towards the contraction band should also play a role for the internal flows . In the future , this prediction has to be confirmed experimentally by imaging actomyosin flow in the plane of the cortex , as has been done earlier for C . elegans [32] . In contrast to the case of C . elegans , however , here the spatial actomyosin distribution varies strongly in time ( as shown explicitly in S12 Video ) [21] . Similar dynamical distributions might be at work in other organisms and our model with both radial and tangential displacements should be applicable to any approximately spherical cell . Cytoplasmic flows have been shown to be important in the development of multiple organisms [17 , 32] . Recently Boquet-Pujades et al . have introduced a method to calculate an internal pressure field by numerically solving Stokes equation [27] . The great advantage of our analytical model is the feasibility to compute the flow field for videos with several thousand frames , which is computationally very demanding for a numerical procedure . Additionally , different contributions to the flows can be studied individually using our method . For instance , we are able to directly discriminate the effects of radial flows from tangential flows . We note that our model can be used in the future to predict experimentally inaccessible internal flows for non-transparent cells , as it is the case of the Xenopus oocyte , by tracking its surface movement . This is especially interesting because the cytoplasm of Xenopus seems to be characterised by a spatially highly structured distribution of determinants . Our hydrodynamic model can explain the shape and strength of the internal flow field during the SCW . This clearly shows that the flows are a physical consequence of the surface movement . The surface movement is explained by the model resulting from a band of locally increased surface tension moving over a contractile surface [21] . It is known that the cytoplasm of oocytes can behave as an active fluid [40] . However , these contributions to the flows are much smaller compared to the flows driven by the SCW and that we are able to predict with our hydrodynamic model . This shows that during the SCW internal cytoskeletal forces are smaller than the forces due to the surface movement . This is in contrast to findings in other species such as Drosophila [14] and C . elegans [22] , where a complex internal biochemical machinery is involved in flow generation . These cells do not have the possibility to undergo large surface deformations as they are surrounded by a stiff egg shell , in contrast to oocytes of starfish , Amphibia , fish , mouse and other mammals . Indeed , the shape of the oocytes is determined by the cortex in these species of the deuterostome group of animals . Therefore cortical deformations can lead to strong internal flows in oocytes of these species . Hamaguchi and Hiramoto explained polar body extrusion as the result of a locally increased pressure produced by the SCW [4] . By quantifying the pressure field inside the oocyte we demonstrated that it is five orders of magnitude smaller than the pressure difference between inside and outside of the oocyte . Importantly , we have used a value of 5mPa s for the viscosity that we have measured directly with FCS . The fact that this value is close to the one of an aqueous solution again suggests that no relevant cytoskeletal structures are present in the cytoplasm during the stage of the SCW . We conclude that the pressure gradient generated by hydrodynamic flow is too weak as to explain polar body extrusion . We tested this conclusion experimentally by myosin II inhibition , which reduced both shape changes and hydrodynamic flow , but did not abolish polar body extrusion . This result agrees with previous studies based on mechanical manipulation [23] and results known from other species [41 , 42] . We further showed that actin polymerisation is the most likely mechanism for polar body formation . The main function of the SCW then might be coordination of the pinching off of the polar body , which depends on actomyosin contractility . Mixing or moving substances inside the cytoplasm is a common biological function of cytoplasmic flows [16 , 17 , 32] . This becomes especially interesting for systems at low Reynolds number . In this regime hydrodynamic flows driven by movement obeying time reversal symmetry cannot lead to any mixing . Thus mixing is only driven by diffusion or by internally driven flows . Diffusion is a slow process in large systems , though . The SCW of starfish in contrast clearly breaks time reversal symmetry , because it proceeds in one direction only , from the V-pole to the A-pole , and therefore cytoplasmic mixing becomes a potential function of the SCW . In particular , the kymograph from Fig 6E shows that the two periods of upflow and downflow are not simply mirror images of each other , because the downflow does not result from squeezing from the top ( downward slope in the kymograph ) , but rather from relaxation at the bottom ( upward slope in the kymograph ) . Here we could show by simulated particle-tracking that nevertheless after the full SCW most particles end up where they started , because the directional movement of the SCW combines with the spherical symmetry of the soft shell such that the second half of the SCW counteracts the effect of the first half . This excludes intracellular transport of particles of the size of the yolk particles as a potential function of the SCW . Significant displacement of particles can only be detected near the animal pole , where the cell remains deformed . In addition , trajectories of nearby particles are to a large extent parallel to each other . Therefore we can exclude that the SCW contributes to significant mixing of the cytoplasm . However , the wave could still lead to accumulation of constituents much smaller than the yolk particles , which would need to be tested in the future by tracking the movement of fluorescently labelled markers . It could well be that on that smaller scale , the cytoplasm does not behave like a simple Newtonian fluid any more , but like a poroelastic medium as suggested earlier for human cells [35] . Our results show that hydrodynamic flows arise each time when major cell shape changes occur , including cell division , migration and spreading . This also implies that hydrodynamic flow might be induced by artificially effecting cell shape changes , e . g . by optogenetics [43] . At the other extreme , one could induce hydrodynamic flows by purely physical means , thus decoupling them from the biochemistry related to cell shape changes [22] . Together , these new developments open up many perspectives to elucidate if hydrodynamical flows in specific systems have a biological function or are simply by-products of cell shape changes .
For the study of SCWs the oocytes of bat star ( Patiria miniata ) were used . They are obtained once a year from Southern California Sea Urchin Co . , Marinus Scientific , South Coast Bio-Marine , or Monterey Abalone Co . and kept in sea water tanks at the EMBL marine facility . They are kept at 15 °C and fed once a week with shrimp . Sea water was made by dissolving red sea salt mix in water and filtering the water . Ca-free seawater was made from 437 mM NaCl , 9 mM KCl , 22 . 9 mM MgCl2 · 6 H2O , 25 . 5 mM MgSO4 · 6 H2O , 2 . 1 mM NaHCO3 . Oocytes are isolated from a biopsy of the ovaries that is obtained from the dorsal side of a starfish arm using a surgical biopsy puncher . The biopsy is put into calcium-free sea water at pH 8 with 50 mM L-Phenylalanine ( Sigma ) added for 15 min to 20 min . This prevents unwanted oocyte maturation . The biopsy is transferred to filtered sea-water with 100 μm Acetylcholine ( Sigma ) . This leads to a contraction of the ovaries , so that the oocytes are expunged from the ovary . Oocytes are kept in plastic Petri-dishes with filtered sea water at 14 °C for up to 3 days . For most cells in the data set , the jelly coat was removed by either treatment with actinase ( 30 min in 0 . 1 mg/ml and then 3 times washing with FSW ) or by Filtered Seawater adjusted to pH4 ( 3 min incubation and 3 washes ) . Maturation is induced by adding 10 μM 1-Methyladenin ( Sigma ) . To inhibit non-muscle myosin II activity , oocytes were treated with 300 μM ( - ) -Blebbistatin ( Abcam , stock in DMSO ) . Cells were pre-treated for 1h before application of hormone . To visualise actin filaments , oocytes were injected with mRNA for Utrophin-CH domain-mEGFP [44] ( construct gift from Bill Bement ) . Injections were performed as described in [45] . Injected oocytes were incubated overnight at 13°C to allow for expression of fluorescent proteins before being matured and imaged as described above . Full 3D imaging of the oocyte was performed on a Zeiss LSM 880 , using the Airyscan fast mode , 600 Hz scan speed , with a 25x water-immersion objective . For this , cells were injected with Alexa647-labelled 10kDa Dextran as a contrast agent , treated with maturation hormone and placed freely into a glass-bottom dish before imaging . Z-slices with a distance of 2 μm were taken through the whole oocyte continuously throughout meiosis . The surface reconstruction was performed with the commercial microscopy image analysis software Imaris ( Imaris 9 . 2 . 0 , Bitplane AG , Oxford Instruments , https://bitplane . com ) . For this purpose the raw data was loaded into Imaris as a series of z-stacks . The pixel intensity showed a global decrease with increasing z-stack direction . To compensate this loss of signal for deeper optical sections , we applied an attenuation correction . We used a ratio of 256 to 30 from front to back ( in z-direction ) , to linearly interpolate pixel intensities , resulting in overall homogeneous pixel intensity levels for all regions of interest containing a signal ( foreground ) . The two values stand for the mean intensities in the front layer compared to the back layer of the full z-stack . Then we used this corrected data set to use Imaris’ surface reconstruction algorithm . This algorithm applies the marching cubes algorithm from Lorensen and Cline [46] and uses an absolute pixel intensity criterion for the underlying thresholding . The surface roughness of the reconstruction was set to 2 μm which provided the most stable solutions in the range from 0 . 1 μ1 to 15 μm . Internal surfaces with surface area A < 1000 μm2 were excluded to only extract the cell surface . Imaging of cell shape changes and hydrodynamic flow was mainly done in 2D using custom-made imaging chambers in a Zeiss Cellobserver transmission light microscope at one frame per second temporal resolution at 21 °C . A Plan-Apochromat 20x/0 . 8 objective was used together with an exposure time of 1 s . The cells are chosen such that the AV-axis of the oocyte visually lies in the imaging plane . Oocytes for which it was found in image analysis that the AV-axis does not lie in the imaging plane were not taken into account for further evaluation . 2D fluorescent images and most of the cells used in Fig . S3 Fig were acquired with a Leica SP5 confocal microscope with a 1 . 1 NA HC PL APO ×40 water immersion objective . It was equipped with a fast Z-focusing device ( SuperZ Galvo stage ) ( Leica Microsystems ) . Initially , each frame is Gaussian filtered and edge filtered by applying a two-dimensional Sobel filter . Next , a preliminary contour is initialised that is iteratively adjusted for the edges in the image . The active contour algorithm of the Python package scikit-image [47] is used for this step . The contour is optimised by weighting contour length , smoothness and the edges in the image . The surface quantification is sketched in Fig 2B . The surface points are expressed in planar polar coordinates with the centre of mass as the origin . Thus different cells and time points can be compared irrespective of the position in the frame . This results in a radius function r ( θ , t ) . By spatially Fourier transforming this function and cutting off high spatial frequencies , the radius function is smoothed . The AV-axis is detected by finding the points that develop from a local minimum to a local maximum back to a local minimum during the SCW as it is shown in Fig 2C and 2D . The radius function is shifted so that the animal pole lies at θ = 0 . Only the symmetric parts are kept in the Fourier transform and therefore the symmetrised radius function is expressed as r ˜ ( θ , t ) = ∑ k = 0 6 a k ( t ) cos ( k θ ) . The radial surface velocities are determined by calculating the numerical derivative of two subsequent time steps . In order to reduce noise , the coefficients ak ( t ) are smoothed by cutting off high temporal frequencies . The cut off frequency is set for each order of k individually . The radial surface velocity is then calculated as: v ( θ , t ) = r ˜ ′ ( θ , t + Δ t ) - r ˜ ′ ( θ , t ) Δ t = ∑ k = 0 6 ( a ˜ k ( t + Δ t ) - a ˜ k ( t ) ) cos ( k θ ) Δ t , where a ˜ k ( t ) is the temporally smoothed Fourier coefficient for mode k and time point t . Cytoplasmic flows are experimentally measured by particle image velocimetry ( PIV ) [33 , 34] . In PIV each image is segmented and each segment is correlated with the subsequent image . The image segment is assumed to have flown where correlation is highest . We use the algorithm from the Python package openpiv [34] . Especially in regions with little contrast ( the nucleus , outside of the oocyte ) the algorithm encounters problems . We apply several sorting steps to the vertices found by the algorithm in order to identify ill-detected vertices . Different filters are applied for this task in the following order: At first a local-median filter is applied . A vector is discarded , if its strength and direction deviates too much from the surrounding flows . Next , a global value filter is applied , discarding flows that are unphysically strong . For each flow a signal to noise ratio of detection is stored that can be used to exclude flows . As we are only interested in the flows inside the oocyte , other flows can be discarded . In the end , we apply a global standard deviation filter , that means that if a flow deviates too much from the global stream pattern , it is sorted out . We parametrise the rotationally symmetric surface as f → ( φ , θ ) = R ( θ ) ( cos θ cos φ cos θ sin φ sin θ ) , ( 1 ) where θ ∈ [ - π 2 , π 2 ) is the polar angle defined between the xy-plane and the point of consideration . E . g . θ = - π 2 defines the point x → = ( 0 0 - 1 ) . φ ∈ [0 , 2π ) is the azimuthal angle measured between xz-plane and the point of consideration . The coordinates θ = 0 , φ = π 2 define the point x → = ( 0 1 0 ) . We then get the following entries for the Weingarten matrix: aφφ=−1− ( ∂θR ) tanθRR2+ ( ∂θR ) 2 , aθθ=−1+ ( − ( ∂θR ) 2+R ( ∂θ∂θR ) ) ( R2+ ( ∂θR ) 2 ) R2+ ( ∂θR ) 2 . ( 2 ) From the Weingarten matrix mean curvature ( H ) and Gaussian curvature ( K ) are computed as: H = 1 2 tr ( a ) = a φ φ + a θ θ 2 , K = det ( a ) = a φ φ a θ θ . ( 3 ) More details of the calculations are given in S1 Appendix . Volume and surface area can then be calculated directly from the Fourier coefficients up to kmax = 4 as: V = ∫ Ω d Ω = ∫ 0 π d θ ∫ 0 R ( θ ) d R 2 π R 2 sin ( θ ) = 2 π 3 ∫ 0 π d θ ( ∑ k = 0 4 cos ( k θ ) a k ) 3 sin ( θ ) = 2 π 3 ( 2 a 0 3 - 2 5 a 0 2 ( 5 a 2 + a 4 ) + 2 105 a 0 ( 105 a 1 2 + 147 a 2 2 - 126 a 1 a 3 + 153 a 3 2 - 114 a 2 a 4 + 155 a 4 2 ) - 2 15015 ( 3861 a 2 3 + 8437 a 2 a 3 2 - 429 a 1 2 ( 7 a 2 - 13 a 4 ) - 9581 a 2 2 a 4 + 2 15015 ( 5369 a 3 2 a 4 + 7943 a 2 a 4 2 + 777 a 4 3 - 286 a 1 a 3 ( 45 a 2 + 49 a 4 ) ) ) . For the surface area A we get: A = ∫ ∂ Ω d A = ∫ 0 π d θ R 2 2 π sin ( θ ) = 2 π ∫ 0 π d θ ( ∑ k = 0 4 cos ( k θ ) a k ) 2 sin ( θ ) = 4 π 315 ( 315 a 0 2 - 42 ( 5 a 2 + a 4 ) a 0 + 105 a 1 2 + 147 a 2 2 + = 4 π 315 ( 153 a 3 2 + 155 a 4 2 - 126 a 1 a 3 - 114 a 2 a 4 ) . We calculate hydrodynamic flow for a Newtonian , incompressible fluid at low Reynolds number . Reynolds number of starfish internal flows is about 2 × 10−4 , thus we can use the Stokes equation . Even though the volume is changing over time , the assumption of local incompressibility can still be made for the fluid . For each time point the internal flows are calculated anew from the current surface movement . This is valid because the Stokes equation is time-independent , thus the past history of the flow is not required . We solve the standard form of the Stokes equation [48 , 49]: μ ∇ → 2u i = ∂ p ∂ x i ( 4 ) ∇ → · u → = 0 ( 5 ) with the standard no-slip boundary condition at the cell surface . Thus the movement of the cell envelope serves as boundary condition for the internal flows . Lamb has given a solution for spherical coordinates [48] that Happel and Brenner have adjusted to external flows emerging from boundary flows on a sphere [50] . The solution to the Stokes equation for the flows inside a moving spherical shell is given by: u → = ∑ n = - ∞ ∞ [ ∇ → ϕ n + ∇ → × ( r → χ n ) + A r 2 ∇ → p n + B r → p n ] , ( 6 ) with: A =n + 3 2 μ ( n + 1 ) ( 2 n + 3 ) ( 7 ) B =- n μ ( n + 1 ) ( 2 n + 3 ) ( 8 ) and three fields of solid harmonic functions of order n that have to determined from the boundary conditions: pn ( r , θ , φ ) , ϕn ( r , θ , φ ) and χn ( r , θ , φ ) , where p = ∑ - ∞ ∞ p n is the pressure field . For the flows inside a sphere of radius a it holds: ϕ n = χ n = p n = 0 . n < 1 ( 9 ) In order to calculate the other orders , three fields of spherical harmonics have to be calculated from the velocity field at the surface u → ( a , θ , φ ) . u r ( a , θ , φ ) = ∑ n = 0 ∞ X n ( θ , φ ) ( 10 ) - r ∇ → · u → ( a , θ , φ ) = ∑ n = 0 ∞ Y n ( θ , φ ) ( 11 ) r → · ( ∇ → × u → ( a , θ , φ ) ) = ∑ n = 0 ∞ Z n ( θ , φ ) . ( 12 ) They are then matched up for n > 0 in the following way: p n ( r , θ , φ ) = μ ( 2 n + 3 ) n a r n a n ( Y n - ( n - 1 ) X n ) ( 13 ) ϕ n ( r , θ , φ ) = a 2 n r n a n ( ( n + 1 ) X n - Y n ) ( 14 ) χ n ( r , θ , φ ) = 1 n ( n + 1 ) r n a n Z n . ( 15 ) This result has been generalised by Brenner in a perturbation ansatz for the flows outside a slightly deformed moving sphere [51] . We here give it for the flows inside the surface . If the radius for 0 < ϵ < 1 is given as: r ( θ , φ ) = a ( 1 + ϵ f ( θ , φ ) ) , ( 16 ) where f ( θ , φ ) is a linear combination of spherical harmonics , then the velocity field and the pressure field can also be expanded in orders of ϵ: u → = ∑ k = 0 ∞ ϵ k u → ( k ) , p = ∑ k = 0 ∞ ϵ k p ( k ) . The zeroth order is set to the velocity field of the perturbed sphere U → ( θ , φ ) : u → ( 0 ) | r = a = U → ( θ , φ ) and for higher order k > 0 this leads to: u → ( k ) | r = a = - ∑ j = 1 k ( a f ( θ , φ ) ) j j ! ∂ j u → ( k - j ) ∂ r j | r = a ( 17 ) For each perturbation order the velocity field and pressure field is calculated separately as for an unperturbed sphere , from the equations given previously . By assuming rotational symmetry of the system , numerous terms in this derivation simplify . More details of the calculations are given in S1 Appendix . The diffusion constant of 50 kDa fluorescently labelled dextran was determined by fluorescence correlation spectroscopy ( FCS ) ( see [52] for characterisation of Dextrans ) inside the starfish oocyte Dcyt and in plain water DH20 . From Stokes-Einstein equation D = k B T 6 π μ R we get the ratio of viscosities as μ c y t μ H 20 = D H 20 D c y t . 3D simulations of the oocyte shapes were performed following Bischof et al . [21] using the software SurfaceMaster [53] . The surface is modelled by a triangular mesh with locally varying surface tension . Surface tension was locally set by a Gaussian shaped band σ ( θ ) = A exp ( - ( θ - θ 0 ) 2 s 2 ) , with the amplitude A , position θ0 and width s . For each time step independently a surface Hamiltonian with the contraction band at a specific position is minimised in steady state due to overdamped dynamics of the system . The Hamiltonian reads: H = ∫ A d A ( κ b H 2 + σ ( x → ) ) + ∫ A 0 d A 0 ( K α 2 α 2 + μ ˜ β ) + k V ( V - V 0 ) 2 , where κ b , K α , μ ˜ and k V are the bending rigidity , stretch modulus , strain modulus and volume modulus respectively . They are fixed for the whole simulation . α and β are the strain invariants computed from stretch and shear of the undeformed sphere A0 . V0 is the reference volume . The parameters were chosen following Bischof et al . [21]: Initial radius of the oocyte: 90 μm , μ˜=0 . 5nNμm−1 , Kα = 1 nN μm−1 , κb = 0 . 2 pN μm and kV = 0 . 01 nN/μm2 . In order to model the cytoplasmic flows due to a tangential movement of the cell surface , we need to calculate the effect of a localised contraction band on the tangential surface movement . The results of these calculations are shown in Fig 5D . For this task we follow the work by Mayer et al . [32] who modelled the cortex of C . elegans as a viscoelastic medium . We also neglect the elastic contribution due to the slow dynamics of the system and thus consider a purely viscous medium . It is assumed that the microscopic restructuring of the cortical network is on timescales faster than the flow dynamics . They have shown that the viscous flow v in a 1D system with contraction C ( x ˜ ) is described by the following differential equation: - ∂ C ∂ x ˜ = ∂ 2 v ∂ x ˜ 2 - b 2 v , with x ˜ a normalised length scale and b = L l the ratio between the length of the system and the spatial decay length of the velocity . The contraction field C ( x ˜ ) thus determines the local strength of isotropic contraction and models the decrease in surface area by locally increased surface tension in our case . The boundary conditions are set to v ( 0 ) = v ( L ) = 0 and v′ ( 0 ) = v′ ( L ) = 0 , because of symmetry of the oocyte . We solved it for a Gaussian contraction band C ( x ˜ ) = A exp ( - ( x ˜ - x ˜ 0 ) 2 s 2 ) . This equation is solved by v ( x ˜ ) = w 1 v 1 + w 2 v 2 , where v 1 ( x ˜ ) = exp ( b x ˜ ) , v 2 ( x ˜ ) = exp ( - b x ˜ ) , w 1 ( x ˜ ) = 1 2 b ∫ x ˜ d x ′ ∂ C ( x ′ ) ∂ x ′ exp ( - b x ′ ) and the function w 2 ( x ˜ ) = - 1 2 b ∫ x ˜ d x ′ ∂ C ( x ′ ) ∂ x ′ exp ( b x ′ ) . We solved this system of equations analytically using the commercial software Mathematica and the resulting tangential flows are used as surface movement in the hydrodynamic calculations . Due to linearity of the Stokes equation , different contributions to the internal flows can be calculated separately . To account for CM movement during imaging rigid body movement with varying velocity along the AV-axis is added to the flows emerging from radial surface movement . The shift velocity is fitted to the experimentally observed flows for each time step individually . Tangential surface movement is modelled by adding the flows emerging from a contraction band at position x ˜ 0 and strength A . The width of the contraction band is set to s = 0 . 63 , which was a result of the shape simulation . The ratio of length scales is set to b = 3 following the result from Mayer et al . [32] . This is justified as we have observed that the flows emerging from the contraction band vary only little for varying b in the range of consideration . The parameters x ˜ 0 and A are found by a brute force approach . The parameter set that showed minimal root mean squared deviation between model and experiment is chosen . Particle trajectories were calculated by integrating the experimentally obtained flow field numerically . | As already noted by Aristotle , life is motion . On the molecular scale , thermal motion leads to diffusive transport . On cellular scales , however , diffusion starts to become inefficient , due to the general property of random walks that their spatial excursions grow less than linear with time . Therefore more directed transport processes are needed on cellular scales , including transport by molecular motors or by hydrodynamic flows . This is especially true for oocytes and eggs in early animal development , which often have to be large in order to store sufficient amounts of nutrients . Here we use starfish oocytes as a convenient model system to investigate the nature and function of cytoplasmic flows in early development . These cells are very large and optically transparent , and therefore ideal for live cell imaging that here we combine with image processing and mathematical modelling . This approach allows us to demonstrate that the experimentally observed cytoplasmic flows during early development are a direct consequence of surface contraction waves that deform the soft and contractile eggs . Additionally we show that despite its microscopic complexity , the cytoplasm behaves like a Newtonian fluid on the cellular scale . Our findings impose strong physical limits on the potential biological function of these flows and suggest that also other cellular systems that are soft and contractile might experience large cytoplasmic flows upon cell shape changes , for example during cell migration or division . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"cell",
"physiology",
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"classical",
"mechanics",
"fluid",
"mechanics",
"geometry",
"animals",
"cell",
"polarity",
"germ",
"cells",
"oocytes",
"mathematics",
"echinoderms",
"starfish",
"cellular",
"structures",
"an... | 2018 | Cytoplasmic flows in starfish oocytes are fully determined by cortical contractions |
We evaluated the sensitivity of the dengue surveillance system in detecting hospitalized cases in ten capital cities in Brazil from 2008 to 2013 using a probabilistic record linkage of two independent information systems hospitalization ( SIH-SUS ) adopted as the gold standard and surveillance ( SINAN ) . Sensitivity was defined as the proportion of cases reported to the surveillance system amid the suspected hospitalized cases registered in SIH-SUS . Of the 48 , 174 hospitalizations registered in SIH-SUS , 24 , 469 ( 50 . 7% ) were reported and registered in SINAN , indicating an overall sensitivity of 50 . 8% ( 95%CI 50 . 3–51 . 2 ) . The observed sensitivity for each of the municipalities included in the study ranged from 22 . 0% to 99 . 1% . The combination of the two data sources identified 71 , 161 hospitalizations , an increase of 97 . 0% over SINAN itself . Our results allowed establishing the proportion of underreported dengue hospitalizations in the public health system in Brazil , highlighting the use of probabilistic record linkage as a valuable tool for evaluating surveillance systems .
Dengue is the most important arboviral disease in the world due to the associated morbidity , mortality and economic burden [1–4] . In Brazil , the disease has become a major public health challenge , with 5 . 8 million probable cases , 555 thousand hospitalizations and 3 , 000 deaths reported from 2002 to 2014 . During this period , noteworthy epidemiologic shifts were observed in the country , including an increase in the number of smaller cities experiencing transmission , changes in affected age groups and increases in the proportion of severe cases [5–8] . Dengue fever is a mandatorily reportable disease in Brazil since the reintroduction of the virus to the country in 1986 [7] . The dengue surveillance system relies on passive reporting from healthcare facilities ( outpatient and hospital ) , with uniform standardized forms used throughout the country . The data from these forms are entered into the National Reportable Disease Information System—SINAN ( Sistema de Informação de Agravos de Notificação ) , which is the main source for dengue related information in Brazil . As expected , passive surveillance likely results in under-reporting , especially with regard to undifferentiated febrile and atypical forms of dengue [9 , 10] . This limitation may lead to an underestimated burden of the disease , which can result in inappropriate allocation of resources for prevention and control activities[11–13] . The continuous evaluation of surveillance systems with respect to their attributes , is critical for maximizing the efficacy and the utility of such systems and for producing more reliable indicators [14] . The sensitivity of a surveillance system is the capacity to identify cases of the disease and , therefore , is a crucial attribute for a system to reach its goals . However , the evaluation of the sensitivity is usually a challenge due the lack of a gold standard to provide the true number of cases for comparison [15] . The record linkage of different information systems is an alternative to improve disease estimates and evaluate the sensitivity of surveillance systems [16–18] . Hospitalizations in the public health system in Brazil ( National Unified Health System-SUS ) are registered in a specific information system that is independent of the surveillance system . The objective of this study was to estimate the sensitivity of the national dengue surveillance system ( SINAN ) for detecting hospitalized dengue cases in the National Unified Health System ( SUS ) in 10 state capitals between 2008 and 2013 in Brazil .
This study was approved by the Committee for Ethics in Research of the Federal University of Goiás in accordance with the ethics principles established in Resolution 466/12 of the National Council for Health of Brazil and all data analyzed were anonymized . This an observational , descriptive and cross-sectional epidemiologic study based on the probabilistic record linkage between the databases of the National Unified Health System’s Hospital Information System ( we use the Portuguese acronym SIH-SUS ) and the National Reportable Disease Information System ( SINAN ) . Study area and period: We selected ten state capitals located in the four dengue endemic regions of Brazil for the study . These municipalities contributed 10% of the total hospitalized dengue cases from 2008 to 2013 , the study period . Although the municipalities vary in size , they were similar with regard to epidemiological aspects of dengue such as the historical circulation of the four viral serotypes ( DENV 1 , 2 , 3 , 4 ) and the occurrence of epidemics . The selected capitals were the following ( city ( state ) ) : North Region—Manaus ( AM ) , Boa Vista ( RR ) ; Northeast Region–Fortaleza ( CE ) , Natal ( RN ) , São Luis ( MA ) , Teresina ( PI ) ; Southeast Region—Rio de Janeiro ( RJ ) , Belo Horizonte ( MG ) ; and Central-West Region–Goiânia ( GO ) , Campo Grande ( MS ) . Dengue hospitalized cases: The organization of the Public Health System ( SUS ) in Brazil has been described in detail elsewhere [19] . Briefly , this system provides universal healthcare to all persons residing in Brazil , outpatient and inpatient , at no charge to patients . Approximately 70% of inpatient medical services in the country are provided by SUS [20] . Hospitalizations in SUS requires completion of a standard form that captures patients’ personal data , symptoms , and the initial diagnosis coded according to the 10th revision of the International Classification of Disease ( ICD-10 ) . This form and further information on diagnoses , treatment , test results , and billing are the main data recorded by the SIH-SUS , which is an administrative database standardized throughout Brazil . This system captures data on all hospitalizations paid by the public health system for public and contracted hospitals . The resulting data is checked and validated by local health authorities and subsequently transmitted to regional and national levels . For this study , we extracted the records for suspected dengue cases that were hospitalized in SUS , using admission ICD10 codes A90 and A91 for Dengue and Dengue Hemorrhagic Fever . We used the SIH-SUS database , updated in January 2014 , for the years 2008 to 2013 . Reported dengue Cases: The organization of the surveillance system has been previously described [7] . In summary , dengue is a mandatorily reportable disease and the system relies on the notification of all suspected cases at public and private health facilities based on the attending clinician’s initial clinical diagnosis ( not laboratory confirmed ) . SINAN is the official information system for entering and processing the data for reported dengue cases throughout Brazil . It uses uniform standardized forms that capture data related to patient identification as well as the main characteristics of the illness . Data on patient hospitalization during each dengue disease episode is also recorded and this procedure is independent of SIH-SUS routines . During the study period , the Ministry of Health ( MoH ) of Brazil adopted the case definitions proposed by the Pan American Health Organization ( PAHO/WHO ) for suspected and confirmed cases of Dengue Fever ( DF ) and Dengue Hemorrhagic Fever ( DHF ) . Additionally , the MOH adopted an intermediate final classification “dengue with complications” ( DwC ) that includes all the cases that did not fulfill the DHF diagnosis criteria and ones for which the DF classification was not satisfactory due to the severity of the clinical and laboratory outcomes presented . The final classification of each case is only performed after the patient’s discharge or conclusion of the case investigation . Identification and deletion of duplicated records are conducted at the local level using SINAN’s automated routine and the resulting data is transferred to state and federal levels . Failure to transmit data from the local level for a period of two months is penalized by cancellation of financial resources destined to the municipality . Initially we identified and excluded duplicate records and inconsistencies in both databases . During the process of standardization of databases a total 25 , 047 duplicate records were excluded ( 23 , 232 from SINAN and 1 , 815 from SIH-SUS ) . Using the cleaned databases , we generated descriptive findings of dengue related hospitalizations by gender , age group and diagnosis according to ICD-10 in each of the two systems . The number of dengue-related hospitalizations in each state capital derived from SIH-SUS was defined as the gold standard for the subsequent sensitivity analysis . A probabilistic record linkage of all dengue related-hospitalized recorded in SIH-SUS and all reported cases in SINAN was performed . We used RecLinkIII software , which implements the probabilistic record linkage methodology and is widely used for this purpose in Brazil [21 , 22] . The output of this software includes a score for the links formed to assess the agreement and disagreement of the variables selected for the linkage . The higher the score , the greater the probability of finding a true matched pair . Prior to record linkage , both databases underwent a pre-processing stage of quality analysis to minimize errors and increase the likelihood of finding matched records . These procedures comprised mainly standardization of the variables selected as matching and/or blocking variables . The record linkage process consisted of the following steps: The sensitivity of the surveillance system for detecting hospitalized cases in SIH-SUS was calculated using two approaches . In method 1 , the numerator consisted of the total number of true pairs , which were also described as being hospitalized in SINAN; the denominator consisted of the total number of dengue related hospitalizations in SIH-SUS . In method 2 , the numerator consisted of the total number of true pairs regardless of the reported hospitalization status in SINAN; the denominator consisted , again , of the total number of dengue related hospitalizations in SIH-SUS . To estimate the overall number of hospitalization was adopted the formula: ( SINAN hospitalization X SIH-SUS ) /Matched pairs[23] Data was processed and analyzed using Tabwin , Reclink III version 3 . 1 . 6 and Microsoft Office 2010 .
During the study period , 1 , 203 , 212 suspected dengue cases were reported in the 10 selected municipalities and of these 36 , 145 ( 3 . 0% ) were hospitalized according to SINAN . In SIH-SUS , 48 , 174 dengue hospitalizations were registered during the same period . Overall , the number of hospitalizations recorded in SIH-SUS was 33 . 3% higher than those recorded in SINAN ( Fig 1 ) . However , this pattern was not observed in the municipalities of São Luis ( MA ) , Rio de Janeiro ( RJ ) and Campo Grande ( MS ) where the number of hospitalizations was higher in SINAN exceeded those recorded in SIH-SUS by 5 , 207 ( 31 . 2% ) ( S1 Table ) . The distribution of hospitalizations by sex showed a similar pattern in both information systems , with females accounting for about 51% of the records . The proportion of hospitalized cases in children under 15 years was higher in SIH-SUS ( 45 . 0% ) compared to SINAN ( 38 . 5% ) . This pattern was observed in most capitals except in Boa Vista ( RR ) and Teresina ( PI ) where the proportion of hospitalizations in children were 55 . 8% and 29 . 6% in SINAN and 44 . 0% and 28 . 3% in SIH-SUS respectively . Patients with DF accounted for 43 . 4% and 83 . 1% of hospitalizations in SINAN and SIH-SUS , respectively . Almost twice as many hospitalizations due to suspected DHF were observed in SIH-SUS compared with those in SINAN: 8 , 123 ( 17 . 0% ) vs . 3 , 346 ( 9 . 2% ) records , respectively . However , 14 , 030 hospitalizations ( 38 . 8% ) were classified as DwC in SINAN , highlighting that some of these cases may include suspected DHF inpatients . Only in Boa Vista ( RR ) there was a higher number of hospitalizations due to DHF in SINAN ( Table 1 ) . The probabilistic record linkage identified 24 , 469 records common to both databases . Overall , among the total number of pairs found , 12 , 995 ( 53 . 1% ) had the field for the variable “hospitalization” completed in the SINAN record , with the highest proportion observed in 2008 ( 26 . 6% ) ( Table 2 ) . However , different results were observed for the years of 2010 and 2009 , with , respectively , 23 . 5% and 2 . 2% of the pairs with information on hospitalization available in SINAN . The combination of the two systems allowed identification of 71 , 161 hospitalizations , which represented increases of 97 . 0% and 47 . 7% in the number previously registered in SINAN ( 36 , 145 ) and SIH- SUS ( 48 , 174 ) , respectively ( Table 2 ) . The sensitivity of the surveillance system in detecting cases hospitalized in SUS was 27 . 0% ( 95% CI 26 . 6 to 27 . 4 ) , when considering only records with information regarding hospitalization in both systems ( method 1 ) . Using this approach , the lowest sensitivity was observed in 2009 at 8 . 3% ( 95%CI: 7 . 4 to 9 . 3 ) and the highest in 2013 at 41 . 9% ( 95% CI 40 . 2–43 . 6 ) . Among the municipalities the highest and lowest sensitivities were observed in Campo Grande ( MS ) at 78 . 5% ( 95%CI: 67 . 5 to 86 . 6 ) in 2012 and 0% in 2008 , when none of the 15 hospitalizations were registered in SINAN . The sensitivity of the surveillance system including all records in SINAN regardless of the hospitalization status ( method 2 ) was almost twice as high as that calculated by the first approach . Using this method , the cumulative sensitivity was 50 . 8% ( 95%CI: 50 . 3–51 . 2 ) , with the lowest value observed in 2009 at 41 . 2% ( 95%CI: 39 . 6–42 . 9 ) and the highest in 2013 at 79 . 1% ( 95%CI: 77 . 7–80 . 4 ) . Among the municipalities , the highest value was observed in Campo Grande ( MS ) at 99 . 2% ( 95%CI: 98 . 3–99 . 6 ) in 2010 and the lowest in Teresina ( PI ) at 18 . 1% ( 95%CI: 14 . 8–21 . 9 ) in 2008 ( Table 2; Fig 2 ) . The comparison of matched pairs according to the initial clinical suspicion of dengue from SIH-SUS and the final classification according to the surveillance system is presented in Table 3 . Among the 4 , 515 pairs hospitalized in SIH-SUS and classified as DHF ( A91 ) , 35 . 5% ( 1604 ) were classified in SINAN as DF , 31 . 2% ( 1 , 407 ) as DwC and 15 . 1% as DHF / DSS . Of the 19 , 954 patients hospitalized with a classification of DF ( A90 ) , 57 . 5% ( 11 , 482 ) had a final classification of DF , 17 . 2% ( 3 , 456 ) were reclassified as DwC and 4 . 9% DHF and DSS . The percentage of pairs that lacked a classification by the surveillance system was 11 . 6% .
In this study , we demonstrated the occurrence of a larger number of hospitalized dengue fever cases in Brazil than that captured in the national surveillance system ( SINAN ) . The use of probabilistic record linkage of SINAN data and the national hospitalization system ( SIH-SUS ) database expanded the estimate of dengue hospitalizations by over 49 . 2% ( 35 , 016 ) hospitalizations in the ten cities of the study , compared to the data available from SINAN alone . The dengue surveillance system should capture all reported cases from both public and private health systems . We therefore expected that the total number of cases in the SINAN surveillance database would exceed that of SIH-SUS , which only covers the public health system . However , this was not observed in general , except for three municipalities , where the higher number of hospitalizations observed in SINAN when compared to SIH-SUS could reflect an improved participation from private hospitals in surveillance activities . The surveillance system did not allow identifying if the reporting health unit is public or private , but the inclusion of this data would improve the quality and representativeness of the surveillance system . The adoption of two different approaches to evaluate the sensitivity allowed a more comprehensive analysis of the current operational aspects of the surveillance system . Most hospitalized suspected cases of dengue were not reported to the surveillance system or were reported before hospitalization . Only 12 , 995 ( 53% ) of the reported dengue cases had data on the hospitalization status in the surveillance information system . The completion of the appropriate data field for hospitalization status in the SINAN surveillance form would enable the surveillance system to capture the burden of the disease and its trends over time . Interpretation of the results of our analysis for 2008 and 2009 requires caution . A new version of SINAN software was implemented nationwide in January 2007 . The updated version did not allow filling the field for the hospitalization variable for cases classified as DF , even if the data was available from the investigation form . As this limitation was a technical oversight , it was corrected with a software patch in late 2009 . Since the attributes of a surveillance system are intrinsically interconnected , the sensitivity in this case was influenced by the limited capacity to adapt to changing information needs or operating conditions , in other words , the flexibility of the system . Although SINAN and SIH-SUS are completely independent of each other , the distribution of sex and age groups of hospitalized patients was very similar for the cases captured by each of the two systems . However , the observed differences in initial clinical diagnosis at the moment of the admission and the final classification performed after the full course of the illness highlight some of the difficulties discussed over time in the disease classification [24 , 25] . This disagreement reinforces the importance of the surveillance routines adopted by the Brazilian surveillance system . The final classification of cases reflects the investigation conducted locally by public health professionals based on chart review and clinical and laboratory findings , in accordance national guidelines of the Ministry of Health[26] . Annually , 30 to 40% of reported cases are discarded by the surveillance system in Brazil based on these investigation routines; confirmed cases are classified according to their clinical outcome . During dengue outbreaks , the health system is usually overwhelmed and mild cases may be reported without follow up , but additional information is mandatory for cases with severe outcomes . The results of these investigation efforts by the public health system serve to guide the adoption of control measures and organization of the healthcare network for present and future transmission periods . In our study , the final classification of cases available in SINAN presented a low concordance with the initial diagnosis by physicians at the time of hospitalization . Only 15 . 1% of those hospitalized as a suspected case of DHF in the public health system met the proposed criteria for this definition; the high proportion ( 22 . 1% ) of hospitalized DF cases that were reclassified to more severe forms , DwC or DHF , underscore the difficulties of using the WHO protocol in routine epidemiological surveillance [27–29] . Greater accuracy in the identification of severe cases was attempted in SINAN by including classification of Dengue with Complications ( DwC ) . DwC is not a classification option in SIH-SUS . Additionally , the strain of virus , specific sequence of dengue virus infection , comorbidities , the age and possibly the ethnic composition of patient groups may also influence the different patterns observed in different cities and in different years . Better knowledge of the indicators of morbidity and mortality is essential for assessing the burden of dengue and for measuring the impact of intervention strategies [30 , 31] . The availability of different data sources in Brazil is a most welcoming scenario as it has the potential to increase the representativeness of the surveillance system [32–35] . In this context , the integration of these different sources should be seamless , with automated reports from SIH-SUS to SINAN as suspected cases of mandatorily notifiable diseases are hospitalized . Efforts like this should also be extended to the health insurance companies that also rely on information system for payment purposes . Other studies have emphasized the need for evaluating the underreporting of hospitalized dengue cases . In Cambodia , data from the routine surveillance system was compared with data from active surveillance; 1 . 1- to 2 . 4-fold more hospitalized cases were detected by active surveillance [36] . Similar results were found in another study in Cambodia and Thailand with 1 . 4 and 2 . 6-fold more hospitalizations , respectively [37] . In Puerto Rico , the estimated underreporting for inpatients was 42 . 0% [23] . In Belo Horizonte , Brazil an evaluation of the surveillance system found similar results with a sensitivity of 63% in detecting hospitalized patients [38] . Underreporting of non-hospitalized patients is even more pronounced . The use of active surveillance in Thailand and Cambodia detected 8 . 7- and 9 . 1- fold more cases , respectively , than routine surveillance [37] . Other studies showed higher levels of underreporting ranging from 3 . 9–29 times in Cambodia and 14 . 0–28 times in Nicaragua [36 , 39] . The significant undercounting of non-hospitalized patients is likely do to the fact that many infected persons , especially those with non-severe dengue , do not seek healthcare . The following limitations may have influenced the estimates of sensitivity in our study . We used the data from SIH-SUS as a gold standard , but this system includes only hospitalizations in the public health system , and excludes private healthcare facilities . The private health system is not required to make databases available to research initiatives this was the reason we could not perform an evaluation of underreporting in the private sector . A second potential limitation lies in the methodology of the probabilistic record linkage . Although we have processed a manual review of the pairs with the objective of minimizing errors , it is possible that some pairs were considered true but in fact consisted of different individuals ( i . e . , false positives ) , while others may not have been correctly identified ( i . e . , false negatives ) [40] . Although cohort studies are considered the best method for epidemiological estimates of disease incidence , our study confirms the practicality of comparing different databases by using probabilistic methods as a viable alternative for evaluation of surveillance systems . To our knowledge , this is the first study that uses the probabilistic linkage of the databases of SINAN and SIH-SUS in the evaluation of the surveillance of hospitalized dengue cases in multiple cities . Some of our findings reinforce the usefulness of such methodology . The first concerns the revised WHO dengue classification , adopted by Brazil in 2014 . Our study may serve as a reference for comparisons in the future on some attributes of the surveillance system employing the new classification . Additionally , the introduction of a dengue vaccine requires a stable , robust surveillance system that provides reliable counts of hospitalized dengue cases , among other indicators , in order to define priority areas and populations for vaccination trials and cost effectiveness studies . | This manuscript address essential issues regarding the current and future challenges of a dengue surveillance system . Dengue fever is one of the major public health threats in a large area of the world , including Brazil which represents around 70% of the cases in the Americas . The dengue surveillance system in Brazil was established in 1986 , but evaluations of this surveillance system were rarely conducted . The need for accurate data is of paramount importance due to not only the increase of severe cases in the past decade , but also to monitor the impact of a vaccine as soon as it occurs . The evaluation of the sensitivity of surveillance system is a challenge , due to the lack of a gold standard . In our study , we took advantage of two very well structured and independent information systems with nationwide coverage . The hospitalization information system was defined as the gold standard for hospitalized cases in the public health system and compared to the notifiable diseases information system using a probabilistic record linkage . Therefore , we were able to evaluate the sensitivity of the dengue surveillance system in detecting hospitalized cases in ten capital cities from 2008 to 2013 in Brazil . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"medicine",
"and",
"health",
"sciences",
"clinical",
"laboratory",
"sciences",
"geographical",
"locations",
"tropical",
"diseases",
"health",
"care",
"age",
"groups",
"inpatients",
"neglected",
"tropical",
"diseases",
"infectious",
"disease",
"control",
"patients",
"pub... | 2016 | Sensitivity of the Dengue Surveillance System in Brazil for Detecting Hospitalized Cases |
Apoptosis is a form of programmed cell death critical for development and homeostasis in multicellular organisms . Apoptosis-like cell death ( ALCD ) has been described in several fungi , including the opportunistic human pathogen Cryptococcus neoformans . In addition , capsular polysaccharides of C . neoformans are known to induce apoptosis in host immune cells , thereby contributing to its virulence . Our goals were to characterize the apoptotic signaling cascade in C . neoformans as well as its unique features compared to the host machinery to exploit the endogenous fungal apoptotic pathways as a novel antifungal strategy in the future . The dissection of apoptotic pathways revealed that apoptosis-inducing factor ( Aif1 ) and metacaspases ( Mca1 and Mca2 ) are independently required for ALCD in C . neoformans . We show that the apoptotic pathways are required for cell fusion and sporulation during mating , indicating that apoptosis may occur during sexual development . Previous studies showed that antifungal drugs induce ALCD in fungi and that C . neoformans adapts to high concentrations of the antifungal fluconazole ( FLC ) by acquisition of aneuploidy , especially duplication of chromosome 1 ( Chr1 ) . Disruption of aif1 , but not the metacaspases , stimulates the emergence of aneuploid subpopulations with Chr1 disomy that are resistant to fluconazole ( FLCR ) in vitro and in vivo . FLCR isolates in the aif1 background are stable in the absence of the drug , while those in the wild-type background readily revert to FLC sensitivity . We propose that apoptosis orchestrated by Aif1 might eliminate aneuploid cells from the population and defects in this pathway contribute to the selection of aneuploid FLCR subpopulations during treatment . Aneuploid clinical isolates with disomies for chromosomes other than Chr1 exhibit reduced AIF1 expression , suggesting that inactivation of Aif1 might be a novel aneuploidy-tolerating mechanism in fungi that facilitates the selection of antifungal drug resistance .
Apoptosis is a form of programmed cell death critical for development and homeostasis in multicellular organisms [1] . The mechanisms of apoptosis are complex and energy-dependent . Numerous pro-and anti-apoptotic signals have to be integrated to activate apoptotic effectors only when necessary . Mitochondria play an important role in apoptosis by releasing several proteins , such as cytochrome c , which triggers the proteolytic maturation of caspases [2] . Caspases are a family of cysteine proteases that are recognized as key components of the apoptotic machinery . Caspases proteolytically cleave specific substrates , leading to the ordered dismantling of intracellular components during apoptotic death [3] . Caspase-independent apoptotic pathways involve the translocation of two nucleases , apoptosis-inducing factor ( AIF ) and endonuclease G ( EndoG ) , from the mitochondria to the nucleus [4] . Features typical of apoptotic cell death , including chromatin fragmentation and condensation and translocation of phosphatidylserine to the outer layer of the cellular membrane , were described for the first time in fungi in a Saccharomyces cerevisiae cdc48 mutant [5] . Heterologous expression of mammalian pro- or anti-apoptotic Bax or Bcl2 in yeast causes or prevents apoptosis-like cell death ( ALCD ) , respectively [6]–[8] . Subsequently , several fungal species have been reported to undergo ALCD as a consequence of interactions with other organisms and during development and aging ( reviewed by [9]–[13] ) . The identification and functional analysis of core components of the apoptotic machinery has revealed partial conservation , along with substantial differences in function and mode of action between fungal and human proteins . Consequently , apoptotic pathways are considered suitable targets for novel antifungal therapies . Cryptococcus neoformans var . neoformans undergoes ALCD when co-cultivated with the bacterium Staphylococcus aureus and also in response to oxidative stress induced by hydrogen peroxide [14] . Conversely , purified capsular polysaccharides from C . neoformans are known to induce apoptosis in host immune cells , including rat splenocytes [15] , [16] , activated human T cells [17]–[19] , and murine macrophages [20]–[22] . Apoptosis induction is a strategy that C . neoformans may deploy to evade host innate immune responses and contributes to the powerful immunosuppression that accompanies cryptococcosis . Because C . neoformans both induces and undergoes apoptosis in response to different signals , its apoptotic machinery must have unique features compared to the apoptotic-signaling cascade of the host , which are potential new therapeutic targets . C . neoformans has emerged as an important pathogen of immunocompromised and immunocompetent patients , estimated to cause 1 million new cryptococcal disease cases globally each year , with up to 600 , 000 fatalities yearly [23] . Cryptococcosis is a systemic mycosis that commonly involves the lungs and central nervous system . Untreated cryptococcal meningitis has a mortality rate of 100% . The recommended therapy for cryptococcal meningitis , which is intravenous amphotericin B ( AMB ) combined with flucytosine [24] , is associated with high toxicity and serious adverse reactions in patients . Moreover , the high costs associated with AMB therapies limit its availability in resource-limited regions such as sub-Saharan Africa . Instead , a less effective agent , fluconazole ( FLC ) , is widely used in these areas , and mortality remains high . FLC is a triazole antifungal drug that inhibits the fungal cytochrome P450 lanosterol 14α-demethylase ( encoded by ERG11 ) , thereby blocking the production of ergosterol , the main sterol of fungal membranes . Here we studied the apoptotic pathways of C . neoformans aiming to identify unique features to specifically induce ALCD of this pathogen as a novel treatment . We found an unexpected connection between ALCD and antifungal drug resistance: the elimination of the apoptosis-inducing factor ( Aif1 ) stabilizes aneuploidy and promotes the generation of FLC-resistance . Selection of FLCR populations with Aif1 defects occurs in vivo , and the inactivation of Aif1-mediated ALCD in C . neoformans may explain cases of FLC treatment failures in the clinical setting . In addition , apoptotic pathways are involved in sexual development and the formation of spores , which serve as infectious propagules of C . neoformans .
ALCD has been described in several fungal species and is induced by different compounds and stress conditions ( reviewed by [9] , [11]–[13] ) . In particular , hydrogen peroxide has been reported to trigger ALCD in C . neoformans var . neoformans [14] , and so we tested if similar conditions would induce cell death in C . neoformans var . grubii , which is responsible for more than 90% of cryptococcal infections worldwide [25] . Treatment of the H99 sequence reference strain with 2 mM hydrogen peroxide for 3 hours at 37oC induced characteristic apoptotic markers , such as DNA fragmentation indicated by TUNEL ( TdT-mediated dUTP nick end labeling ) . TUNEL detects DNA fragmentation by using terminal deoxynucleotidyl transferase ( TdT ) to incorporate dUTP tagged with FITC into the blunt ends of double-stranded DNA breaks , and is the standard method for identification and quantification of apoptotic cells [26] . While TUNEL-FITC positive cells were only rarely observed in untreated controls ( Figure 1A , upper panels ) , ∼45% of wild-type cells showed DNA degradation after treatment with hydrogen peroxide ( Figure 1B ) . We also used Annexin V-FITC to bind to phosphatidylserine ( PS ) , a plasma membrane phospholipid that is externalized to the outer layer during apoptosis . Annexin V-FITC does not bind cells with an intact plasma membrane , but can falsely detect the PS present in the inner membrane of lysed ( necrotic ) cells . Therefore , simultaneous staining with the vital dye propidium iodide ( PI ) allows the discrimination of early apoptotic ( FITC+ , PI− ) and late apoptotic or necrotic yeast cells ( FITC+ , PI+ ) . The percentage of WT cells staining positive with Annexin V-FITC but negative with PI ( apoptotic cells ) rose to ∼42% after treatment with hydrogen peroxide ( Figure S1A , green bars ) . To find potential regulators of ALCD in C . neoformans , functional homologs of the major apoptotic machinery identified from S . cerevisiae were used to search the H99 genome database . The following homologs were identified as first reciprocal BLASTp hits: cytochrome c oxidase subunit 1 ( COX1 , CNAG_09009 ) ; inhibitor of apoptosis protein ( IAP1 , CNAG_04708 ) , and the caspase-independent nucleases apoptosis-inducing factor ( AIF1 , CNAG_04521 ) and endonuclease G ( ENDOG CNAG_02204 ) . Two metacaspases were identified in the H99 genome using the S . cerevisiae Mca1p: MCA1 ( CNAG_04636; 60% identity and 4e-103 ) and MCA2 ( CNAG_06787 , 44% identity and 1e-88 ) . Interestingly , no apparent high-temperature resistance A homolog ( HtrA2 , also called Nma111p in yeast for nuclear mediator of apoptosis ) was found . C . neoformans does not have apparent homologs of the apoptotic protease activating factor ( APAF ) and the poly ( ADP-ribose ) polymerase ( PARP ) , which are present in filamentous fungi but absent in S . cerevisiae . Of note , multiple attempts to disrupt the COX1 gene were unsuccessful ( Toffaletti and Perfect , unpublished observations ) and it may therefore be an essential gene in C . neoformans . We generated two independent aif1 , mca1 mca2 , and aif1 mca1 mca2 null mutants in congenic α and a C . neoformans var . grubii backgrounds to address whether C . neoformans has parallel caspase-dependent and caspase-independent apoptotic pathways . After treatment with 2 mM hydrogen peroxide , the aif1 mutants showed a 41% reduction , and mca1 mca2 showed a 47% reduction in TUNEL-positive cells when compared to wild-type ( Figure 1B ) , while the aif1 mca1 mca2 triple mutant showed a reduction of 79% . In addition , all mutant strains showed reduced externalization of PS revealed by Annexin V assay ( Figure S1A ) and did not show increased vacuolization observed during autophagic cell death ( not shown ) . Finally , the aif1 , mca1 mca2 , and aif1 mca1 mca2 null mutants were more resistant to 2 mM hydrogen peroxide than the wild-type ( Figure S2 ) . Our results indicate that Aif1 and the two metacaspases independently promote ALCD in C . neoformans . Amphotericin B ( AMB ) and fluconazole ( FLC ) are antifungal drugs widely used to treat major fungal diseases . Recent treatment guidelines recommend the more toxic AMB-based regimens for induction therapy instead of the better-tolerated FLC-based regimens to treat cryptococcal meningitis , because the latter is fungistatic and because Cryptococcus strains repeatedly exposed to FLC can acquire direct drug resistance [27] . AMB was shown to induce ALCD in Aspergillus fumigatus [28] and Candida albicans in vitro [29] , and in biofilms of C . albicans , Candida krusei , and Candida parapsilosis [30] . To test if antifungal drugs would induce ALCD in C . neoformans , we first determined if the apoptotic mutants have increased resistance to AMB and FLC . Using Epsilometer test strips ( Etest ) , the aif1 and mca1 mca2 mutants showed no significant difference in sensitivity to AMB ( Table 1 ) and we could not detect ALCD induction by treating wild-type cells with the minimal inhibitory concentration of 0 . 125 µg/ml of AMB ( Figure S1B ) . On the other hand , the aif1 and mca1 mca2 mutants showed increased resistance to FLC ( Table 1 ) . Even though FLC is a fungistatic agent , we tested if this azole was able to induce ALCD in C . neoformans . We could not detect ALCD induction by treating wild-type cells with up to 64 µg/ml of FLC ( Figure S1B and data not shown ) and therefore , the increased resistance of the aif1 and mca1 mca2 mutants is likely not attributable simply to a lack of cell death induction . We also noticed that considerably more FLCR colonies grew within the zone of inhibition around the FLC strip in Etest assays of two independent aif1 deletion mutants , aif1::NAT and aif1::HYG ( Figure 2A ) , a phenomenon called heteroresistance . Heteroresistance to azoles , or the presence of drug-resistant cells within a drug-sensitive population , was previously observed as an intrinsic resistance mechanism in all serotypes of C . neoformans [31]–[33] and C . gattii [34] . During these Etest experiments , resistant colonies within the zone of inhibition were observed up to 10 times more frequently in the aif1 deletion mutants ( Figure 2A ) . Fluctuation analysis using 12 independent colonies estimated that the resistance rate to 32 µg/ml FLC is about 5 times higher in the aif1 mutant compared to H99 ( Table 2 ) . This is not due to a higher mutation rate in the aif1 strain because it showed similar 5-FOAR rates compared to the wild-type stain ( 5 . 1 and 4 . 5×10−8 respectively ) . H99 was recently reported to be able to adapt to FLC concentrations higher than its minimal inhibitory concentration ( MIC ) by becoming aneuploid [35] . Chromosome 1 ( Chr1 ) disomy was a common signature of FLCR strains , and two of the genes resident on this chromosome were shown to be important for FLC resistance: ERG11 ( cytochrome P450 lanosterol 14α-demethylase ) , which is the target of FLC , and AFR1 ( antifungal resistance 1 , [36] ) , which is the major ATP binding cassette ( ABC ) transporter of azoles in C . neoformans [35] . Therefore , we determined the copy number of Chr1 in several FLCR colonies isolated from independent Etest assays ( Table 3 ) . We used a quantitative polymerase chain reaction ( qPCR ) assay and three probes distributed along Chr1: ERG11 , ACT1 , and AFR1 ( Figure S3 ) . A subset of the colonies was karyotyped by array comparative genomic hybridization ( aCGH; Figure 2B-E and Figure S4 ) . DNA content analyses by flow cytometry indicate that all strains described in Table 3 are haploid or near haploid ( n+1 ) ( data not shown ) . In the wild-type H99 background , four out of eight colonies ( 50% ) isolated from Etest halos had MICs greater than 48 µg/ml for FLC and also had Chr1 disomy ( Table 3 , Figure 2B ) . The four colonies that presented MICs less than 48 µg/ml were euploid ( Table 3 ) . Resistant colonies isolated from a second round of Etests presented higher MICs than the original four strains , and in all cases Chr1 was duplicated ( Table 3 , Figure 2C ) . Notably , all aif1 FLCR colonies ( 100% ) isolated directly from Etest experiments became completely resistant to FLC and also had whole or partial Chr1 disomy ( Table 3 ) . From our aCGH analysis , it was clear that some FLCR colonies in the aif1 background had whole or partial chromosome duplications in addition to Chr1 , which were not observed for the H99 wild-type background ( Figure 2D-E , Table 3 , Figure S4 ) . A complemented strain , in which the AIF1 gene with its native promoter and terminator was ectopically integrated into the genome , behaved like the wild-type strain in Etest assays ( Figure 2A ) and FLC fluctuation analysis ( Table 2 ) . Only two out of ten colonies ( 25% ) isolated from Etest halos had MICs greater than 48 µg/ml for FLC and also had Chr1 disomy ( Table 3 ) , while the remaining colonies were euploid . In conclusion , aneuploidy and FLC resistance emerge at higher frequency in the absence of Aif1 . FLC resistance acquired through Chr1 disomy was reported to be lost during passage in drug-free media , and the clones returned to their original euploid state [33] , [35] . To test if FLCR aneuploid isolates would lose their adaptive phenotype upon removal of the selective pressure imposed by the drug , we chose to test two isolates with Chr1 disomy [n+ ( 1 ) ] and FLC resistance in wild-type H99 , the aif1 mutant , and the aif1+AIF1 complemented strain . Upon growth in FLC-free liquid medium for about 20 generations , the isolates in the wild-type ( CPS106 and CPS149 ) and complemented ( CPS178 and CPS179 ) backgrounds returned to the native heteroresistant phenotype ( Table 4 ) . However , the isolates in the aif1 background ( CPS18 and CPS24 ) did not lose the FLC resistance phenotype during nonselective growth ( Table 4 ) . These results indicate that the aneuploid FLCR cells are stabilized and tolerated in the absence of Aif1 . Aneuploid S . cerevisiae strains bearing an extra copy of various chromosomes displayed decreased resistance to stress , such as increased sensitivity to high temperatures and to protein synthesis inhibitors [37] . We tested whether the FLCR aneuploid isolates that we isolated had a similar phenotype . As shown in Figure 3A , none of the tested aneuploid strains ( including isolates of different strain backgrounds and containing various chromosomal amplifications ) showed increased sensitivity to high temperatures or to the protein synthesis inhibitor rapamycin . The only difference that we noticed in the aneuploid strains compared with euploid strains was the presence of elongated and larger cells ( data not shown ) , which was also observed by Sionov et al . [35] . Because there was no difference between aneuploid strains in the AIF1 and aif1 backgrounds , we did not further investigate this phenotypic characteristic . We considered that the increased frequency of FLCR subpopulations in the aif1 mutant could be explained by two possible mechanisms . First , it is possible that Aif1-mediated ALCD would eliminate aneuploid cells from the population , and in the absence of Aif1 , aneuploid cells persist within the population and emerge as FLCR isolates . Another hypothesis is that Aif1 has non-apoptotic functions related to proper chromosomal segregation . To test if aif1 has mitotic spindle checkpoint defects , we examined its sensitivity to benomyl . In S . cerevisiae , hypersensitivity to the microtubule-destabilizing drug benomyl of “budding uninhibited by benzimidazole” ( bub ) mutants correlates with their mitotic checkpoint defect [38] , [39] . As shown in Figure 3B , the aif1 mutant did not exhibit sensitivity to benomyl . To eliminate the possibility that benomyl is simply not taken up by C . neoformans , we deleted the gene CNAG_03184 , which was the first reciprocal BLASTp hit with S . cerevisiae Bub1 , in the H99 background . As expected , the bub1 mutant showed an increased sensitivity to benomyl , indicating that this drug is active in C . neoformans . In conclusion , because aif1 has an intact mitotic spindle checkpoint , we propose that Aif1-mediated ALCD eliminates aneuploid cells from the population . C . neoformans strains with FLC resistance levels greater than 32 µg/ml were previously reported to have higher virulence in mice [33] . To test if that would also be the case for isolates in the aif1 mutant background , we tested the virulence of two strains completely resistant to FLC using an inhalation murine model of cryptococcosis . Mice were intranasally infected with the indicated strains , and their survival was monitored and plotted against time . The aif1 mutant has a virulence level similar to wild-type ( see below ) . As shown in Figure 3C , FLCR strains CPS18 [FLCR256 n+ ( 1 ) ] and CPS24 [FLCR256 n+dup ( 1 ) ( 3 ) ] showed similar virulence to the H99 strain ( untreated cohort of animals; p-values of 0 . 5739 and 0 . 5252 , respectively ) . Next , we tested the efficacy of antifungal therapy when infection is caused by FLCR strains . Treatment with 20 mg/kg/day of FLC was initiated 24 hours after infection and was continued for 14 days , as described by [33] . As shown in Figure S5 , treatment with 20 mg/kg/day of FLC for 14 days modestly prolonged the survival of mice inoculated with wild-type H99 ( p-value of 0 . 0043 ) and aif1 mutant ( p-value of 0 . 0023 ) strains , but it did not prolong the survival of cohorts inoculated with the CPS18 ( p-value of 0 . 1443 ) and CPS24 ( p-value of 0 . 2237 ) resistant isolates . In a second set of experiments , we evaluated the efficacy of a higher FLC dosage , and animals received 100 mg/kg/day of FLC , without treatment interruption . All mice inoculated with wild-type H99 and aif1 mutant strains and treated with 100 mg/kg/day of FLC survived for 60 days ( p-values <0 . 0001 ) . On the other hand , while treatment with 100 mg/kg/day of FLC prolonged the survival of mice inoculated with the CPS18 and CPS24 resistant isolates ( p-values <0 . 0001 ) , 100% mortality was observed in these cohorts ( Figure 3C ) . Additionally , we examined if resistant strains would appear during FLC treatment and if aneuploidy would be stable with in vivo passage . We randomly selected three mice treated with 100 mg/kg/day of FLC for each strain . For CPS18 and CPS24 cohorts , the mice were sampled post-mortem between 30 and 35 days post-infection . For H99 and aif1 cohorts , we used asymptomatic mice sacrificed on day 60 , when the experiment was terminated . Colonies recovered on YPD medium from lungs and brains were tested for resistance to FLC . About 8% of wild-type and 29% of aif1 colonies developed FLCR in vivo during treatment , while more than 96% of CPS18 and CPS24 colonies retained their initial FLC resistance acquired in vitro ( Figure 4A ) . The aCGH analysis of a wild-type H99 FLCR colony ( CPS129 ) selected randomly showed that this isolate acquired Chr1 disomy ( Figure 4B ) . Ploidy analysis by FACS indicated that CPS129 was diploid [2n+ ( 1 ) , data not shown] . Increased ploidy has recently been described in C . neoformans to be associated with the formation of giant yeast cells in infected animals [40] , [41] . The aCGH analysis of an aif1 FLCR colony ( CPS134 ) showed duplication of chromosomes 1 , 4 , and 6 ( Figure 4C ) , while FLCR colonies in the CPS18 ( CPS145 ) and CPS24 ( CPS146 ) backgrounds maintained their initial state of aneuploidy ( Figure 4D and 4E ) . These in vivo results can be correlated to our in vitro experiments ( Table 2 and Table 4 ) , but it is clear that the interaction of C . neoformans with the murine host environment promotes genomic plasticity , which serves to generate new phenotypic adaptations in this pathogen with disease management consequences . To evaluate and validate if defects in the apoptotic pathway of C . neoformans have any impact on azole resistance in the human clinical setting , we tested the susceptibility to FLC of seven clinical isolates taken directly from frozen cerebrospinal fluid ( CSF ) of individual patients with cryptococcal meningitis . Microevolution and phenotypic variability of C . neoformans isolates occurs in the laboratory through multiple in vitro passages ( reviewed by [42]–[44] ) . Therefore , an important factor considered in our experimental design when testing clinical isolates was the use of colonies isolated directly from CSF . We also targeted initial clinical isolates , collected before the commencement of antifungal therapy . It is important to note that the RCT isolates originated from South Africa , an HIV endemic area , and FLC is sometimes used as a prophylactic antifungal treatment in HIV/AIDS patients , and thus we cannot completely rule out prior exposure to azoles in these isolates . As shown in Table 1 , the susceptibility to FLC of the seven direct clinical isolates varied , with MICs ranging from 4 to >256 µg/ml . Specifically , isolate RCT 17 showed FLCR colonies within the zone of inhibition around the FLC strip in Etest assays , similar to the aif1 mutants ( Figure 5A ) . Moreover , such colonies were completely resistant to FLC and exhibit Chr1 disomy ( Figure 5B , and Table 3 ) . RCT 17 showed a FLC resistance rate that was about 20 times higher than H99 and 4 times higher than the aif1 strain ( Table 2 ) . The FLCR isolates CPS104 and CPS157 in the RCT 17 background did not revert to FLC sensitivity during growth on FLC-free medium for 20 generations ( Table 4 ) , showing that aneuploid FLCR cells are stable in the RCT17 background , as was observed for the aif1 mutant . Given the similarities between the clinical isolate RCT 17 and the aif1 mutant , we hypothesized that the AIF1 gene might be defective in this clinical isolate . Quantitative RT-PCR ( qRT-PCR ) revealed that AIF1 expression is approximately 80% reduced in the RCT 17 isolate compared to the H99 strain ( Figure 5C ) . DNA sequence analysis determined the presence of five nucleotide polymorphisms in the AIF1 promoter and coding sequence from RCT 17 compared to the H99 ( data not shown ) . Of note , the primer pair used for the qRT-PCR was designed to encompass the boundaries of exons 1 and 2 , a region that has no polymorphisms in the RCT 17 isolate . Complementation of the RCT 17 strain with the AIF1 gene and its native promoter and terminator restored normal AIF1 expression levels ( Figure 5C ) , reduced the FLC heteroresistance rate ( Figure S6A and Table 2 ) , and reversed the stability of the aneuploid Chr1 in the absence of FLC ( Table 4 ) , indicating that low AIF1 expression levels are responsible for the RCT 17 aif1-like phenotypes . Additionally , we evaluated AIF1 expression in two other clinical isolates from the USA that were reported to be disomic for other chromosomes besides Chr1 . Isolates HC-4 and HC-6 show partial duplication for chromosomes 9 and 6 , respectively ( Figure 5D and 5E and Hu and Kronstad et al . , manuscript in preparation ) . AIF1 expression is also reduced in isolates HC-4 and HC-6 ( Figure 5C ) , indicating that downregulation of AIF1 is not restricted to cryptococcal cells under strong selective drug pressure and might be a novel aneuploidy-tolerating mechanism . During the genetic manipulations of our apoptotic mutants , we noticed that they displayed defective mating phenotypes . Mating in C . neoformans is initiated by fusion between α and a cells to produce dikaryotic filaments , and the process culminates in the formation of basidia decorated with four spore chains . Spores formed by mating represent infectious propagules for C . neoformans . Mating filament production was reduced but not abolished when aif1 , mca1 mca2 , and aif1 mca1 mca2 independent mutants of opposite mating types were subjected to bilateral mutant crosses ( Figure S7A ) . In unilateral crosses ( mutant × wild-type ) , the aif1 , mca1 mca2 , and aif1 mca1 mca2 mutants had delayed formation of filaments compared with a wild-type cross , but were still fertile ( data not shown ) . Bilateral crosses of the aif1 mca1 mca2 triple mutant showed more severe defects in filamentation , indicating that both Aif1 and the metacaspases have a role in C . neoformans mating . In confrontation assays between the aif1 , mca1 mca2 , and aif1 mca1 mca2 mutants and a crg1 mutant strain , which is hypersensitive to mating pheromone and produces abundant conjugation tubes when confronted with wild-type cells of the opposite mating type [45] , no conjugation tubes were produced , indicating a decrease in pheromone secretion by the aif1 , mca1 mca2 and aif1 mca1 mca2 mutants ( Figure S7B ) . Decreased hyphal growth during mating can be a result of defective fusion between α and a cells . To determine if this was the case in the aif1 and metacaspase mutants , cell fusion assays using opposite mating type strains with dominant genetic markers in bilateral crosses were conducted . The formation of double-resistant diploids by cell fusion was quantified after 48 h of incubation on mating medium . Many diploid isolates were produced from the wild-type control cross compared with the bilateral mutant crosses ( Figure S8 ) . We also examined fusion in unilateral crosses between the mca1 mca2 mutant and wild-type strains . The number of diploid isolates produced from these crosses was considerably lower than those with wild-type strains , demonstrating that one copy of each metacaspase gene is not sufficient for fusion ( Figure S8 ) . Hence , Aif1 and the metacaspases are required for the cell fusion step of mating in C . neoformans . Additionally , a significant reduction in basidiospore production was observed in bilateral mutant crosses ( Figure 6A ) . The aif1 and mca1 mca2 mutants produce defective basidia with aberrant spore chains ( Figure 6A-D ) , although these strains were not sterile , and we were able to isolate recombinant spores in which the mutations were recovered in the opposite mating type . Sporulation in the aif1 mca1 mca2 bilateral mutant cross was impaired and most basidia produced few spores , in contrast to the four long spore chains observed in a wild-type cross ( Figure 6 ) . Basidiospore production defects were also observed in unilateral crosses between aif1 mca1 mca2 mutant and wild-type strains ( data not shown ) . We conclude that ALCD is required for the proper sporulation of C . neoformans . Apoptosis plays an essential role in maintaining homeostasis in multicellular organisms by eliminating permanently damaged cells . In the unicellular yeast S . cerevisiae , ALCD has been proposed to eliminate the less fit individuals from a monoclonal population of cells , thereby improving the chance of survival of the fitter clones in the genetic pool [46] , [47] . To determine whether the apoptotic pathways are relevant to the ability of C . neoformans populations to compete for resources in a common environment , the fitness of apoptotic null mutants was measured in pair-wise growth competition assays with wild-type . The same amount of cells from wild-type and apoptotic null mutant strains grown to exponential phase were mixed in liquid YPD medium and incubated at 30°C with agitation . Each strain was identified based on differing genetic markers , including a control sample using two wild-type strains . To reduce cell fusion and mating during the experiment , all strains used were of the α mating type . Aliquots of each culture were taken at the indicated times and plated on selective media to calculate the survival ratios . As shown in Figure 7A , the apoptotic defective mutants were outcompeted by the wild-type cells . Because the aif1 , mca1 mca2 , and aif1 mca1 mca2 mutants had similar division rates as wild-type in liquid YPD medium at 30°C during solo culture ( Figure S9 ) , we eliminated the possibility that the wild-type cells were simply more fit to begin with , and thus concluded that wild-type cells exhibit enhanced fitness during co-culture with apoptotic mutant cells . To determine whether the apoptotic null mutants had any growth defect in the host environment , we performed virulence assays in a murine inhalation model of cryptococcosis . Mice were intranasally infected with wild-type and mutant strains , and their survival was monitored and plotted against time ( Figure 7B ) . Compared to the wild-type strain , the apoptotic defective mutants showed no significant difference in virulence ( Figure 7B ) , fungal burden , or organ dissemination pattern ( data not shown ) . Therefore , inactivation of ALCD had no apparent impact on the virulence of C . neoformans in these tested conditions .
Alteration in gene copy number is a major mechanism for environmental adaptation of asexual yeast populations . In S . cerevisiae , aneuploidy was shown to confer improved growth and fitness advantages under stressful conditions [37] , [56] , [57] . Aneuploidy was also associated with FLC resistance in Candida species [58]–[61] . In C . albicans , a specific segmental aneuploidy , isochromosome 5L [i ( 5L ) ] , is commonly found in FLCR isolates , and the loss of i ( 5L ) is correlated with reduced FLC resistance [58] . Gain of i ( 5L ) , which is comprised of two identical chromosome 5 left arms flanking a centromere , amplifies two genes that contribute additively and independently to FLC resistance: ERG11 ( the target of FLC ) and TAC1 ( a transcription factor that activates expression of the drug efflux pumps CDR1 and CDR2 [59] ) . The acquisition of aneuploidy was recently reported in strains of C . neoformans that display heteroresistance to FLC [35] . Heteroresistance in C . neoformans is defined as heterogeneous FLC susceptibility within a population with resistant subpopulations being able to adapt to higher concentrations of the drug in a stepwise and reversible manner [31]–[33] . Heteroresistance is considered different from trailing phenomena or incomplete growth inhibition by azoles observed in Candida species . Trailing growth can cause the MICs of azoles for some isolates to be low ( susceptible ) after 24 h of growth but much higher ( resistant ) after 48 h [62] . The relevance of such discordant interpretive categories at the two time points is as yet unclear , but current evidence suggests that the lower MIC correlates most closely with the outcome in vivo [62] . Bacterial populations were reported to produce persister cells that neither grow nor die in the presence of microbicidal antibiotics . Persisters are largely responsible for high levels of biofilm tolerance to antimicrobials , but do not exhibit an increased MIC . C . albicans persister cells that survived killing by AMB were detected only in biofilms and not in exponentially growing or stationary-phase planktonic populations [63] . Even though we used Etest assays to isolate FLCR colonies in this study , our results support previous findings in which the H99 strain exposed to FLC in liquid cultures became resistant in a step-wise pattern to increasing concentrations of the drug after acquisition of specific chromosomal duplications [33] , [35] . In our experiments , Chr1 disomy was observed in H99 derivatives that became resistant to more than 48 µg/ml of FLC and was more common on second Etest passages , while disruption of aif1 stimulated the selection of completely FLCR ( MIC >256 µg/ml ) subpopulations . Amplification of chromosomes other than Chr1 , particularly Chr4 , was observed in aif1 derivatives , but not in the H99 or AIF1 complemented derivatives isolated in Etest experiments ( Table 3 ) . In the aif1 derivatives , segmental chromosomal duplications were also observed; isolate CPS24 had the left arm of Chr1 duplicated , which harbors ERG11 , and the end of the right arm of chromosome 3 ( Figure 2E and Figure S3A ) . C . gattii isolates resistant to 64 µg/ml of FLC showed different patterns of chromosomal amplifications , including disomy and segmental duplication of several chromosomes besides Chr1 [64] . Clinical isolate RCT 52 had a FLC MIC >256 µg/ml ( Table 1 ) but did not show any chromosomal amplification ( data not shown ) , indicating that not all FLCR isolates are aneuploid . These findings emphasize the existence of FLC resistance mechanisms that are independent of Chr1 disomy . After about 20 generations of non-selective growth , more than 90% of FLCR colonies in the H99 background and 75% in the complemented strain returned to being FLC sensitive , while less than 10% in the aif1 background lost resistance to FLC upon removal of the drug pressure ( Table 4 ) . The lack of benomyl sensitivity in the aif1 mutant excludes the possibility that the observed increased aneuploidy in this strain is a result of a defective mitotic spindle checkpoint . Our results show that the lack of a functional Aif1 was able to stabilize the aneuploid state selected by FLC treatment , and we propose that aif1 is an aneuploidy-tolerating mutation in C . neoformans . Torres et al . [56] described the existence of aneuploidy-tolerating mutations in yeast , such as the deubiquitinating enzyme Ubp6 , which improves growth rates in aneuploid yeast strains by attenuating the protein stoichiometry imbalances caused by chromosomal amplifications . We did not observe that FLCR aneuploid strains display an increased sensitivity to stresses , such as reported in S . cerevisiae by Torres et al . [37] . The results shown in Figure 3A suggest that aneuploidy does not inevitably result in decreased resistance to stress , and are in agreement with a recent study by Pavelka et al . [57] , which showed that aneuploidy directly confers phenotypic variation that can result in a growth advantage under stress conditions that are suboptimal for euploid cells . The conflicting results in S . cerevisiae were explained by the different methods used to generate the aneuploid strains; while Torres et al . [37] constructed their aneuploid strains through selection with a combination of drug and nutrient markers , Pavelka et al . [57] generated aneuploid strains through random meiotic segregation , suggesting an impact of both karyotype and continuous selection with drugs and nutrient markers in the phenotypic variation conferred by aneuploidy [57] . Interestingly , as in Pavelka et al . [57] , we observed that certain aneuploid strains were able to grow better in the presence of drugs such FLC and rapamycin ( strain WT FLCR256 , Figure 3A ) . Rancati et al . [65] reported that S . cerevisiae cells lacking MYO1 , which encodes the only myosin-II normally required for cytokinesis , rapidly acquired aneuploidy that led to specific changes in the transcriptome that restored growth and cytokinesis . Therefore , one possibility is that aif1 mutants have more stable Chr1 aneuploidy because genes present on Chr1L might complement Aif1 functions . It should be noted that no growth defect ( besides the decreased competitiveness with wild-type ) was observed for aif1 mutant strains , and that Chr1 aneuploidy did not appear to improve aif1 mutant strains growth . We speculate that apoptosis orchestrated by Aif1 eliminates aneuploid cells from the population . Why didn't the metacaspase mutants have an increased rate of aneuploidy and consequent resistance to FLC ? Based on the fact that Aif1 and the metacaspases are independently required for apoptosis in C . neoformans , it is possible to speculate that the elimination of aneuploid cells from the population is regulated by a caspase-independent apoptotic pathway . In yeast , the Yca1 caspase-like protease participates in approximately 40% of the investigated cell death scenarios , while Aif1 and endonuclase G execute caspase-independent cell death [66] . It is also possible that ploidy maintenance is a non-death function of Aif1 . FLCR colonies isolated from the H99 and aif1 strains during our in vivo experiments showed more complex patterns of ploidy increase and amplifications ( Figure 4B and 4C ) . The interaction of C . neoformans with mammalian hosts has been shown to promote genomic plasticity in this pathogen ( reviewed by [67] ) . Comparative hybridization studies characterized disomy for chromosome 13 in two independent clinical isolates [68] . Additionally , the formation of giant cells with diameters up to 100 µm was observed during murine cryptococcal infection [40] , [41] . These giant yeast cells were found mainly in lung tissue and were polyploid and uninucleate , suggesting that they arise from DNA endoreplication without concomitant cell division . It will be interesting to investigate if ALCD is also important in ploidy increase and the formation of giant yeast cells . In preliminary analysis of FLC Etests using a diploid wild-type strain , we did not observe an increase in heteroresistance while a diploid aif1 strain still showed increased heteroresistance ( Figure S6B ) . In our in vivo virulence experiments , we evaluated the efficacy of FLC therapy for wild-type and FLCR strains using a murine model of cryptococcal meningitis . Treatment with 100 mg/kg/day of FLC had antifungal activity against H99 and aif1 strains ( but did not clear infection in the brains of the animals ) and doubled the survival time of mice infected with FLCR strains . Pharmacodynamic studies , using the area under the concentration-time curve ( AUC ) , showed that FLC therapy in doses of 100 mg/kg/day given to mice mimics doses of 400 mg/day given to humans [69] . Therefore , the use of higher dosages of FLC might be beneficial in patients , and may help to reduce morbidity and mortality due to cryptococcal meningitis in Africa and other resource-limited regions . We found that RCT 17 , one of seven direct clinical isolates , had increased heteroresistance to FLC , similar to the heteroresistance observed in the aif1 mutants ( Table 1 , Figure 5A ) . Furthermore , FLCR colonies isolated from the RCT 17 background also became completely resistant to FLC through the acquisition of aneuploidy ( Figure 5B , and Table 3 ) and did not revert to being FLC-sensitive during growth on FLC-free medium for 20 generations ( Table 4 ) . AIF1 expression was shown to be downregulated in RCT 17 by qRT-PCR ( Figure 5C ) , and complementation with AIF1 with its native promoter and terminator recovered normal expression levels and reverted the increased heteroresistance to FLC ( Table 2 and Table 3 ) . AIF1 expression was also reduced in the HC-4 and HC-6 clinical isolates ( Figure 5C ) , which show disomies for chromosomes other than Chr1 and are not resistant to FLC ( Table 1 , Figure 5D and 5E ) . Therefore , we propose that the inactivation of AIF1 is a novel aneuploidy-tolerating mechanism in fungi that might not be restricted to cells under strong selective FLC drug pressure . Future investigations will test the mechanisms that result in AIF1 downregulation . Our results suggest that triggering fungal ALCD should be further investigated as a new antifungal approach . Potential antifungal targets include those that can help another antifungal agent by blocking resistance mechanism ( s ) in vivo used by the microorganisms . This principle has previously been demonstrated by beta-lactamase inhibitors in antibacterial combination drugs [70] but has yet to be applied in therapies including new antifungal agents . An AIF1-inducing agent could have a profound effect on the ability of C . neoformans to become resistant in vivo to FLC , and the addition of such inducer to the antifungal regimen could make FLC fungicidal in the host . In conclusion , our results bring to light the importance of the Aif1 in maintaining the euploidy of yeast populations , possibly through the elimination of aneuploid cells via apoptosis . Inactivation of AIF1 might increase genomic plasticity of C . neoformans during infection and promote the generation of phenotypic adaptations such as new virulence traits and antifungal drug resistance .
This study was carried out in strict accordance with The National Research Council's Guide for the Care and Use of Laboratory Animals , Public Health Service Policy on Humane Care and Use of Laboratory Animals , and AAALAC accreditation guidelines . The protocol was approved by the Duke University and Duke University Medical Center Institutional Animal Care and Use Committee ( protocol number: A266–08–10 ) . All efforts were made to minimize suffering . The clinical isolates were collected and stored in the Duke Infectious Disease Repository under an IRB-approved protocol entitled “Database and specimen repository for infectious diseases-related studies” ( #CR3_Pro00005314 ) . The specimens were de-identified discarded samples collected as part of routine clinical practice and were exempt from patient written informed consent . C . neoformans strains used in this study are listed in Table S1 . YL99a strain was obtained by an extra round of H99 and KN99a backcrossing . The tested clinical isolates were taken directly from the CSF of patients with cryptococcal meningitis from USA ( HC isolates , Duke patients ) and from South Africa ( RCT isolates , obtained from Tihana Bicanic and Tom Harrison ) . The indicated mating types were determined by crosses and confirmed with PCR , and the serotyping of clinical isolates was determined by PCR using STE20 mating-type- and serotype-specific primer pairs [71] . Strains were maintained in −80°C glycerol stocks and grown on YPD medium ( 1% yeast extract , 2% Bacto Peptone , and 2% dextrose ) . To perform mating assays , cells of opposite mating type were mixed in water , spotted on 5% V8 juice agar medium ( pH 5 . 0 ) or Murashige and Skoog ( MS ) medium minus sucrose and incubated at room temperature in the dark [72]–[75] . For spot assays , cultures grown in liquid YPD were diluted to 2×106 cells/ml and serially diluted tenfold . Then , 5 µl of each culture was spotted onto YPD or YPD plus 2 mM H2O2 ( Sigma ) or YPD plus 20 ng/ml rapamycin ( LC Laboratories ) plates and incubated at the indicated temperatures . For benomyl sensitivity assays , 5 µl of fivefold serial dilutions were spotted onto YPD plates with DMSO ( control ) or with 10 or 20 µg/ml benomyl ( Sigma ) . Growth differences were imaged following incubation of the plates for 72 h . Strains with gene disruptions were generated using an overlap PCR approach and biolistic transformation in serotype A congenic strains H99 and KN99a as previously described [76] . For gene deletions , the 5' and 3' flanking regions of apoptotic genes and BUB1 were amplified from H99 genomic DNA and the dominant selectable markers NAT , NEO , or HYG were amplified with the universal M13F and M13R primers from plasmids pJAF13 , pJAF12 and pJAF15 respectively . A second overlap PCR amplified a full-length deletion cassette containing the three previous fragments . The products of overlap PCR were purified , precipitated onto gold microcarrier beads ( 0 . 6 mm , Bio-Rad ) , and the H99 strain was biolistically transformed [5] , [26] . Stable transformants were selected for resistance to nourseothricin ( 100 µg/ml ) , G418 ( 100 µg/ml ) , or hygromycin B ( 300 µg/ml ) , screened by PCR , and confirmed by Southern blot . The α mca1 mutant background was used for disruption of MCA2 by biolistic transformation . The resulting α mca1 mca2 mutant was crossed to strain KN99a to generate single and double mutants with opposite mating types ( see Table S1 ) . The aif1 mca1 mca2 triple mutants are progeny obtained from the cross of α aif1with a mca1 mca2 ( see Table S1 ) . For complementation , an overlap PCR product with the NEO marker and the wild-type AIF1 gene containing its native promoter and terminator from strain H99 was generated . The PCR product was biolistically transformed into aif1 mutant and RCT17 backgrounds . Transformants were selected for resistance to G418 ( 100 µg/ml ) and confirmed by PCR . The sequences of primers used are listed in Table S2 . TdT-mediated dUTP nick end labeling ( TUNEL ) assays were performed as described previously [5] , [26] with the following modifications . Log-phase cultures were diluted to an OD600 of 0 . 1 and treated with 2 mM H2O2 ( Sigma ) for 3 h at 37°C . Cells were fixed for 1 h at room temperature with 3 . 7% formaldehyde ( Sigma ) and washed with PBS . After treatment with 5% β-mercaptoethanol in SPM ( 1 . 2 M sorbitol; 50 mM K-phosphate , pH 7 . 3; and 1 mM MgCl2 ) for 1 h at 37°C with gentle agitation , cells were washed with SPM and then digested with 10 µl of zymolyaseTM ( Zymo Research , 4 U/µL ) , 100 mg of lysing enzymes from Trichoderma harzianum ( Sigma ) , and 0 . 1% bovine serum albumin ( Sigma ) in 1 ml of spheroplasting buffer ( 1 M sorbitol; 10 mM EDTA; and 100 mM sodium citrate , pH 5 . 8 ) for 40 min at 37°C with gentle agitation . Spheroplasts were gently harvested at 1600 rpm for 5 min at 4°C and washed three times with SPM to remove the enzymes . Samples were permeabilized with 0 . 5 ml of fresh prepared 0 . 1% Triton X-100 , 0 . 1% sodium citrate solution in cold SPM for 2 min in ice , washed twice with SPM , and incubated with 50 µl TUNEL reaction mixture ( In Situ Cell Death Detection Kit , Roche ) for 60 min at 37°C in dark and humid conditions . Cells were washed with PBS and immediately analyzed by epifluorescence microscopy ( see next section ) or flow cytometry . TUNEL positive cells were quantified by collecting 10 , 000 events on the FL1 ( FITC ) channel of a FACSCalibur flow cytometer ( Becton Dickinson ) using CellQuest software ( Becton Dickinson ) . Cell debris were excluded using the side and forward scatter dot-plot and a negative control ( fixed and permeabilized wild-type cells incubated in 50 µl of TUNEL solution without terminal transferase ) and a positive control [fixed and permeabilized wild-type cells treated with 0 . 1 µg/µl DNase I ( Sigma ) for 10 min at room temperature to induce DNA strand breaks prior to TUNEL labeling] were used to determine the percentage of TUNEL positive cells . Exposed phosphatidylserine was detected by reaction with FITC-coupled Annexin V ( Apoptosis Detection kit , Oncogene Research Products ) . Log-phase cultures diluted to an OD600 of 0 . 1 were treated with 2 mM H2O2 ( Sigma ) for 3 h at 37°C . Yeast cells were washed with PBS , treated with 5% β-mercaptoethanol , and spheroplasted as described above for the TUNEL assay . Spheroplasts were washed two times with SPM and once with Binding Buffer diluted to 1x with 2 M sorbitol , and double stained with FITC-Annexin V and propidium iodine ( PI ) for 20 min at room temperature in the dark dark as described by [77] . Annexin V-positive , PI-negative cells were quantified by collecting 10 , 000 events on the FL1 ( FITC ) and FL2 ( PE ) channels of a FACSCalibur flow cytometer ( Becton Dickinson ) using CellQuest software ( Becton Dickinson ) . To evaluate TUNEL assays , cells were mounted in ProLong Gold antifade reagent with DAPI ( Molecular Probes ) . Brightfield , differential interference microscopy ( DIC ) , and fluorescence images were captured with a Zeiss Axio Imager widefield fluorescence microscope ( Carl Zeiss ) equipped with an Orca cooled-CCD camera ( Hamamatsu ) using DAPI and FITC channels . Images were interfaced with MetaMorph software ( Universal Imaging ) and processed with PhotoShop ( Adobe ) . Mating structures were captured with an Eclipse E400 microscope ( Nikon ) equipped with a DXM1200F digital camera ( Nikon ) . Images were interfaced with ACT-1 software ( Nikon ) and processed with PhotoShop ( Adobe ) . For scanning electron microscopic analysis , mating plates were processed at the North Carolina State University Center for Electron Microscopy , Raleigh , NC , USA . Areas of interest were excised intact from the agar and fixed with 0 . 1 M sodium cacodylate ( pH 6 . 8 ) buffer containing 3% glutaraldehyde for several days at 4°C . Samples were then washed for 1 h in cold sodium cacodylate buffer three times , dehydrated through a graded series of cold 30% and 50% ethanol for 1 h each , and held overnight in 70% ethanol . Dehydration was completed with 1-h incubations with cold 95% and 100% ethanol at 4°C , warming to room temperature in the 100% ethanol , followed by two additional 1-h washes with 100% ethanol at room temperature . The dehydrated samples were critical-point dried with liquid CO2 in a SAMDRI-795 ( Tousimis ) for 15 min and mounted on stubs with silver paint to ensure good adhesion and conductivity . Finally , the samples were coated with 50 Å of gold/palladium using a Hummer 6 . 2 sputter coater ( Anatech ) and held in a vacuum desiccator until viewed at 15 kV on a JSM 5900LV scanning electron microscope ( JEOL ) , photographed with a Digital Scan Generator ( JEOL ) image acquisition system , and processed with PhotoShop ( Adobe ) . In vitro antifungal susceptibility to FLC and AMB was determined by Etest ( AB Biodisk ) according to the manufacturer's instructions . Cells grown to mid-log phase were washed in 0 . 85% NaCl and diluted to an OD600 of 0 . 5 ( ∼1 McFarland turbidity ) . Inocula were applied with cotton swabs in RPMI-1640 agar ( Sigma ) supplemented with 2% glucose and buffered to pH 7 . 0 with MOPS and allowed to dry completely before applying the Etest strip . Plates were incubated at 35°C for 72 h and the MIC was determined as the first growth inhibition ellipse . Distinct colonies within the inhibition halo around the FLC strip during Etest assays were replica plated onto YPD agar and YPD agar plus 32 µg/ml FLC and incubated at 30°C . Isolates able to grow on the FLC plate after 2 days were inoculated into 5 ml YPD liquid and grown overnight at 30°C with agitation . Liquid cultures were tested by Etest assays and used to make gDNA ( later used for ploidy analysis ) and −80°C glycerol stocks . For virulence experiments and FLC resistance stability assays , −80°C glycerol stocks were inoculated into 5 ml YPD liquid and grown overnight at 30°C with agitation . Genomic DNA was extracted using alkyltrimethyl ammonium bromide buffer ( CTAB ) as described by [78] . The quality of the purified DNA was examined on an agarose gel and quantified in a Qubit fluorometer ( Invitrogen ) using the Quant-iT dsDNA BR kit according to Invitrogen's instructions . Then 2 . 5 µg of DNA was sonicated to generate ∼500-bp fragments , which were labeled with Perkin Elmer Cy3-dUTP ( reference DNA from H99 wild-type strain ) or Cy5-dUTP ( strain of interest ) using the Random Primer Reaction of BioPrime Array CGH Genomic Labeling System ( Invitrogen ) . Labeled reference and sample DNA were combined in a 1∶1 ratio , purified using Microcon 30K Centrifugal Filter Units ( Millipore ) , and concentrated in a Speed Vac Plus SC110 ( Savant Instruments ) . Samples were resuspended in 50 µl of Long Oligo Hybridization Solution ( Corning ) , denatured at 95°C for 5 min , and applied to the C . neoformans whole genome 70-mer oligonucleotide spotted array version II ( Washington University Genome Sequencing Center ) covered with a LifterSlip Microarray Coverslip ( Erie Scientific Company ) . Arrays were hybridized at 42°C for 16 h and subsequently washed twice with 1X SSC , 0 . 3% SDS heated to 42°C , followed by two washes in 0 . 2X SSC , and two final washes with 0 . 05X SSC at room temperature for two minutes each . Slides were dried by centrifugation , scanned with a GenePix 4000B scanner ( Axon Instruments ) , and analyzed using GenePix Pro 6 . 0 software ( Molecular Devices ) . GeneSpring GX 7 . 3 ( Agilent Technologies ) was used to perform Lowess normalization of the data , and an S-plus script was used to remove exon-junction probes and to log2 transform the red/green ratios . The transformed data from experimental arrays and from four control hybridizations ( Cy3-reference strain DNA x Cy5-reference strain DNA ) were analyzed with CGH-Miner software [79] , which uses the Clusters Along Chromosomes method to identify DNA copy number alterations . Raw data for the regions of gain/loss identified by the algorithm were then manually inspected to confirm the calls or eliminate occasional errors induced by noise . The plots of log2 data as a function of chromosomal nucleotide position were created using a sliding-window of 10 consecutive probes and a targeted false discovery rate ( FDR , i . e . , the expected proportion of false positive results ) of 0 . 01 . Vertical bars were plotted in red for amplifications and green for deletions that are statistically significant . Vertical bars in gray indicate no DNA copy number alteration or amplifications or deletions that did not meet the statistical settings . FLC resistance rates were calculated by fluctuation analysis using the method of the median [80] . To measure FLC resistance rates , individual colonies of each strain were grown non-selectively in YPD agar from −80°C glycerol stocks . Twelve independent cultures from isolated colonies were grown in YPD liquid medium overnight at 30°C . Yeast cells were collected by centrifugation , washed with water , and serial dilutions were plated on YPD agar and YPD agar with 32 µg/ml FLC to determine the number of viable cells and the number of FLCR cells , respectively . The median number of FLCR cells in the different cultures was corrected for dilution factors , fractions plated , and number of viable cells . The numbers of FLCR cells were ranked , and the 3rd and 10th ranks were used to calculate the lower and upper limits of the 95% confidence interval [81] . Strains resistant to FLC were used to calculate the reversion rates after approximately 20 generations . Non-selectively grown cultures from -80°C glycerol stocks were incubated at 30°C for 24 h in YPD liquid . Culture dilutions were plated on YPD agar , and 100 individual colonies were replica plated onto YPD agar and YPD agar plus 32 µg/ml FLC , followed by incubation at 30°C . Colonies unable to grow on the FLC plate after 2 days were considered revertants . Virulence assays were conducted using a murine inhalation model of cryptococcosis . Cohorts of 4- to 6-week-old female A/JCr mice were anesthetized by intraperitoneal injection of Nembutal ( 37 . 5 mg/kg ) and infected intranasally with 5×104 cells in 50 µl of PBS pipetted slowly into the nares . Inoculum concentration was confirmed by plating serial dilutions in YPD agar and counting colony forming units ( CFU ) after 2 days . Mice were monitored twice daily , and those that showed signs of severe morbidity ( weight loss , extension of the cerebral portion of the cranium , abnormal gait , paralysis , seizures , convulsions , or coma ) were sacrificed by CO2 inhalation . The survival rates of animals were plotted against time , and p-values were calculated with the Mann-Whitney U test . In the first experiment using FLC treatment ( Figure S5 ) , 10 mice per strain were treated with 20 mg/kg/day FLC intraperitoneally , while control groups of 10 mice received only saline . Treatment started 24 h after inoculation of yeasts and was continued for 14 days . In the second experiment using FLC treatment ( Figure 3C ) , 10 mice of about 20 g each were treated with 100 mg/kg/day FLC in the sole source of drinking water as described by [35] . Water intake was recorded daily for each cage ( all 5 mice were assumed to drink the same volume ) and averaged 0 . 23 ml/g . Treatment started 24 h after infection and was continued through day 60 with water being replaced every 2 to 3 days . FLC concentrations for each cage were recalculated based on weights of the animals and the water intake measured during the preceding days . Three euthanized mice in each treated group were dissected on the days indicated post-infection; their brains and lungs were removed , weighed , and homogenized in 1 ml sterile PBS . Serial dilutions of the organ samples were plated on Sabouraud-dextrose agar plus 100 µg/ml chloramphenicol and incubated at 30°C . Up to 50 individual colonies were replica plated onto YPD agar and YPD agar plus 32 µg/ml FLC and incubated at 30°C . Colonies unable to grow on the FLC plate after 2 days were considered revertants . Isolates able to grow on FLC were further tested by aCGH , qPCR , and Etest assays as described above . Yeast cells for RNA extraction were grown in YPD liquid culture overnight and were harvested the following day and washed . RNA preparations were isolated with the RNAeasy Mini Kit ( Qiagen ) , including the DNase treatment step , and reverse transcriptase reactions were performed using AffinityScript RT-RNase ( Stratagene ) . Quantitative PCR reactions were performed in an Applied Biosystems 7500 Real-Time PCR System using Brilliant SYBR Green qRT-PCR master mix ( Stratagene ) . PCR thermal cycling conditions were an initial step at 95°C for 5 min followed by 40 cycles at 95°C for 15 s , 60°C for 20 s and 72°C for 20 s . In each assay , "no-template" controls were included and melting curve analysis was performed to confirm a single PCR product . All experiments were done in triplicate for both the gene of interest and the control gene ( GPB1 or ACT1 , see figure legends ) . Data were normalized to the control genes and relative expression was determined by the 2−ΔΔCT method . The sequences of primers used are listed in Table S2 . Detection of Chr1 disomy by qPCR was adapted from the method described by [81] . To perform the mating cell fusion assays , wild-type and mutant strains containing HYG , NAT , or NEO resistance genes were grown in YPD liquid overnight . Yeast cells were washed and adjusted to 5×106 cells/ml , mixed 1∶1 , and grown on V8 medium ( pH 5 . 0 ) in the dark for 48 h . The mating colonies were collected by scraping , resuspended in sterile water , and serial dilutions were plated onto YPD medium containing hygromycin , nourseothricin , and/or G418 to select for cell-cell fusion products that have two drug-resistance markers ( one from either parent ) . Plates were incubated at 30°C for 3 days until colonies formed and were then counted . The average number of colonies in triplicate assays was calculated and the fusion efficiency in apoptotic-mutants was determined as a function of the wild-type . For competition assays , overnight cultures in YPD were washed three times with sterile water and adjusted to a density of 1×107 cells/ml . The following strains were paired in a 1∶1 ratio: H99 and CPS76 ( H99 HYG ) ; H99 and CPS3; H99 and YPH104; and H99 and CPS89 . Each mixture was inoculated onto fresh YPD medium and incubated at 30°C for the indicated times . Fifty isolated colonies from each time point were replica plated onto selective media and survival rates were calculated . Growth curves were determined by diluting the strains to OD600 of 0 . 01 in YPD and incubation at 30°C without agitation . The turbidity of quadruplicate assays was measured at OD600 every hour with a microdilution plate reader ( Sunrise , Tecan ) . Readings were corrected for background ( YPD media , no cells added ) , averaged , and plotted versus time in Excel ( Microsoft Office ) . All quantitative data are reported as means ± standard deviation and are derived from at least three independent experiments . The significance of the data was assessed using two-tailed t-tests to compare mutants with the wild-type strain . For comparisons of mice survival data , the logrank test was performed using Prism 4 ( GraphPad Software ) . A p-value of less than 0 . 05 was considered significant . AIF1 ( CNAG_04521 ) ; BUB1 ( CNAG_03184 ) ; COX1 ( CNAG_09009 ) ; ENDOG ( CNAG_02204 ) ; IAP1 ( CNAG_04708 ) ; MCA1 ( CNAG_04636 ) , MCA2 ( CNAG_06787 ) , and AIF1 from RCT 17 strain ( JN606083 ) . | Fungal pathogens can cause life-threatening diseases , and the infections that they cause are notoriously difficult to treat . Despite the availability of antifungal drugs , most inhibit fungal growth but do not consistently or efficiently eliminate the pathogen . In addition , fungal cells are very similar to human cells , and therefore , many of the available antifungal agents have toxic side effects . Thus , more efficient drugs with less adverse effects are clearly needed . We investigated apoptosis , a process in which cells become programmed to commit suicide , in the pathogenic fungus Cryptococcus neoformans . We studied genes that regulate apoptosis in C . neoformans and , after inactivating three genes involved in this pathway , we observed defects in sexual reproduction . Such mating defects decrease the production of spores , which are inhaled and cause cryptococcal disease . We also showed that the absence of one investigated apoptotic gene , aif1 , resulted in the selection of antifungal-resistant pathogens ( when the fungal cells no longer respond to the drug ) , which makes treatment of the disease more difficult . The discovery of drugs that kill fungal cells specifically without affecting the cells of the patient being treated holds great potential . Therefore , triggering apoptosis should be further investigated as a new approach to treat fungal pathogens . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"medicine",
"infectious",
"diseases",
"drugs",
"and",
"devices",
"genetics",
"biology",
"molecular",
"cell",
"biology",
"genetics",
"and",
"genomics"
] | 2011 | Deletion of Cryptococcus neoformans AIF Ortholog Promotes Chromosome Aneuploidy and Fluconazole-Resistance in a Metacaspase-Independent Manner |
CTP synthase is an essential enzyme that plays a key role in energy metabolism . Several independent studies have demonstrated that CTP synthase can form an evolutionarily conserved subcellular structure termed cytoophidium . In budding yeast , there are two isoforms of CTP synthase and both isoforms localize in cytoophidium . However , little is known about the distribution of CTP synthase isoforms in Drosophila melanogaster . Here , we report that three transcripts generated at the CTP synthase gene locus exhibit different expression profiles , and three isoforms encoded by this gene locus show a distinct subcellular distribution . While isoform A localizes in the nucleus , isoform B distributes diffusely in the cytoplasm , and only isoform C forms the cytoophidium . In the two isoform C-specific mutants , cytoophidia disappear in the germline cells . Although isoform A does not localize to the cytoophidium , a mutation disrupting mostly isoform A expression results in the disassembly of cytoophidia . Overexpression of isoform C can induce the growth of the cytoophidium in a cell-autonomous manner . Ectopic expression of the cytoophidium-forming isoform does not cause any defect in the embryos . In addition , we identify that a small segment at the amino terminus of isoform C is necessary but not sufficient for cytoophidium formation . Finally , we demonstrate that an excess of the synthetase domain of CTP synthase disrupts cytoophidium formation . Thus , the study of multiple isoforms of CTP synthase in Drosophila provides a good opportunity to dissect the biogenesis and function of the cytoophidum in a genetically tractable organism .
Nucleotides not only serve as building blocks to make up DNA and RNA , but also play critical roles in many additional biological processes . For example , ATP acts as the most widely used biological energy carrier , GTP participates in intracellular signaling and is used as an energy reservoir , and CTP is involved in phospholipid and sialoglycoprotein synthesis . Compartmentation is important for the efficiency of biological processes in cells [1] , [2] . Mitochondria and chloroplasts are specialized organelles that contain ATP synthase , the enzyme that makes ATP . The mitochondrion serves as the main site for ATP synthase to generate ATP , and has been considered the ‘power house’ of a cell because it is responsible for producing 90% of the cellular energy . Defects in mitochondria have been linked to a wide range of human diseases such as mitochondrial myopathy , Leigh syndrome , Parkinson's disease and diabetes [3] , [4] . Thus , it is not surprising that ATP synthase and the mitochondrion have been extensively studied [5] . By contrast , until very recently , little was known about the subcellular distribution of CTP synthase . CTP can be synthesized through either the salvage pathway or the de novo pathway in many cells [6] , [7] , [8] , [9] . The rate-limiting step of de novo CTP biosynthesis is catalyzed by the CTP synthase ( CTPsyn ) enzyme [6] , [7] , [8] , [9] . CTPsyn catalyses a set of three reactions: a kinase reaction being Mg2+-ATP-dependent phosphorylation of the UTP uracil O4 atom; a glutaminase reaction being rate-limiting glutamine hydrolysis to generate ammonia; and a ligase reaction being displacement of the uracil O4 phosphate by ammonia 10 , 11 , 12 , 13 , 14 . In 1978 , Weber and co-workers found that CTP synthase activity in hepatomas was elevated [15] . Subsequent studies demonstrated that unregulated CTP levels and increased CTP synthase activity are features of many forms of cancer such as leukemia , hepatomas , and colon cancer [15] , [16] , [17] , [18] , [19] , 20 , 21 , 22 , 23 , 24 , 25 , 26 , 27 . Furthermore , CTP synthase is an attractive target for drug development against viral [28] and parasitic disease ( e . g . African sleeping sickness [29] , [30] , malaria [31] , and infectious blindness [32] ) . In 2010 , three independent studies reported that CTPsyn is compartmentalized in filamentary structures in bacteria , yeast , fruit flies and rats [33] , [34] , [35] . More recently , studies have shown that CTPsyn can form filaments in human cells as well [36] , [37] . These CTPsyn-containing structures were termed cytoophidia ( meaning ‘cellular serpents’ in Greek ) [33] , [36] , CtpS filaments [34] , CTP synthase filaments [35] , or rods and rings ( RR ) [37] . For simplicity , the term ‘cytoophidia’ is used in this paper to refer to the apparently equivalent structures that contain CTPsyn ( for review see [38] ) . In addition to CTPsyn , some other metabolic enzymes can form filamentous structures in budding yeast [35] and Drosophila melanogaster [39] . Recent studies on the cell biology of CTPsyn suggest that: 1 ) the cytoophidium is strikingly conserved across prokaryotes and eukaryotes; 2 ) the cytoophidium is likely to represent a novel type of subcelluar structure; and 3 ) the conservation of the filament-forming property of CTPsyn offers an exciting opportunity to study the cell biology of metabolic pathways . Two genes , URA7 and URA8 , encode two isoforms of CTPsyn in Saccharomyces cerevisiae , and both isoforms have been shown to form filaments [35] . In humans , two isoforms of CTPsyn , which share 74% identity , are encoded by two genes , CTPsyn1 and CTPsyn2 . The two human CTPsyn isoforms share 44%–55% identity with the two isoforms of CTPsyn in S . cerevisiae . In Drosophila melanogaster , CTP synthase is encoded by the gene CG6854 . Bioinformatic data suggest that the CG6854 gene locus can produce three transcript variants that correspond to three isoforms . The aim of this study was to determine the expression and subcellular distribution of the individual isoforms , which have so far remained elusive . Here , we report that three isoforms produced in the CG6854/CTP synthase locus in D . melanogaster localize to distinct subcellular compartments , while only one isoform ( isoform C ) forms the cytoophidium . Overexpression of isoform C can not only increase the length and thickness of cytoophidia in ovarian cells in which cytoophidia are abundant , but also promote cytoophidium assembly in embryos where cytoophidia are less abundant or not detectable . In addition , we identify that a short N-terminal segment that is only present in the cytoophidium-forming isoform is necessary but not sufficient for cytoophidium formation . Finally , we demonstrate that cytoophidium formation can be disrupted by excessive expression of the synthetase domain of CTP synthase . Together , our study of multiple isoforms of CTP synthase in Drosophila provides a good opportunity to investigate the biogenesis and function of the cytoophidium in a genetically tractable organism .
In Drosophila melanogaster , the CTPsyn gene ( CG6854 , FBgn0262707 ) localizes at Chromosome 3L:15091235 . . 15106103 ( www . flybase . org ) . The Drosophila CG6854/CTPsyn gene locus produces three transcripts ( Figure 1A ) . The first transcript , which is 1867 nt in length , encodes isoform A , a 429-aa protein . Isoform A consists of two different domains: the N-terminal myb/SANT-like domain in Adf-1 ( MADF ) and the C-terminal BEAF , Suvar ( 3 ) 7 and Stonewall ( BESS ) motif ( Figure 1B ) . This architecture is conserved in at least 14 Drosophila proteins [40] . The MADF domain , consisting of 93 aa , is thought to direct the sequence-specific DNA binding to a site consisting of multiple tri-nucleotide repeats . In D . melanogaster , 46 genes contain the MADF domain , which is frequently associated with the BESS domain [40] , a domain consisting of approximately 40 aa with two predicted alpha helices . The BESS domain , which is predicted to be a DNA binding domain , appears to be specific to Drosophila [40] . The second transcript , 2554 nt in length , produced by the CG6854/CTPsyn gene locus encodes the B isoform , a 623-aa protein . The transcript for isoform B contains six exons , among which two exons are shared with the transcript for isoform A . The first 51 aa of isoforms A and B are identical ( Figure S1 ) . Isoform B contains two predicted incomplete domains ( MADF and CTPsyn synthetase domain ) and a type 1 glutamine amidotransferase ( GAT ) domain ( Figure 1A , 1B ) . The incomplete MADF domain present in this isoform is only 41 aa long , and the truncated CTPsyn synthetase domain consists of 224 aa . The third transcript of the CG6854/CTPsyn , which is 2406 nt long , produces a 627-aa protein , the isoform C . It contains two domains: the synthetase domain and the GAT domain ( Figure 1A , 1B ) . Isoform C has 5 UTP binding sites that are required for the synthesis of CTP from UTP , while isoform B only has 4 UTP binding sites ( Figure 1B ) . This suggests that only isoform C has the full function of a typical CTPsyn . Orthologous proteins present in organisms from yeast to mammals are mostly aligned to CTPsyn isoform C in Drosophila ( Figure S2 ) . It appears that two protein trap lines ( CA06746 and CA07332 ) described previously [33] , [36] , [41] have green fluorescence protein ( GFP ) trapped in the first and second exons of CTPsyn isoform C ( Figure 1B ) . To determine the expression pattern of the transcripts that encode these three isoforms we performed quantitative PCR ( qPCR ) using isoform-specific primers ( Table S1 ) . We found that the transcript for isoform B was expressed at a very low level in most developmental stages except early embryos ( 0–4 h ) ( Figure 2 ) . In contrast , the isoform A and C transcripts were both shown to be very abundant throughout all developmental stages ( Figure 2 ) . However , we observed that isoform A and C have very distinct expression profiles . During embryogenesis , both transcripts showed the highest levels in the 0–4 h embryos . However , in embryos at 4–8 h , while the isoform C transcript stayed at a similar level to that in 0–4 h embryos , the transcript for isoform A decreased about 8-fold . In late-stage embryos , the expression of the isoform C transcript decreased 1 . 5 to 2-fold , while isoform A showed similar levels to those in 4–8 h embyos ( Figure 2A ) . The isoform C transcript showed modest expression in larval and pupal stages , while the isoform A transcript showed very strong expression in third instar larvae that reached a peak in the early pupal stage ( Figure 2B ) . In adult flies , very little expression of isoform B could be detected , while isoforms A and C showed similar and abundant expression , especially in heads and gonads , with the highest level in the ovary ( Figure 2C ) . Our previous studies using multiple antibodies against CTPsyn and two independent CTPsyn-GFP protein trap lines demonstrated that CTPsyn localizes to the cytoophidium in many tissues in Drosophila ( Figure 3A–3C ) [33] , [36] , [41] . However , it is unclear whether all three isoforms produced by the Drosophila CTPsyn gene locus localize to cytoophidia . To determine the distribution of individual isoforms , we made both tagged and untagged constructs of each isoform . Isoform A , when tagged by Venus either at the N-terminus or C-terminus , localized to punctate structures in the nucleus ( Figure 3D–3F ) . Ubiquitous expression of isoform A transgene by an actin-GAL4 driver resulted in lethality in Drosophila at a very early stage . Expression of isoform A transgene in the germline led to small ovaries and sterility in female flies ( data not shown ) . However , cytoophidia , as revealed by immunostaining using antibodies against CTPsyn , did not show any obvious morphological change in germline cells when the isoform A transgene was expressed ( Figure 3D–3F ) . The CTPsyn isoform B transgene , either with an N-terminal tag or a C-terminal tag , showed a dispersed distribution in the cytoplasm ( Figure 3G–3I ) . The morphology of cytoophidia , as analyzed by antibody staining against CTPsyn , remained unchanged in nurse cells when the isoform B transgene was overexpressed with actin-GAL4 . However , compared to flies overexpressing isoform A , overexpressing isoform B in the germline did not cause sterility . Flies overexpressing CTPsyn isoform B ubiquitously were also viable . The CTPsyn isoform C transgene , when tagged by Venus either at its C-terminus or N-terminus , localized to cytoophidia ( Figure 3J–3O ) . The Venus signal showed an almost identical pattern to that labeled by antibodies against CTPsyn ( Figure 3J–3O , Figures S3 and S4 ) . When CTPsyn isoform C with a C-terminus-tagged GFP was overexpressed in female germline cells , we observed that the number and length of macro-cytoophidia increased dramatically ( Figure 3J–3L , Figure S3 ) . Many long cytoophidia also appeared very thick . Overexpressing CTPsyn isoform C without a tag showed a similar pattern as that with a C-terminus-tagged Venus ( Figure S5 ) . Flies overexpressing isoform C ubiquitously were still viable and fertile , in contrast to those flies overexpressing isoform A . When Venus was tagged to the N-terminus of isoform C , cytoophidia appeared to be straight . The number of cytoophidia increased considerably; the length and thickness of each individual cytoophidium however did not change dramatically ( Figure 3M–3O ) . These short and straight cytoophidia very often tangled up with each other , making them appear spiky ( Figure S4 ) . In this case , the Venus tag may obstruct a critical cytoophidium-forming region at the N-terminus of CTPsyn isoform C , as described below . Previous transposon screens yielded a number of inserts in and around the CTPsyn locus . We found that three mutants CTPsynd06966 , CTPsyne01207 and CTPsynd07411 ( Figure 4A ) were homozygous lethal . Complementation analysis showed that CTPsynd06966 and CTPsyne01207 failed to complement each other . However , CTPsynd07411 were able to complement to either CTPsynd06966 or CTPsyne01207 , suggesting that CTPsynd07411 affects different isoform from the other mutants . This is confirmed by qPCR analysis . In comparison with wild-type flies , CTPsynd07411 flies showed a 6-fold decreased level of the transcript for CTPsyn isoform A , while no change in isoform B and only 20% decrease in isoform C expression ( Figure 4B ) . On the contrary , isoform C expression decreases significantly in both CTPsynd06966 and CTPsyne01207 flies ( Figure 4B ) . We balanced these mutants with TM6B which carries the dominant recessive mutation Tubby ( Tb ) that makes the larvae and pupae have short bodies which are easily distinguishable [42] . Eggs from all three CTPS mutants were collected in an apple juice plates analysed every day from 3 days after egg deposition . From early on , all the mutant larvae were significantly smaller compared to wild-type control ( Figure 5 ) . Both CTPsynd06966 ( Figure 5A–5E ) and CTPsyne01270 ( Figure 5F–5J ) survive and continue to grow until 7 days after egg deposition . However , CTPsynd07411 only survives until 5 days after egg deposition with very little growth ( Figure 5K–5M ) . Even at 5 days after egg deposition , the CTPsynd07411 mutant larvae are still very small . All three CTPsyn mutants failed to develop into proper pupation stage , although they formed pseudo-pupa occasionally . To study the effect on cytoophidia formation , mitotic clones were generated in female germline cells . As expected , egg chamber clones from two isoform C-specific mutants ( CTPsynd06966 and CTPsyne01207 ) showed disruption in the formation of cytoophidia ( Figure 6A–6F ) . To our surprise , we were unable to detect cytoophidia in the germline cells even from the third mutation CTPsynd07411 in which isoform A expression was disrupted , while clear cytoophidia could be observed in adjacent wild-type egg chambers ( Figure 6G–6I ) . In addition to 15 nurse cells and one oocyte , an egg chamber contains several hundred to a thousand follicle cells , which form a monolayer epithelium surrounding the large nurse cells and the oocyte [43] . Our previous studies indicate that each follicle cell contains only one cytoophidium . Cytoophidia exhibit a similar length in follicle cells within the same egg chamber [33] . This unique feature of cytoophidia makes the follicle cell epithelium an ideal model to study the biogenesis of the cytoophidium . To better understand the role of CTPsyn isoform C in cytoophidium formation , we analyzed cytoophidia in follicle cells from flies in which this isoform was overexpressed . We predicted several possible outcomes: the number of cytoophidia could increase from one to many , the length of the cytoophidia could increase , or both changes could happen . Our results showed that the length of cytoophidia increased dramatically in the follicle cells ( Figure 7A–7D ) . In CTPsyn protein trap CA06746 flies , cytoophidia are less than 2 µm in early-stage follicle cells ( average 1 . 92±0 . 23 µm at stages 4 to 6 , n = 54 ) and they reach average 3 . 23±0 . 47 µm at stage 9 ( n = 196 ) and 3 . 81±0 . 47 µm at stage 10A ( n = 80 ) . When the expression of CTPsyn isoform C transgene was driven by actin-GAL4 , we observed that the average lengths of cytoophidia can reach to 5 . 30±0 . 77 µm at stage 6 ( n = 36 ) , 9 . 33±1 . 54 µm at stage 9 ( n = 69 ) , and 9 . 77±1 . 36 µm at stage 10A ( n = 54 ) , respectively ( Figure 7A–7D ) . However , in many cases , there was still only one cytoophidium per follicle cell even when overexpressing isoform C ( Figure 7E , 7F ) . To test if CTPsyn isoform C cell-autonomously affects cytoophidium formation , we generated mitotic clones in which nuclear GFP labeled a patch of cells overexpressing isoform C . We observed that cytoophidia in cloned cells were much thicker and longer than in the neighboring wild-type cells ( Figure 7G ) . Follicle cells within a clone contained cytoophidia of similar lengths . The extended cytoophidia curled up around the nucleus in many cells , as they were presumably restrained by the cytoplasm . Wild-type follicle cells near the clones had cytoophidia of similar lengths to those cells that were further away from the clones , suggesting that the level of CTPsyn isoform C regulates the formation of cytoophidia in a cell-autonomous manner . The above results in the ovary suggest that CTPsyn isoform C can dramatically affect the length of cytoophidia in cells that normally contain cytoophidia . Previous studies indicate that not every cell contains a detectable cytoophidium [33] , [35] , [36] . We have observed that Drosophila embryos do not exhibit detectable large cytoophidia ( i . e . macro-cytoophidia ) , either by immunostaining with antibodies against CTPsyn or by CTP syn::GFP protein trap line ( Figure 8A , 8C , 8E ) . To determine if CTPsyn isoform C can induce de novo cytoophidium assembly , we ectopically expressed this isoform in Drosophila embryos . Our results showed that in embryos overexpressing CTPsyn isoform C , cytoophidium formation was promoted throughout embryogenesis , starting from stage 1 ( Figure 8B , 8D , 8F ) . We observed more cytoophidia in stage 15 embryos than in stage 12 embryos ( Figure 8F ) , which could be due to differential expression of the actin-GAL4 driver during embryogenesis . Embryos with highly abundant cytoophidia developed normally . These results indicate that CTPsyn isoform C plays a critical role in the de novo assembly of cytoophidia , which seemingly do not impair embryogenesis in Drosophila . Most regions of CTPsyn isoform B are identical to those of isoform C . The only difference between these two isoforms lies in their amino ( N ) terminus: 56 amino acids in isoform C and 52 amino acids in isoform B ( Figure 1A , 1B ) . To identify the critical regions for cytoophidium formation , we generated transgenic flies that carried Venus-tagged constructs based on various regions of CTPsyn isoform C ( the cytoophidium-forming isoform ) ( Figure S6 ) . Each of these transgenes were induced by a maternal triple driver ( MTD ) so their cytoophidium-forming ability could be monitored in the female germlines ( Figure 9 ) . We found that a truncated isoform C without a 56-aa N-terminal segment ( N-term ) was unable to form cytoophidia , suggesting that N-term plays an important role in cytoophidium formation ( Figure 9D–9L ) . As shown above , we noted that Venus tagging at the N-terminal of full-length isoform C affected the morphology of cytoophidia . This may be because the Venus tag interferes with this critical cytoophidium-forming segment . To test if N-term alone is sufficient for cytoophidium formation , we tagged Venus with N-term and found that N-term-GFP did not localize to the cytoophidum ( Figure 9A–9C ) . Our results indicate that N-term of CTPsyn isoform C is necessary but not sufficient for the formation of cytoophidia . The concentration of CTPsyn affects the equilibrium between its monomeric , dimeric and tetrameric forms [44] . Each monomer contains two functional domains: the synthetase domain and the GAT domain [45] , [46] , [47] . The GAT domain catalyses GTP-activated glutamine hydrolysis , while the synthetase domain mediates Mg2+-ATP-dependent phosphorylation of the UTP uracil O4 atom and displacement of the uracil O4 phosphate by ammonia [48] , [49] . CTPsyn activity requires oligomerization , and each synthetase active site and essential ATP- and UTP-binding surfaces are contributed by three monomers [13] . The triphosphate moiety of the CTP product overlaps the binding site for the UTP substrate , while the CTP cytosine ring resides at a separate site [14] . Using time-lapse microscopy , Gitai and co-workers have shown that mCherry-CTPsyn can grow from a focus to a long filament in curved bacterium Caulobacter crescentus [34] . In the same study , they have also shown that purified E . coli CTPsyn molecules can form filaments in vitro [34] . Therefore , it is likely that CTPsyn in a cytoophidiun is in its polymeric form . Overexpression of isoform C induces cytoophidium formation , suggesting that the concentration of CTPsyn also affects the equilibrium between its oligomeric and polymeric forms . Since the synthetase domain is critical for oligomerization of CTPsyn , we hypothesized that an excess of the synthetase domain , by competitively binding full-length CTPsyn to form oligomers , can disrupt the polymerization of CTPsyn to form cytoophidia . To test this hypothesis , we generated a synthetase domain transgene tagged with Venus . When this transgene was overexpressed in female germline cells , we found that the Venus signal was dispersed in the cytoplasm and did not localize to the cytoophidium ( Figure 9M–9O ) . As predicted , the endogenous CTPsyn failed to form detectable cytoophidia in germline cells while cytoophidia in follicle cells were not affected . Consistent with this idea , we found that overexpression of the GAT domain , which would be dispensable for tetramerization of CTPsyn , showed no obvious effect on endogenous cytoophidium formation ( Figure 9P–9R ) . To better understand the effect of ectopic expression of these two domains , we analysed cytoophidia in follicle cells . Using inducible driver , we generated mitotic clones in which nuclear GFP labeled those cells expressing the transgene of interest . In our case , the transgenes were synthetase domain or GAT domain , both of which were tagged with Venus . When follicle cells contained excess synthease domain-Venus , cytoophidia were not longer maintained , while wild-type follicle cells in the same egg chamber have obvious cytoophidia ( Figure 10A–10F ) . Similar to the results obtained from germline cells , overexpression Venus-GAT domain did not show obvious effect on cytoophidium formation in follicle cells ( Figure 10G–10I ) .
Three transcripts generated at the CTPsyn gene locus give rise to three protein products in Drosophila melanogaster ( CTPsyn-PA , PB and PC , Figure 1A which distribute differently within the cell . Do the three protein products belong to three isoforms of the same protein or should they be considered as products from different genes ? The arguments that suggest that the three products are derived from different genes as follows: CTPsyn-PA and CTPsyn-PC do not share any amino acid sequence , and they have very distinct localizations , with PA in the nucleus and PC in cytoplasmic cytoophidia . While orthologous proteins of CTPsyn-PC are much conserved from bacteria to human , CTPsyn-PA and PB seem to be unique to Drosophila ( Figure S2 ) . Among those three protein products only CTPsyn-PA contains the BESS domain , which appears to be specific to Drosophila . The BESS domain can bind to DNA , which may explain the nuclear localization of CTPsyn-PA . Thus it appears reasonable to suggest that the Drosophila CTPsyn gene locus contains three genes , with only one encoding full-length CTPsyn , of which cytoophidium are composed . However , there are several lines of evidence to support those three protein products being considered as isoforms derived from the same gene . First , three transcripts from the CTPsyn gene locus in Drosophila melanogaster are overlapping , not only at untranslated regions , but also at coding sequences . Second , at the protein level , although CTPsyn-PA and PC are non-overlapping , they both overlap with CTPsyn-PB at different regions . It appears that CTPsyn-PB is the bridge that links PA and PC at the same locus . Third , even though PA , PB and PC have distinct subcellular localizations , there are examples of different isoforms of the same protein showing distinct localizations in the cell . Chan and colleagues [37] have shown that another nucleotide metabolic enzyme , IMP dehydrogenase ( IMPDH ) , colocalizes with CTPsyn in the rods and rings ( i . e . cytoophidia ) in human and mouse cell lines . A more recent study in Drosophila suggests that IMPDH can localize both in the cytoplasm as a metabolic enzyme and in the nucleus as a transcriptional factor [50] . Finally , in the CTPsynd07411 mutant , the formation of cytoophidia is disrupted . Since the transcript of CTPsyn-PC only has 20% decrease in CTPsynd07411 ( in comparison , the transcript of CTPsyn-PA decreases its expression by 83% ) , it is unlikely that the disassembly of cytoophidia in CTPsynd07411 could be explained by the change of CTPsyn-PC alone . A possible explanation is that cytoophidia not only requires CTPsyn-PC , but also requires CTPsyn-PA for assembly , even though the latter does not localize on the cytoophidium . This suggests that there is a functional link between CTPsyn-PA and PC in Drosophila . Further analysis of multiple alleles at the CTPsyn gene locus in Drosophila will shed new insights into the relationship of these three isoforms . There are an abundant of cytoophidia in female germline cells , i . e nurse cells and oocytes , especially during mid-oogenesis . There are also two types of cytoophidia in nurse cells and oocytes: macro- and micro-cytoophidia . The numbers of macro-cytoophidia are less at late stage oogenesis , suggesting that the assembly and disassembly of cytoophidia is linked to developmental status . Overexpressing CTPsyn isoform C can not only promote cytoophidium assembly in cells with abundant cytoophidia , such as nurse cells , oocytes and follicle cells in the ovaries , but also induce the formation of cytoophidia in tissues that usually have few , if any , macro-cytoophidia . This suggests that CTPsyn isoform C is a major factor promoting cytoophidium assembly . CTPsyn isoform A is distributed in the nucleus , yet a mutation primarily disrupting isoform A affects cytoophidium formation in the cytoplasm . How does a nuclear protein affects a structure in the cytoplasm ? One possibility is that CTPsyn isoform A might affect isoform C post-transcriptionally . Another possible explanation could be that the biogenesis of cytoophidia requires a phase in the nucleus , which somehow is dependent on isoform A . Alternatively , the maintenance of cytoophidia might require one or a few factors which are regulated by isoform A . Further detailed studies of isoform A are necessary to get insight into the assembly of cytoophidia . The sequence of CTPsyn isoform B is 91% identical with that of isoform C , yet only the latter forms the cytoophidium . It suggests that the 56-aa present in the N-terminus of CTPsyn isoform C but not in isoform B is critical for the formation of cytoophidia , as confirmed experimentally in this study . This N-terminal region is highly conserved among CTPsyn molecules from bacteria to humans; thus our results in Drosophila would most likely be acceptable across species . When these 56 aa were tagged with Venus at its N-terminus we did not observe cytoophidium localization . Two possibilities can be considered . One possible explanation is that the large Venus tag ( about 260 aa ) interferes with cytoophidium formation sites at this short peptide ( 56 aa ) . It is also possible that the formation of cytoophidia requires multiple binding sites including the ones presenting at those N-terminal 56 aa . Overexpressing the synthetase domain of CTPsyn isoform C disrupts the formation of endogenous cytoophidia . The excessive synthetase domain might competitively bind and block some critical cytoophidium-forming sites that present in the full-length CTPsyn isoform C . Those critical cytoophidium-forming sites could be uncoupled from catalytic sites in CTPsyn isoform C . Gitai and colleagues have shown that a point mutation at the synthetase domain ( G147A ) of Caulobacter crescentus CTPsyn that inactive a catalytic site does not cause any detectable change in the frequency or morphology of filamentous structure [34] . In summary , the assembly of cytoophidia appears to be a multiple-step process . While CTPsyn isoform A is primarily localizing in the nucleus , a mutation specific to this nuclear isoform disrupts the formation of cytoophidia in the cytoplasm . Further studies on the relationship of different CTPsyn isoforms in Drosophila would be helpful to understand the biogenesis of the cytoophidium . Note added in proof: While this paper was under review , a new release from the Flybase ( www . flybase . org; FB2012_06 , released November 6th , 2012 ) showed a fourth transcript of at the CG6854/CTPsyn gene locus , which is 2062 nt in length , encoding CTPsyn isoform D ( 429 aa ) . While the CTPsyn isoform D protein has the exact sequence as isoform A , the first exon of the CTPsyn isoform D transcript overlaps with the first and second exons of the CTPsyn isoform C transcript .
All stocks were raised at 25°C on standard cornmeal media . y w flies were used as a control in all experiments if not indicated . CTP synthase GFP protein trap lines , CA06746 and CA07332 , were gifts from Michael Buszczak and Allan Spradling [41] . The inducible GAL4 driver stock hsFLP , UAS-GFPnls; tub-{GAL80}-GAL4 was used to generate mosaic overexpression and hsFLP , UAS-GFPnls; UAS-Dicer-2; tub-{GAL80}-GAL4 was used to generate mosaic overexpression of transgenes [51] . CTPsyn mutant stocks ( CTPsyne01207 ( PBac{RB} ) , CTPsynd06966 ( P{XP} ) , CTPsynd07411 ( P{XP} ) used in this study were obtained from Bloomington stock centre and the Harvard Exelixis collection [52] . RNA extraction was performed by homogenizing samples using the Qiagen QIAshredder ( Cat . no . 79654 ) and RNA was extracted using the Qiagen RNeasy Plus Mini Kit ( Cat . No . 74134 ) as per the manufacturer's instructions . Samples were kept at −80°C . Reverse transcription was carried out using the Qiagen QuantiTect Rev . Transcription Kit ( Cat . no . 205311 ) with the gDNA removal step following the manufacturer's instructions . The cDNA were then further diluted 1∶10 with nuclease-free water and kept at −20°C . About 1 µl of diluted cDNA from the reverse transcription was mixed with Fast SYBR Green Master Mix ( Applied Biosystems Cat . no . 4385612 ) and 1 µM of primers ( Table S1 ) for each 10 µl qPCR reactions . The reactions were carried out using the 7500 Fast Real-Time PCR System ( Applied Biosystems ) on the Fast setting: initial denaturation at 95°C for 20 s , denaturation at 95°C for 3 s , primer annealing and elongation at 60°C for 30 s , repeated for 40 cycles , then final denaturation at 95°C for 15 s , and final primer annealing and elongation at 60°C for 1 min . Expression values were normalized using reference gene RP49 . To make the untagged constructs , the gene sequence was amplified by PCR from cDNA using Pfx polymerase ( Invitrogen 11708-021 ) and specific primers with additional restriction site sequences ( Table S2 ) . Both PCR products and the pUASp-K10-attB vector were then digested individually using specific restriction enzymes and ligated using T4 DNA ligase ( NEB M202 ) . The product was then sequenced before being injected into embryos . All tagged constructs were made using an enhanced GFP or Venus [53] tagged UASp vector from the Gateway clones ( Terrance Murphy collection ) , and the untagged constructs were made using the modified pUASp vector which contains attB sites for PhiC31 integrase-mediated site-specific injection [54] . To make the constructs , these cDNA clones were used: LD27370 ( isoform A ) , LD27537 ( isoform B ) and LP25005 ( isoform C ) , which were acquired from the Drosophila Genomics Resource Center gold collections . To make the Venus tagged construct , the gene sequence was amplified by PCR from specific cDNA using Pfx polymerase ( Invitrogen 11708-021 ) and isoform-specific primers ( Table S3 ) that have additional sequences as described by the Drosophila Gateway Vector Collection protocol . The PCR products were then cloned into a pENTR using the pENTR/D-TOPO cloning kit ( Invitrogen K240020 ) . Then , to make the destination construct , the sequences cloned into the pENTR vector were recombined into the UASp-Venus ( N-terminal tagged or C-terminal tagged ) vector using LR clonase II ( Invitrogen 11791020 ) . The final constructs were then sequenced before being sent to be injected . Embryo injections were carried out by GenetiVision Inc . ( Texas , USA ) . Constructs with the attB sequence were injected using the PhiC31 integrase-mediated site-specific technique with specific landing site [55] . P-element constructs were injected and the progenies were scored using eye marker . To generate clones overexpressing CTPsyn transgenes , hsFLP , UAS-GFPnls; sp; tub {GAL80} GAL4/SM5 , Cy-TM6 Tb flies were crossed to UASp-CTPsyn isoform C . To generate follicle cell clones , eggs were collected in vials for 24 h and heat-shocked after 4 days in a 37°C water bath for 1 h . The cells expressing the construct were marked by the presence of GFP in the nucleus . Three CTPsyn mutants ( CTPsynd06966 , CTPsyne01207 and CTPsynd07411 ) were used in this study . A FRT site located at the 79D-F region ( FRT2A ) was recombined to all three lines . Mitotic clones were generated by crossing y w; CTPsynmutant e FRT2A/TM3 Ser males to hsFLP; ubi GFP FRT2A/TM3-Ser females . The resulting progeny were heat-shocked at 37°C for 1 h during the third instar larval stages . Ovaries from the female progeny of genotype y w/hsFLP ; CTPsynmutant e FRT2A/Ubi-GFP FRT2A were harvested and stained for GFP . Egg chambers homozygous for the mutation were negative for GFP in the germline , whereas wild-type egg chambers were marked with GFP . Drosophila ovaries were dissected in Grace's insect medium ( Invitrogen Cat . no . 11605045 ) and fixed with 4% paraformaldehyde for 10 min , washed with PBT ( PBS+0 . 4% Triton X-100 ) , blocked with 5% horse serum for 1 h and incubated in primary antibodies at room temperature overnight . Samples were washed with PBT and then incubated overnight with the DNA dye Hoechst 33342 and secondary antibodies . Primary antibodies used in this study included rabbit anti-CTPsyn ( 1∶1000; y-88 , sc-134457 , Santa Cruz BioTech Ltd , Santa Cruz , CA , USA ) , mouse anti-Hu-li tao shao ( Hts ) ( 1∶20; 7H9 1B1 , Developmental Studies Hybridoma Bank , Iowa City , IA , USA ) and mouse anti-Engrailed ( 1∶1000; 4D9 , Developmental Studies Hybridoma Bank ) . Secondary antibodies used in this study were anti-mouse , rabbit , goat or guinea pig antibodies that were labeled with Alexa Fluor 488 , 546 or 633 dyes ( Molecular Probes ) , or with Cy5 or Dylight 649 ( Jackson ImmunoResearch Laboratories , Inc . ) . All samples were examined and captured under laser-scanning confocal microscopes ( Zeiss LSM 510 META , Oberkochen , Germany; and Leica TCS SP5II , Leica Microsystems CMS GmbH , Mannheim , Germany ) . The lengths of cytoophidia in follicle cells were measured by tracing with straight or segmented lines using ImageJ ( v1 . 43 U ) ( http://rsbweb . nih . gov/ij/ ) . | DNA and RNA are made up from basic building blocks called nucleotides . Those nucleotides also play essential roles in many other biological processes . To separate biological processes within a cell is an important feature of all cell types . For example , mitochondria are specialized structures that contain ATP synthase , the enzyme that makes the nucleotide ATP . While mitochondria and ATP synthase have been studied for about 100 years , it was only very recently that we realized that there are specialized subcellular structures that contain CTP synthase , the enzyme that makes up another basic nucleotide CTP . Several independent studies have shown that CTP synthase molecules can form a filamentous structure called the cytoophidium ( meaning “cellular snake” in Greek ) or CTP synthase filament in bacteria , budding yeasts , fruit flies , and rat and human cells . In budding yeast , there are two isoforms of CTP synthase and both isoforms localize in the cytoophidium . Here , we report that three CTP synthase isoforms in fruit flies show a distinct subcellular distribution and only one isoform forms the cytoophidium . Thus , the study of multiple isoforms of CTP synthase in the fruit fly gives us a good way to begin to learn how and why CTP synthase molecules form this snake-like structure . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"animal",
"models",
"developmental",
"biology",
"cellular",
"structures",
"drosophila",
"melanogaster",
"model",
"organisms",
"molecular",
"cell",
"biology",
"cell",
"biology",
"genetics",
"biology",
"genetics",
"and",
"genomics"
] | 2013 | Only One Isoform of Drosophila melanogaster CTP Synthase Forms the Cytoophidium |
Kaposi's sarcoma-associated herpesvirus ( KSHV ) is the infectious cause of several AIDS-related cancers , including the endothelial cell ( EC ) neoplasm Kaposi's sarcoma ( KS ) . KSHV-infected ECs secrete abundant host-derived pro-inflammatory molecules and angiogenic factors that contribute to tumorigenesis . The precise contributions of viral gene products to this secretory phenotype remain to be elucidated , but there is emerging evidence for post-transcriptional regulation . The Kaposin B ( KapB ) protein is thought to contribute to the secretory phenotype in infected cells by binding and activating the stress-responsive kinase MK2 , thereby selectively blocking decay of AU-rich mRNAs ( ARE-mRNAs ) encoding pro-inflammatory cytokines and angiogenic factors . Processing bodies ( PBs ) are cytoplasmic ribonucleoprotein foci in which ARE-mRNAs normally undergo rapid 5′ to 3′ decay . Here , we demonstrate that PB dispersion is a feature of latent KSHV infection , which is dependent on kaposin protein expression . KapB is sufficient to disperse PBs , and KapB-mediated ARE-mRNA stabilization could be partially reversed by treatments that restore PBs . Using a combination of genetic and chemical approaches we provide evidence that KapB-mediated PB dispersion is dependent on activation of a non-canonical Rho-GTPase signaling axis involving MK2 , hsp27 , p115RhoGEF and RhoA . PB dispersion in latently infected cells is likewise dependent on p115RhoGEF . In addition to PB dispersion , KapB-mediated RhoA activation in primary ECs caused actin stress fiber formation , increased cell motility and angiogenesis; these effects were dependent on the activity of the RhoA substrate kinases ROCK1/2 . By contrast , KapB-mediated PB dispersion occurred in a ROCK1/2-independent manner . Taken together , these observations position KapB as a key contributor to viral reprogramming of ECs , capable of eliciting many of the phenotypes characteristic of KS tumor cells , and strongly contributing to the post-transcriptional control of EC gene expression and secretion .
Kaposi's sarcoma-associated herpesvirus ( KSHV ) , a . k . a . human herpesvirus-8 ( HHV-8 ) is the infectious cause of Kaposi's sarcoma ( KS ) , the most common malignancy of untreated AIDS patients , and two rare lymphoproliferative disorders , multicentric Castleman's disease ( MCD ) and primary effusion lymphoma ( PEL ) [1]–[3] . Like all herpesviruses , KSHV establishes persistent , life-long infection of its human host . The primary proliferative elements of KS lesions are latently infected endothelial cells ( ECs ) with an abnormal spindle-shaped morphology , commonly known as ‘spindle cells’ . In latency , the viral episome persists in a reversible latent state and viral gene expression is limited to 6 consensus protein products ( LANA , v-cyclin , v-FLIP , Kaposins A , B , and C ) and 12 pre-miRNAs that are processed into at least 17 mature miRNAs ( reviewed in [4] , [5] , [6] ) . Spindle cells display actin cytoskeleton rearrangements , enhanced cell motility and an aberrant angiogenic phenotype ( recently reviewed in [7] , [8] ) ; all of these features can be recapitulated during in vitro infection of primary ECs [8]–[11] . Several KSHV latent gene products have been shown to contribute to these dramatic alterations in EC physiology ( reviewed in [8] ) , but our understanding of their relative contribution to tumor-initiating events remains incomplete . During KSHV infection , a complex translational program involving translation initiation at non-canonical CUG codons and decoding sets of GC-rich repeats results in the generation of multiple kaposin protein products , including Kaposin B ( KapB ) . We have previously shown that KapB regulates the expression of pathogenetically important pro-inflammatory cytokines and angiogenic factors at the post-transcriptional level [12] . This is achieved by direct binding and activation of mitogen-activated protein kinase ( MAPK ) -associated protein kinase-2 ( MK2 ) , a nodal kinase that regulates the turnover of mRNAs bearing AU-rich instability elements ( AREs ) within their 3′UTRs [13] , [14] . AREs are commonly found in labile mRNAs encoding tightly regulated , potent effector molecules , including many cytokines , angiogenic factors and proto-oncogenes [15] , [16] . MK2 phosphorylates a variety of target proteins , including several ARE-binding proteins ( ARE-BPs ) that govern ARE-mRNA stability , with the net effect of causing ARE-mRNA stabilization . For example , the ARE-BP tristetraprolin ( TTP ) normally promotes ARE-mRNA turnover by facilitating interactions between bound mRNAs and the cytoplasmic mRNA degrading enzymes associated within exosomes and processing bodies ( p-bodies , PBs ) ( reviewed in [17] ) . Phosphorylation of TTP by MK2 creates a binding site for 14-3-3 scaffolding proteins , thereby preventing association of ARE-mRNAs with the decay machinery . Accordingly , MK2 activation during latent KSHV infection or in response to ectopic expression of KapB coincided with dramatic stabilization of ARE-mRNAs and increased production of a number of canonical products of ARE-mRNAs , including IL6 and CSF2 [12] . Importantly , KapB also stabilized the ARE-mRNA encoding PROX1 [18] , a master regulator of lymphatic reprogramming of vascular endothelial cells , thereby providing a molecular mechanism for KSHV-mediated cell fate reprogramming [1] . Altogether , these findings suggest that KapB makes key contributions to the reprogramming of ECs in KS lesions . PBs are small ribonucleoprotein ( RNP ) granules that contain the requisite enzymes to mediate rapid mRNA deadenylation , decapping and exonucleolytic degradation in a 5′ to 3′ direction [19]–[23] . PBs also contain RNA induced silencing complexes ( RISC ) , translational repressors ( rck/p54 ) and many RNA-binding proteins such as the ARE-BP , TTP . PBs are constitutively present in most cells , but they are also dynamic structures; PB number and size increase in response to a variety of environmental stresses , when 5′ to 3′-exonucleolytic decay is blocked , or when translation is inhibited [24] . PB formation involves aggregation of RNA binding proteins , utilizing mRNA itself as an organizing scaffold; as such , PBs disperse after treatment of cells with RNase [25] . Not exclusively sites of ARE-RNA decay , PBs have been shown to have important roles in nonsense-mediated decay , RNA interference , and they can also harbor translationally-silenced mRNAs that have the potential to escape PBs and resume translation [26] , [27] . PBs have intimate links to both the microtubule and the actin cytoskeleton ( reviewed in [28] ) . Stationary PBs associate with actin bundles while mobile PBs connect to the microtubule network [29] . They are linked to microtubules via the microtubule-associated protein nesprin-1 and travel along microtubules using the retrograde motor protein dynein [30] , [31] . More recently , it was shown that PB accretion and ARE-mRNA turnover was modulated by the cytoskeletal regulator , RhoA GTPase ( RhoA ) [32] . Relatively little is known about how RhoA and other upstream signaling proteins govern PB assembly and function . The Rho family of small GTPases ( Rho , Rac and Cdc42 ) are molecular switches that cycle between an inactive GDP-bound configuration and an active GTP-bound form with the aid of guanine nucleotide exchange factors ( GEFs ) ( reviewed in [33] , [34] ) . RhoA regulates actin cytoskeleton dynamics to facilitate normal cell attachment , the formation of actin stress fibers , cell migration and angiogenesis ( reviewed in [35]–[37] ) . Inactive cytosolic RhoA translocates to membranes upon activation by G-protein coupled receptors , that link the G-protein Gα13 to RhoA activation via numerous GEFs including p115RhoGEF [34] , [38]–[40] . There , in its active conformation , RhoA can bind numerous downstream effectors . The most extensively studied RhoA effector is the Rho-associated kinase ROCK , a serine/threonine kinase with two isoforms , ROCK1 and ROCK2 , bearing 64% overall sequence identity , and many overlapping activities [41] . Upon binding to RhoA , ROCK1/2 promote actomyosin contractility and stress fiber formation by phosphorylating target proteins including focal adhesion kinase ( FAK ) , LIM kinase 1 ( LIMK1 ) , myosin light chain ( MLC ) and myosin phosphatase target ( MYPT-1 ) [42]–[45] . The effects of MK2 are not limited to regulation of ARE-mRNA turnover; numerous studies have pinpointed a role for MK2 in the actin cytoskeletal remodeling required to promote EC migration and angiogenesis [46]–[51] [52] , [53] . Active MK2 phosphorylates the small heat shock protein ( hsp ) 27 and suppresses its actin filament capping activity , releasing the molecule from the barbed end of the actin fiber to permit actin fiber growth [53]–[55] . MK2 also phosphorylates the serine/threonine kinase , LIM kinase 1 ( LIMK1 ) [56] , which subsequently phosphorylates and inactivates the actin-severing protein cofilin [45] . Thus , the available data indicates that both MK2 and RhoA support EC cytoskeletal rearrangements , migration and angiogenesis . However , to date , little is known about how these two pathways are functionally integrated . By studying KSHV latency , we have elucidated the functional integration of the MK2 and RhoA signaling pathways in ECs . We show that Kaposin B activates a signaling axis involving MK2 , hsp27 , p115RhoGEF and RhoA . The consequences of activation of this pathway in ECs include the formation of actin stress fibers , increased cell migration and angiogenesis , and dispersal of PBs . Because PBs are important sites of ARE-mRNA decay , KapB-mediated PB dispersal supports its role in potent stabilization of ARE-mRNAs . Taken together , these observations position KapB as a key contributor to viral reprogramming of ECs , capable of eliciting many of the phenotypes characteristic of KS tumor cells , and strongly contributing to the post-transcriptional control of EC gene expression and secretion .
Latent KSHV-infected ECs display marked alterations in cytoskeletal morphology . The latent vFLIP protein has been shown to modulate actin and cause spindling of ECs in an NF-kB-dependent manner [9] , but the impact of the remaining latent gene products on the cytoskeleton remains largely unexplored . We previously reported that the latent KapB protein binds and activates MK2 kinase , known to be a major regulator of actin remodeling [12] . For this reason , we investigated the ability of KapB to modulate the actin cytoskeleton in human umbilical vein endothelial cells ( HUVECs ) . Cells ectopically expressing KapB displayed thick parallel actin stress fibers ( Fig . 1B ) . Actin stress fibers have been observed in cells following activation of the small GTPase RhoA [57] , [58] . They have also been observed in cells with increased activation of the p38/MK2 signaling pathway or by expression of the constitutively active form of the kinase , MK2 ( MK2-EE ) or a phosphomimicking form of hsp27 ( hsp27-DDD ) [59] [60] , [61] . Consistent with this , we observed that expression of either MK2-EE or hsp27DDD in HUVECs caused the formation of actin stress fibers ( Figs . 1C , 1D ) . Activation of the p38/MK2 pathway by KapB also causes the increased secretion of pro-inflammatory cytokines that can then act in an autocrine or paracrine fashion to potentiate p38 pathway activation [12] . To address whether the appearance of actin stress fibers in response to KapB expression was mediated largely by KapB-mediated secretion of inflammatory molecules ( ie . paracrine effects ) , we treated HUVECs with conditioned media from KapB-expressing cells . No stress fibers were observed in control cells treated with media from KapB-expressing cells ( S1 Fig . ) , indicating that KapB-mediated rearrangement of the actin cytoskeleton was a cell autonomous effect . Selective chemical inhibitors were used to investigate the mechanism of KapB-mediated actin rearrangements . Treatment of KapB-expressing HUVECs with a selective MK2 inhibitor [62] prevented formation of actin stress fibers , whereas p38 MAPK inhibition had no effect ( Fig . 2A ) . This data is consistent with the notion that KapB binds and activates MK2 downstream of p38 MAPK . Treatment of cells with a selective inhibitor of the Rho kinases ROCK1 and ROCK2 ( hereafter referred to as ROCK ) , which are RhoA substrates known to play a role in actin stress fiber formation , also inhibited KapB-mediated stress fibers ( Fig . 2A ) . Taken together , these data suggest that KapB-mediated actin rearrangements depend upon activity of both MK2/hsp27 and RhoA/ROCK signaling axes . Because previous studies suggested a functional link between MK2/hsp27 and RhoA activation [63] , we measured RhoA activity in KapB-expressing cells using a pull-down assay that isolates only the active ( GTP-bound ) form of RhoA [64] . KapB expression activated RhoA , both in the absence ( Fig . 2B , lane 4 ) and presence ( Fig . 2B , lane 2 ) of the canonical RhoA activator , LPA [65] . Interestingly , we also observed increased pull-down of the active form of Rho from cells transfected with MK2-EE and hsp27-DDD ( Fig . 2B , lanes 5 and 6 ) . Thus , RhoA was activated in cells where MK2 activity was mimicked or stimulated by direct KapB-MK2 binding . This is consistent with previous reports of a non-canonical MK2/hsp27/p115RhoGEF/RhoA signaling pathway in arachadonic acid-treated epithelial cells ( 63 ) ( Fig . 3 ) . In addition to spurring actin stress fiber formation , activation of Rho family GTPases and the p38/MK2/hsp27 MAPK pathway has previously been linked to increased cell migration , and in the case of ECs , increased angiogenesis [33] , [49]–[51] , [56] , [60] . We observed that ectopic KapB expression in HUVECs promoted cell migration in a wound-healing assay; KapB-expressing cells displayed 59% wound closure compared to 25% wound closure by control cells over a 6-hour period ( S2A Fig . ) . KapB also promoted migration of HUVECs across a gelatin-coated semi-permeable membrane ( S2B Fig . ) . Interestingly , KapB-mediated enhancement of cell migration was detectable only in the absence of the potent endothelial angiogenic molecule , vascular endothelial growth factor ( VEGF ) , though VEGF treatment has no appreciable effect on KapB expression level ( S3 Fig . ) . The p38/MK2/hsp27 pathway plays a clearly defined role mediating the migration of ECs in response to VEGF [60]; however , our results suggest that the activation of this pathway by KapB also mediates EC migration when VEGF levels are low . This supports the notion that KapB targets nodal kinases commonly stimulated during EC migration . Endothelial cells form tubules on matrigel in an in vitro angiogenesis assay that mimics the formation of blood vessels [56] , [66] . We examined the effect of KapB expression on tubule formation compared to control HUVECs expressing either empty vector or the constitutively active from of MK2 ( MK2-EE ) . Both KapB- and MK2-EE-expressing HUVECs formed tubules in matrigel ( S4 Fig . ) . This is consistent with the well-described role of the p38/MK2 pathway in promoting EC tubule formation [50] , [56] . In both cases , tubule network formation was reduced after treatment with the ROCK inhibitor Y-27632 ( S4B Fig . ) . Thus , by activating MK2 and RhoA , KapB deregulates several processes in primary ECs contributing to a migratory and angiogenic phenotype . KapB stabilizes labile host cell ARE-mRNAs . Interestingly , a known site of ARE-mRNA translational repression and degradation is the processing body ( PB ) , which has intimate links to both the actin and the microtubule cytoskeleton ( reviewed in [28] ) ; stationary PBs associate with actin bundles while mobile PBs connect to the microtubule network [29] . More recently , it was shown that RhoA GTPase activity modulates PB formation [32] . Because KapB activates RhoA we reasoned that PB disruption might contribute to KapB-mediated ARE-mRNA stabilization . PBs were visualized by immunofluorescent staining for two PB resident proteins , hedls or DDX6 [67] . Indeed , KapB-expressing HUVECs displayed a decrease in the number of cells containing PBs of expected dimensions ( approximately 0 . 3 µm in diameter , [68] ) ( Fig . 4A–B ) . This effect was quantified by counting the number of KapB-expressing cells that contained one or more PBs of normal size compared to control ( empty vector ) transduced cells . We observed that in control cell populations , 64% of cells contained normal PBs , compared to 30% in KapB-expressing cells . HUVECs were also stained for the actin cytoskeleton , and KapB-expressing cells displayed thick parallel actin stress fibers ( Fig . 4B ) , consistent with our previous observations that KapB induces actin polymerization and RhoA activation ( Figs . 1–2 ) . It is not known how RhoA activation modulates PBs . RhoA GTPase can be activated by the microtubule-bound guanine exchange factor GEF-H1; microtubule disruption causes release of GEF-H1 and concomitant RhoA activation [69] . To test whether KapB-mediated activation of RhoA and disruption of PB accretion were related to a disruption of the microtubular network , we stained KapB-expressing HUVECs for α-tubulin . We did not observe altered tubulin staining intensity nor did we observe any striking differences in the appearance of the microtubule network in cells expressing KapB compared to controls ( Fig . 4C ) , indicating that KapB-mediated suppression of PBs is independent of changes to microtubule cytoskeleton , and suggesting that RhoA activation is unlikely a result of the release of GEF-H1 from microtubules . When Takahashi et al . [32] observed that the overexpression of RhoA mediated an alteration to PB dynamics , they concluded that RhoA induced an increase in the number of PBs while causing a marked reduction in average PB size . Using retroviral transduction , we expressed a constitutively active ( CA ) version or dominant negative ( DN ) version of RhoA fused to GFP ( Rho-CA-eGFP or Rho-DN-eGFP ) in primary ECs [37] , [70] . We observed both a marked loss of total PBs and a reduction in remaining PB size in HUVECs expressing Rho-CA-eGFP , whereas PBs in cells expressing Rho-DN-eGFP were similar to eGFP control ( Fig . 5A ) . This is consistent with our previous observations that chemical activators of RhoA , including LPA and nocodazole , disrupted PB accretion [67] . Cells expressing CA-RhoA-eGFP also display marked actin stress fiber formation that is lacking in control cells or cells expressing the Rho-dominant negative ( DN ) construct [37] , [70] ( Fig . 5A ) . Irreversible inactivation of RhoA by C3 transferase , which ADP-ribosylates the Asn41 residue , disrupts RhoA-mediated re-organization of actin filaments [38] , [71] . To determine whether RhoA activity is required for KapB-mediated actin polymerization and disruption of PBs , we treated KapB-expressing HUVECs with C3 . In the absence of C3 , the number of cells containing normal PBs was reduced two-fold by KapB expression; when treated with C3 , this number was restored to control levels ( Fig . 5B , D ) . C3 also abrogated the ability of KapB to promote actin stress fibers . We also co-expressed KapB with RhoA-DN by sequential transduction of HUVECs . Like C3 treatment , Rho-DN expression also prevented the ability of KapB to disperse PBs ( Fig . 5E ) . However , when KapB-expressing HUVECs were treated with ROCK inhibitor there was no effect on KapB-mediated PB disruption ( Fig . 5D ) . By contrast , ROCK inhibition prevented KapB-mediated stress fiber formation ( Fig . 2 ) . This observation uncouples the effect that RhoA activation has on PBs from canonical RhoA effects on actin and migration - processes that clearly require ROCK . Taken together , these experiments show that KapB-mediated PB disruption is dependent on RhoA , but not ROCK , consistent with previous findings [32] ( Fig . 3 ) . We hypothesized that KapB expression activates a little-known signaling axis that links upstream activation of the p38/MK2 MAPK pathway to RhoA activity ( Fig . 3 ) ( 63 ) . To examine the role of MK2 in KapB-mediated RhoA activation , we generated retroviruses expressing short hairpin ( sh ) -RNAs directed against MK2 . Efficient silencing of endogenous MK2 expression ( Fig . 6A ) prevented RhoA activation in KapB-expressing HUVECs ( Fig . 6B ) . Furthermore , MK2 silencing prevented efficient KapB-mediated PB dispersion ( Figs . 6C–D ) . These data confirm a role for MK2 in the mechanism of KapB-mediated RhoA activation and PB dispersion . McCormick and Ganem [12] previously demonstrated that KapB expression caused a dramatic stabilization of AU-rich element ( ARE ) -mRNAs . To examine the role of RhoA in ARE-mRNA turnover , we utilized a reporter assay developed in our lab and described in detail in [72] . Briefly , HeLa Tet-Off cells were co-transfected with an empty plasmid vector or KapB expression vector , along with a doxycycline ( dox ) -responsive reporter plasmid encoding firefly luciferase linked to a canonical ARE derived from the labile CSF2 transcript . After 24 hours , transcription was arrested by the addition of dox , and 24 hours later , lysates were harvested for luciferase assays . Co-transfected dox-responsive Renilla luciferase reporter lacking an ARE served as a normalization control . KapB expression caused a striking increase in normalized luciferase activity , as did expression of MK2-EE , hsp27-DDD and RhoA-CA ( Fig . 7A ) . These results are consistent with previous reports of control of ARE-mRNA turnover by RhoA [32] and p38/MK2 signaling pathways [13] , [14] , [73] . To further elucidate the role of RhoA in ARE-mRNA turnover , dominant negative RhoA ( RhoA-DN ) was introduced into this system . When KapB , MK2-EE and hsp27-DDD were co-expressed with RhoA-DN , the normalized luciferase activity was markedly reduced , indicating reduced stability of the firefly luciferase transcript ( Fig . 7B ) . Phosphorylation of MK2 substrate hsp27 in cells expressing KapB or MK2-EE was confirmed by immunoblotting ( Fig . 7C ) . These data suggest that KapB requires RhoA activation in order to achieve maximal ARE-RNA stabilization; furthermore , they confirm that MK2 activation precedes RhoA activation ( Fig . 3 ) . Thus , KapB causes ARE-mRNA stabilization via the direct binding and activation of MK2 which in turn causes RhoA activation . Previous work demonstrated that p38/MK2-mediated RhoA activation depends on the phosphorylation of hsp27 on serines 15 , 78 and 82 , and the formation of a complex comprising phosphorylated hsp27 ( p-hsp27 ) , RhoA and the guanine exchange factor ( GEF ) , p115RhoGEF [63] . Therefore , we hypothesized that KapB-mediated RhoA activation may require p-hsp27 and p115RhoGEF . To test this , we undertook a series of experiments that examined the importance of hsp27 , p115RhoGEF and RhoA in modulating PB and actin dynamics in response to upstream activators . Expression of MK2-EE and hsp27-DDD in HUVECs promoted the disruption of PBs ( Fig . 8A–C ) . Consistent with our hypothesis , we found that inhibition of RhoA using C3 transferase or co-expression of RhoA-DN restored PB levels to that of controls in MK2-EE-expressing HUVECs ( Fig . 8A , D , E ) . However , when RhoA was inhibited ( by either C3 or RhoA-DN ) in hsp27-DDD-expressing HUVECs , PB levels were not restored ( Fig . 8B–E ) . Treatment of MK2-EE- and hsp27-DDD-expressing HUVECs with the ROCK inhibitor had no effect on PB disruption ( Fig . 8D ) , as previously observed for KapB ( Fig . 5D ) . We reasoned that KapB and MK2 activate RhoA by mediating complex formation between p115RhoGEF , RhoA and p-hsp27 ( Fig . 3 ) ( 63 ) . To investigate the role of p-hsp27 , we utilized a dominant negative form of hsp27 ( hsp27-AAA ) in which the three canonical phosphorylation sites ( serines 15 , 78 , 82 ) had been mutated to alanines [74] . Hsp27-AAA was co-expressed with KapB , MK2-EE and hsp27-DDD in HUVECs . In all cases , expression of the hsp27-DN prevented PB disruption , restoring PBs to normal levels ( Fig . 9A–C ) . These data support an important role for hsp27 phosphorylation in KapB-mediated activation of RhoA and resulting PB dispersion . Rho GTPase activity is regulated by numerous GEFs , which are in turn are regulated by upstream signals including G-protein activation ( in the case of G-protein coupled receptors ) , phosphorylation ( such as by receptor tyrosine kinases ) or , as described by Garcia et al . [63] , complex formation with phosphorylated hsp27 . To confirm a role for the p115RhoGEF in our model of KapB-induced RhoA activation and PB disruption ( Fig . 3 ) , we generated lentiviruses expressing short hairpin ( sh ) -RNAs directed against p115RhoGEF and another RhoA GEF , namely GEF-H1 . H1 is bound to microtubules and is released from upon their disruption ( e . g . with nocodozole ) to mediate activation of RhoA . When HUVECs were transduced with these lentiviral vectors , they expressed eGFP and displayed reduced expression of target genes ( S5 Fig . ) . HUVECs expressing three different shRNA constructs targeting p115RhoGEF were unable to mediate PB disruption in cells expressing KapB , MK2-EE or hsp27-DDD ( Fig . 10 , S6 Fig . ) . By contrast , silencing of microtubule-bound GEF-H1 had no appreciable effect on PBs ( Fig . 10 , S6 Fig . ) . This is consistent with our observation that microtubules are not disrupted in response to KapB expression ( Fig . 4 ) . Together , these data indicate that p115RhoGEF is essential for KapB-mediated RhoA activation and PB disruption . HUVECs were infected with KSHV and establishment of latency was confirmed by LANA immunostaining ( Fig . 11A ) . KapB expression during latent infection of HUVECs was confirmed by immunoblotting ( S3 Fig . ) . LANA-positive cells displayed a marked reduction in the number of cells with normal-sized PBs at 24 , 48 and 72 hours post-infection ( Fig . 11 A , B ) . To implicate kaposin gene products in PB disruption , we transduced cells with shRNAs targeting the kaposin transcript ( shKAP1 , shKAP2 ) . These shRNAs would be expected to silence expression of kaposin gene products ( translated from spliced , cytoplasmic kaposin mRNA ) , but have no effect on Drosha-dependent processing of kaposin transcript-derived miRNAs in the nucleus . KSHV infection of cells bearing kaposin shRNAs revealed either a partial or full restoration in PB levels to that observed in uninfected cells ( shKAP1 and shKAP2 , respectively , Fig . 11C , D ) . These data indicate that at least one of the kaposin proteins is required for KSHV to alter PB dynamics during latent infection . Moreover , in parallel shRNA knockdown experiments we demonstrated that p115RhoGEF is essential for PB disruption in latently infected ECs , whereas GEF-H1 was dispensable ( Fig . 12 ) . Taken together , these findings indicate that during KSHV latency a product of the kaposin locus , likely KapB , activates the MK2-hsp27-p115RhoGEF-RhoA signaling pathway , thereby disrupting PBs .
Latent KSHV infection of primary ECs in vitro causes dramatic changes in cellular physiology that largely reflect observations of KS tumor cells . Infected cells display marked alterations in signal transduction and gene expression , extended life span , and enhanced motility and angiogenic properties . Despite intensive efforts , the precise contributions of individual viral gene products to alterations in EC physiology remain incompletely understood . The data presented in the current study position KapB as a key contributor to viral reprogramming of ECs; ectopic expression of KapB leads to actin stress fiber formation and altered cell morphology , increased motility and an angiogenic phenotype; all of which are characteristic of the KS tumor cells ( see model , Fig . 13 ) . Moreover , KapB was sufficient to disrupt PBs , sites of mRNA translational repression and decay . These diverse phenotypes are linked to a signaling axis , comprising MK2 , hsp27 , p115RhoGEF and RhoA ( Fig . 3 ) , which likely evolved to respond to acute , transient stress and promote cell survival . By encoding the KapB protein that directly binds to the nodal kinase MK2 , KSHV achieves constitutive activation of this pathway . Several viruses have been shown to modulate PBs during infection , but the functional relevance of these observations are not yet clear ( reviewed in [75] ) . Many viral RNAs associate with PB resident proteins [76] [77] , and there have been reports of viral RNA recruitment to PBs [78] . Thus , PB dispersion might be an attractive mechanism for viral evasion of translational repression and RNA decay . Furthermore , some viruses have been shown to co-opt key PB resident proteins to support viral replication . For example , the flaviviruses West Nile virus and hepatitis C virus recruit DDX6 and Lsm1 to viral replication centers [79] , [80] [81] , and in so doing cause PB dispersion . By contrast , little is known about the impact of PBs on herpesvirus infection . Human cytomegalovirus has recently been shown to increase PBs in infected cells using a mechanism that requires active synthesis of cellular mRNA [82] . However , disruption of PB accretion by selective knock-down of several PB component proteins had no effect on virus replication in primary fibroblasts , making is unclear how enhanced PB formation benefits viral fitness . We previously reported PB disruption during lytic KSHV infection , and demonstrated that the lytic viral gene vGPCR was sufficient to disrupt PBs in an ectopic expression model [67] . We now show that PB dispersal is also a cell autonomous feature of KSHV latency , and elucidate the molecular mechanism for KapB-mediated PB disruption . A common feature of these studies was strong correlation between PB dispersal and stabilization of endogenous [67] and exogenous ( [67] and Fig . 7 ) ARE-containing RNAs , and resulting increases in protein production of the encoded products . Thus , multiple KSHV gene products converge on the regulation of ARE-mRNA turnover , providing an attractive mechanism for the overproduction of pro-inflammatory cytokines and angiogenic factors characteristic of KS lesions . Our study reveals a previously unappreciated mechanism for the control of PB formation , involving MK2-mediated activation of RhoA . Our data fits well with a proposed model by Garcia et al . who observed that arachadonic acid-mediated activation of RhoA depended on prior activation of p38/MK2 and phosphorylation of the MK2 substrate hsp27 [63] . Hsp27 phosphorylation is an important ‘switch’ that allows local recruitment of p115RhoGEF and RhoA activation . In our experiments , we investigated the role of each of these signaling molecules in KapB-mediated PB dispersion . RhoA inhibition , either using the irreversible C3 transferase or expression of a dominant negative construct , blocked PB dispersion ( Figs . 5 , 8 ) , as did the expression of a dominant negative version of hsp27 ( Fig . 9 ) . p115RhoGEF silencing using three different shRNA constructs also restored PB numbers to control levels in cells ectopically expressing KapB , MK2-EE , or hsp27-DDD or in latently KSHV-infected HUVECs ( Figs . 10 , 12 , and S6 Fig . ) . Ours is the first study to link MK2 to regulation of PBs , and confirm a previously identified role for RhoA [32] . However , the precise mechanism of PB dispersion downstream of RhoA activation remains to elucidated . Active RhoA signals through numerous effectors , the most extensively studied of which are the Rho-associated kinases ( ROCKs ) , ROCK1 and ROCK2; binding of active RhoA relieves ROCK autoinhibition . PB disruption by KapB is insensitive to treatment with a ROCK inhibitor , suggesting that this process requires RhoA but not its downstream effector ROCK kinases ( Figs . 5 , 8 ) . Overexpression of ROCK and another RhoA effector , the formin-family member mDia , likewise did not alter PBs [32] . PBs associate and traffic along microtubules [29] , [83] and mDia activates microtubule polymerization and cause microtubule bundling [84] , [85] . In isolation , these observations would make mDia an attractive candidate effector for PB dispersion . However , because the microtubule cytoskeleton is unchanged in KapB-expressing cells ( Fig . 4 ) , this suggests that KapB does not activate mDia . Moreover , shRNA knockdown of the microtubule-bound GEF-H1 had no effect on PB dispersion during latent KSHV infection or in cells ectopically expressing KapB ( Figs . 10 , 12 ) , also suggesting that microtubules and mDia do not play a role in the mechanism of KapB-mediated PB dispersion . Future elucidation of the mechanism of MK2/hsp27/p115RhoGEF/RhoA-dependent PB dispersal will be challenging , but may be accelerated by careful inspection of PB resident proteins that might either be recruited by active RhoA , or phosphorylated by MK2 . Latent KSHV infection reprograms gene expression in ECs at transcriptional and post-transcriptional levels , but relative contributions of viral gene products to reprogramming remains incompletely understood . Our studies position Kaposin B as a chief post-transcriptional regulator of gene expression , binding and activating the nodal kinase MK2 , thereby stabilizing and enhancing the translation of a variety of ARE-mRNAs encoding pathogenetically important pro-inflammatory cytokines and angiogenic factors . The effects of KapB on EC physiology are striking; formation of actin stress fibers , accelerated cell migration , and a strong angiogenic phenotype . Furthermore , KapB-mediated ARE-mRNA stabilization coincided with dispersal of PBs , a major site of ARE-mRNA decay in mammalian cells . By studying KapB , we gained new insight into the fundamental regulation of these processes by a recently identified signaling axis involving MK2 , hsp27 , p115RhoGEF and RhoA . Stress fiber formation , cell migration and angiogenesis were dependent on the activity of the RhoA substrate kinase ROCK , whereas PB dispersion occurred in a ROCK-independent manner . Taken together , these observations suggest that KapB is a key contributor to viral reprogramming of ECs , capable of eliciting many of the phenotypes characteristic of KS tumor cells , and strongly contributing to the post-transcriptional control of EC gene expression and secretion .
Doxycycline ( dox ) , blasticidin , puromycin , valproic acid , human AB serum , polyethyleneimine ( PEI ) and polybrene were purchased from Sigma-Aldrich Canada . C3 transferase ( Rho inhibitor I ) was from Cytoskeleton , Inc . MK2 inhibitor-III was purchased from Calbiochem . ROCK inhibitor Y-27632 was from Sigma . To irreversibly inhibit RhoA-GTPase , HUVECs were treated with 1 µg/ml C3 transferase for 6 h in SF medium . Serum-free conditions are important to minimize baseline RhoA activity in the context of C3 exposure . However , since RhoA inhibition by C3 is irreversible after treatment , C3 and serum-free control cells were incubated for 1 h in normal medium to restore baseline PB levels . HeLa Tet-Off ( Clontech ) , Phoenix-Amphotropic ( a kind gift from G . Nolan , Stanford ) , and HEK293T cells ( ATCC ) were maintained at 37°C in a 5% CO2 atmosphere in Dulbecco's modified Eagle's medium containing 100 U of penicillin and streptomycin per ml and 10% heat-inactivated fetal bovine serum . Primary human umbilical vein endothelial cells ( HUVECs ) were purchased from Lonza . Cultures were expanded in EGM-2 medium ( Lonza ) on tissue culture plates coated with 0 . 1% ( wt/vol ) gelatin ( in phosphate-buffered saline [PBS] ) and used between passages 5 and 7 for experiments . The BCBL-1 primary effusion lymphoma ( PEL ) cell line was cultured in RPMI medium containing 10% heat-inactivated fetal bovine serum and 55 µM ß-mercaptoethanol . pcDNA3-MK2EE was a kind gift from Paul Anderson ( Harvard University ) , and its creation is described in [86] . Briefly , to create pcDNA3-Flag-MK2-EE the cDNA encoding constitutively active murine MK2 was amplified with primers G42/G43 from pcDNA3mycMK2T205E/T317E [13] . The amplicons were digested with BamHI and XhoI , and inserted into the BamHI and XhoI sites of pcDNA3-Flag-BAK . An expression vector encoding the phosphomimicking version of small heat shock protein 27 ( hsp27 ) in which serine residues at positions 15 , 78 and 82 have been substituted with aspartic acid ( pcDNA3 . 1-HAhsp27DDD ) was generously obtained from Matthias Gaestel and its creation is described [54] . To generate the pcDNA3 . 1 HA-hsp27-AAA clone , two sequential reactions of Phusion site directed mutagenesis was performed on pcDNA3 . 1-HA-hsp27-DDD according to the instructions of the manufacturer ( NEB ) and using the following primers: Reaction 1 . D15A forward 5′- GGCCCCGCCTGGGACCCC-3′ , D15A reverse 5′- CCGCAGGAGCGAGAAGGGG-3′; Reaction 2 . D78AD82A forward 5′- ACTCGCCAGCGGGGTCTCG-3′ , D78AD82A reverse 5′- TGCCGGGCGAGCGCGCGG-3′ . The expression vector for KapB ( pCR3 . 1-kapB ) and the ARE-RNA reporter plasmids ( pTRE2-Rluc , pTRE2-Fluc-ARE , pTRE2-BBB , pTRE2-BBB-ARE , pTRE2-d1EGFP , and pTRE2-d1EGFP-ARE ) have been previously described [12] , [67] , [72] . pCB6-eGFP-RhoA-CA and pCB6-eGFP-RhoA-DN plasmids were a generous gift from Dr . Roy Duncan , Dalhousie University . To create the pBMN-GFP-IP plasmid , the pEGFP-N1 plasmid ( Clontech ) was digested with NotI , subjected to a standard fill-in reaction with Klenow DNA polymerase ( NEB ) , and further digested with BglII , releasing the eGFP ORF . To prepare the recipient pBMN-IP vector ( G . Nolan lab , Stanford U . ) , XhoI digestion was performed , followed by a Klenow fill-in reaction , and a BamHI digest . These fragments were ligated to create pBMN-GFP-IP , which permits the expression of GFP and the puromycin resistance gene from a single bicistronic mRNA . pBMN-kapB-IP was generated by BamHI/EcoRI digestion of pCR3 . 1-kapB , releasing the 636 bp KapB ORF ( derived from a pulmonary KS isolate ) . This fragment was subsequently ligated into the BamHI/EcoRI digested pBMN-IP vector . To create pBMN-MK2EE-IP , pcDNA3-Flag-MK2EE was amplified with primers G42/G43 from pcDNA3mycMK2T205ET317E [13] . The amplicons were digested with BamHI and XhoI , and inserted into BamHI/XhoI digested pBMN-IP vector . To create pBMN-hsp27DDD-IP , HA-hsp27DDD was excised from pcDNA3 . 1-HAhsp27DDD by sequential XbaI digest , Klenow fill-in to create a blunt end , and EcoRI digest . Following this , the insert was ligated into pBMN-IP , which had been prepared by sequential XhoI digest , Klenow fill-in to create a blunt end , and EcoRI digest . To generate pBMN-IP or pBMN-IB vectors containing RhoA-CA-eGFP and RhoA-DN-eGFP , the ORFs from pCB6-eGFP-RhoA-CA and pCB6-eGFP-RhoA-DN were amplified using Phusion High Fidelity Polymerase ( NEB ) chain reaction and the following primers: eGFP-N forward ( 5′-in database-3′ ) and reverse 5′-ATGCGAATTCTTATTATTACAAGACAAGGCACCCAGATT-3′ . The resulting products were digested with BamHI and EcoRI and ligated into the pBMN-IP or pBMN-IB vector backbone . To generate HA-tagged versions of these clones , the ORFs of pCB6-eGFP-RhoA-CA and pCB6-eGFP-RhoA-DN were amplified using Phusion High Fidelity Polymerase ( NEB ) chain reaction and the following primers: forward 5′- GCATGGATCCACCATGGAGTACCCATACGATGTTCCAGATTACGCTCCCAGAGCTGCCATCCGGAAGAAAC-3′ and reverse 5′-ATGCGAATTCTTATTATTACAAGACAAGGCACCCAGATT-3′ . The resulting PCR products were digested with BamHI and EcoRI and ligated into the pBMN-IP or pBMN-IB vector backbone . To create pBMN-IB-HA-Hsp27-AAA , the ORF of pCDNA3 . 1-HA-HSP27-AAA was amplified using Phusion High Fidelity Polymerase ( NEB ) chain reaction and the following primers: forward 5′- GGTGGAATTCATGGCTTACC-3′ and reverse 5′-ATGCCTCGAGTTATTATTACTTGGCGGCAGTCTCAT-3′ . The resulting PCR product was digested with EcoRI and XhoI and ligated into the pBMN-IB vector backbone . Retroviral shRNA expression vectors were created via PCR amplification of template oligonucleotides , and cloning into XhoI/EcoRI restriction sites in pSMP ( Open Biosystems ) . Briefly , the following 97-mer template oligonucleotides were synthesized , sh-scrambled: TGCTGTTGACAGTGAGCGAGCACAAGCTGGAGTACAACTATAGTGAAGCCACAGATGTATAGTTGTACTCCAGCTTGTGCCTGCCTACTGCCTCGGA , shMK2: TGCTGTTGACAGTGAGCGCGCCTGAGAATCTCTTATACACTAGTGAAGCCACAGATGTAGTGTATAAGAGATTCTCAGGCTTGCCTACTGCCTCGGA . These sequences were PCR amplified with Xho pSMP forward primer ( 5′-CAGAAGGCTCGAGAAGGTATATTGCTGTTGACAGTGAGCG-3′ ) and Eco pSMP reverse primer ( 5′-CTAAAGTAGCCCCTTGAATTCCGAGGCAGTAGGCA-3′ ) and Pfu Ultra High Fidelity DNA polymerase ( Stratagene ) . 110-bp amplicons were then digested with XhoI/EcoRI and ligated into XhoI/EcoRI digested pSMP retroviral vector . pMD . 2G ( envelope ) and pSPAX2 ( packaging ) plasmids were purchased from Addgene . All pGIPZ shRNA constructs used to knock down expression of the RhoA-specific guanine exchange factors ( GEFs ) listed below were purchased from Open Biosystems ( oligo ID in parentheses ) : p115-3 ( V3LHS_317458 ) , p115-4 ( V3LHS_317456 ) , p115-9 ( V2LHS_37090 ) , H1-1 ( V3LHS_317146 ) , H1-2 ( V3LHS_317143 ) , and H1-7 ( V2LHS_36680 ) . pIPZ was created by PCR amplification of the enhanced CMV promoter of pGIPZ using a forward XbaI primer and a reverse NotI primer . After digestion of both the PCR product and pGIPZ vector with XbaI and NotI , the product was ligated into the vector backbone to create pIPZ . This new vector lacks the ORF for turbo eGFP and places the CMV promoter proximal to the IRES . To create pIPZ shRNA constructs used to knock down expression of the kaposin gene , we used previously generated shRNAs against kaposin in the pSM2 vector . Two different shRNA sequences that target the kaposin ORF were used and are named according to the position of the starting nucleotide . The 22-mer sequences for KapB 692 and KapB 746 shRNAs are 5′-TGTCCCGGATGTGTTACTAAAT-3′ and 5′-ACTCGTTTGTCTGTTGGCGATT-3′ , respectively . In order to transfer these from pSM2 to pIPZ , the pSM2 vectors were digested with MluI and XhoI and inserted into pIPZ . Retrovirus stocks were produced in 15-cm cell culture dishes by PEI-mediated transfection using 54 µl of PEI and 18 µg of the gene/shRNA of interest , contained in either the pBMN-IP , pBMN-IB or pSMP vector backbone , into the Phoenix amphotropic packing cell line ( a kind gift from G . Nolan , Stanford ) . The transfection medium was replaced after 6 hours . Virus-containing supernatants were harvested 48 h after transfection . These supernatants were spinoculated onto target HUVEC cell monolayers for 2 h at 2 , 000 rpm in the presence of 5 µg/ml Polybrene ( Sigma ) , and after 24 h , 1 µg/ml puromycin or 10 µg/ml blasticidin was added to select for transductants . Lentivirus stocks were produced in 15-cm cell culture dishes by PEI-mediated co-transfection of three plasmids: 10 µg of the shRNA of interest ( in pGIPZ or pIPZ ) , 3 µg pMD . 2G ( envelope ) , 6 µg pSPAX2 ( packaging ) and 54 µl of PEI into HEK293T cells . Virus-containing supernatants were collected after 48 hours , diluted 1∶2 , and added to target HUVEC monolayers for 4-5 hours at 37°C in the presence of 5 µg/ml polybrene ( Sigma ) . After 24 h , 1 µg/ml puromycin was added to select for transductants . Wild-type KSHV virus was produced from lytic reactivation of the BCBL-1 PEL cell line . Briefly , KSHV was induced to lytically reactivate from BCBL-1 cells at a cell concentration of 2×105/ml using 0 . 3 mM valproic acid . After 7 days of induction , the suspension culture was pre-cleared by centrifugation for 10 minutes at 800×g before being filtered using 0 . 45 µm Millipore filters . The virus-containing filtrate was then centrifuged for 2 hours at 25 , 000×g and the supernatant discarded . The virus pellet was resuspended in 1/100 of the original culture volume of DMEM containing 10% FBS and stored at −80°C . Infectious viral titer was determined by immunofluorescent staining with anti-LANA ( see method below ) . For the experiments reported in this paper , a 1∶25 dilution of stock virus was spinoculated onto HUVECs for 45 minutes at 2000 x g and 30°C without the addition of polybrene . The viral inoculum was left on the cells for an additional hour at 37°C before being replaced with normal EGM-2 media according to the methods of [87] [88] . Latently infected cells were fixed at the indicated times post infection . The luciferase reporter assay for identification of modulators of ARE-mediated mRNA decay is described in detail in [72] . Briefly , 105 HeLa Tet-Off cells were cotransfected with 100 ng of a reporter plasmid master mix ( pTRE2-Fluc-ARE and pTRE-2-Rluc , at a ratio of 9∶1 ) ; 900 ng of an expression vector or an empty vector control; and 3 µl of Fugene HD ( Roche ) according to the instructions of the manufacturer . Twenty-four hours after transfection , Dox was added ( 1 µg/ml ) to stop de novo transcription from the pTRE reporter plasmids . Twenty-four hours after the addition of Dox , transfected cells were lysed in 200 µl of 1× passive lysis buffer , and samples were processed using the dual-luciferase assay kit ( Promega ) according to the instructions of the manufacturer . Firefly and Renilla luminescence was determined using the GloMax 20/20 luminometer ( Promega ) . Firefly luminescent signal , expressed in relative light units ( RLUs ) , was normalized to that of Renilla luciferase , to eliminate off-target effects of our expression plasmids or the transfection procedure . Results are expressed as normalized luciferase activity . Six-well plates of cells were washed once with phosphate-buffered saline and lysed directly in 1× sodium dodecyl sulfate ( SDS ) sample buffer . Equivalent amounts of protein ( 10 or 25 µg ) were subjected to SDS-polyacrylamide gel electrophoresis and transferred to nitrocellulose membranes ( Amersham ) . Membranes were blocked in Tris-buffered saline–Tween-20 ( TBST ) containing 5% bovine serum albumin unless otherwise indicated and probed overnight at 4°C using anti-KapB ( a generous gift from D . Ganem used at 1∶5000 in 5% milk ) , anti-phospho ( ser82 ) Hsp27 ( 1∶1000 ) , anti-p115 ( 1∶1000 ) , anti-H1 ( 1∶1000 ) , anti-GAPDH ( 1∶1000 , Abcam ) , anti-RhoA ( 1∶667 ) , anti-MK2 ( 1∶1000 ) or anti-β-actin ( 1∶2500 ) antibody . Horseradish peroxidase-conjugated goat anti-rabbit and anti-mouse immunoglobulin secondary antibodies were used at a 1∶2000 dilution . All antibodies were purchased from Cell Signaling Technologies unless otherwise indicated . Secondary antibody was detected using ECL Plus detection reagents ( Amersham Biosciences ) according to the manufacturer's instructions . The chemiluminescent signal was detected using the Kodak Image Station 4000 mm PRO with no excitation or emission filter . 6-well cluster dishes of HeLa-Tet Off cells were subjected to PEI-mediated transfection with 100 ng of a green fluorescent reporter protein , 1 . 0 µg of expression plasmids for KapB ( pcr3 . 1 KapB ) , constitutively active MK2 ( pcDNA3 FlagMK2-EE ) , phosphomimicking heat shock protein 27 ( pcDNA3 . 1 HA-hsp27-DDD ) or empty vector pcDNA3 . 1 and 3 . 3 µl of PEI per well . The transfection medium was replaced after 6 hours . 48 hours post transfection , 105 transfected cells were seeded at sub-confluent density in a 10cm tissue culture dish . 24 hours later , cells were treated with starvation medium containing 0 . 1% FBS for 24 hours . The next day , cells were starved in medium without FBS for an additional 3–4 hours before being treated or not treated with LPA for 3 minutes . Alternately , confluent 6-well plates of either KapB- or vector control-expressing HUVECs were starved in low-serum EBM-2 medium for 24 hours and in serum-free medium for 4 hours before use . Cells were then immediately washed in ice-cold PBS , and lysed in 500 ul of 1 x Lysis buffer containing aprotinin , pepstatin , leupeptin , and PMSF on ice following the instructions of the Active Rho Detection Kit ( Cell Signaling Technologies ) . Fresh lysates were clarified by centrifugation at 21 , 000 x g for 5 minutes at 4°C , kept cold at all times , and used immediately for active rho pull downs as recommended by the manufacturer . After binding for one hour , the unbound protein lysate was reserved and used for assay of total protein . For immunoblot analysis 10–20 µg of each total protein lysate and proportional amounts of Rho pull-down reaction ( 1∶10 ratio of lysate:pull-down ) were subjected to 12%- SDS-polyacrylamide gel electrophoresis and transferred to nitrocellulose membranes ( Amersham ) . Membranes were blocked Tris-buffered saline–Tween-20 ( TBST ) containing 5% bovine serum albumin and probed overnight at 4°C using anti-RhoA primary antibody ( 1∶667 , CST ) and horseradish peroxidase-conjugated goat anti-rabbit secondary antibody ( 1∶2000 ) and developed and imaged according to the above protocol . After transduction and selection , HUVECs were seeded on coverslips for microscopy . Twenty-four hours later , cells were either not treated or treated with the appropriate inhibitors or activators as described above and in the figure legends . After treatment , cells were fixed at room temperature in 4% paraformaldehyde ( in PBS ) for 10 min and then permeabilized with 0 . 1% Triton X-100 for 10 min . For staining microtubules , fixation was performed at 37°C in pre-warmed 4% paraformaldehyde ( in D-PBS ) at 37°C before permeabilization as indicated above . Cells were subsequently washed 3 times with PBS and then blocked in 1% human AB serum in PBS for 1 h at room temperature . To stain PB resident proteins , fixed cells were incubated with mouse anti-Hedls antibody ( 1∶1000; Santa Cruz ) or rabbit anti-DDX6 antibody ( C terminus , 1∶1000; Bethyl Laboratories ) in 1% human AB serum overnight at 4°C . When appropriate and according to figure legends , cells were also incubated overnight at 4°C with rabbit anti-HA ( 1∶1600 , CST ) , mouse anti-FLAG ( 1∶1600 , Sigma ) , rabbit anti-LANA ( 1∶1000 , a generous gift from D . Ganem ) , mouse anti-tubulin ( 1∶200 , Santa Cruz ) or rabbit anti-KapB antibody ( 1∶1000 , a generous gift from D . Ganem ) for 30 min at room temperature . Primary antibodies were removed by three 5-minute washes of PBS . Goat anti-rabbit Alexa 555 , chicken anti-mouse Alexa 488 , chicken anti-mouse Alexa 647 or goat anti-rabbit Alexa 647 secondary antibodies ( Molecular Probes ) were added for 1 h at room temperature in the dark . After being washed as described above , indicated cells were incubated with for one hour at room temperature with 1∶100 phalloidin in PBS ( conjugated to either Alexa 555 or 647; Molecular Probes ) . Finally , cells were washed 3 more times and mounted on microscope slides with ProLong Gold antifade mounting medium ( Invitrogen ) and visualized using a Zeiss LSM 510 META laser-scanning confocal microscope and the 40× or 63x objective . HUVECs transduced with vector were examined at 400x or 630x magnification , and the number of cells per field of view that contained normal-sized ( approximately 300 nm in diameter; see [67] ) PBs were counted . After counting between 100 and 200 cells ( usually 4 fields of view ) , the percentage of cells containing normal PBs was determined . The effect of viral infection or ectopic expression on PB accretion was determined by counting only those cells that were positive by immune fluorescence for infection/gene expression . Results are displayed as the average fold reduction in cells with PBs compared to that of the untreated vector control ( ±SE ) . HUVEC cell monolayers were grown on gelatin-coated coverslips that had been etched with a reference marker , transduced with either the KapB retrovirus or an empty vector control and selected with puromycin as described above . After incubating cells in medium devoid of serum or growth factors for 1 hour , cell monolayers were wounded with a p200 pipette tip ( by scraping off cells near to the reference marker ) , washed , and incubated in either complete media or media supplemented with VEGF ( 10 ng/ml ) . The ability of cells to repair the wound was monitored over time . Images at the time of wounding ( t = 0 ) or at six hours ( t = 6 hours ) were captured using an Olympus CKX41 Inverted microscope , and the surface area ( SA ) of the initial and the remaining wound was determined using Image J . The percent of wound closure after the six hours was calculated using the equation SA ( t = 0 ) –SA ( t = 6 ) /SA ( t = 0 ) ×100 . Each experiment was performed in duplicate and the results presented are one representative experiment of three . Cell migration was assayed using a modified Boyden chamber assay [61] . HUVECs , transduced to express either KapB or an empty vector control , were harvested with trypsin , counted , centrifuged and resuspended in supplement-free EBM-2 medium containing 0 . 1% FBS ( 0 . 1%-EBM-2 ) . 7 . 5×104 cells were added to each 8 . 0 um pore size gelatinized polycarbonate membrane ( Corning ) separating the two chambers of a 6 . 5 mm transwell . After one hour of adhesion , either 0 . 1%-EBM-2 alone or media containing VEGF ( 1 or 10 ng/ml ) was added to the lower chamber . After 4 hours , non-migratory cells remaining on the upper side of the membrane were removed by cotton swabbing and the cells on the underside of the membrane were fixed with 4% paraformaldehyde before staining with 0 . 2% crystal violet for 1–2 hours . The filters were dried , removed from the chambers , and visualized by microscopic examination . The number of migrated cells on the lower face of the filter was counted in five random fields at 400x magnification . Assays were done in duplicate and were repeated in three independent experiments . The data represents the average of two technical duplicates and three independent experiments +/− SE . Wells of a 48-well plate were coated with 200 µl Matrigel ( BD Biosciences ) . HUVECs , transduced to express KapB , or constitutively active MK2 ( MK2-EE ) or an empty vector control , were harvested with trypsin , counted , centrifuged and resuspended in basal EBM-2 medium . 5×104 cells were added to the top of each matrigel-containing well in basal media . Cells were incubated at 37°C and observed every hour . Over time , HUVECs progress from individual cells to connected tubules by first forming sprouted cells and then connections between small groups of cells . As tubulogenesis progresses , cells form connected tubes , enclosed polygons and complex meshwork ( layered tubes of cells ) as described in [66] . At 5 hours , extensive tubules , often with the presence of polygons and complex mesh , had formed and were visualized using an Olympus CKX41 inverted microscope . Representative images were captured using Image Pro Plus . To quantify the angiogenic potential of KapB , an angiogenic score was calculated by adapting the methods of [66] . For each condition , 5 random fields of view at 200x magnification were observed and the number of enclosed polygons was counted . Further , the presence of a complex meshwork was given the following score: 1 = no complex mesh , 2 = presence of complex mesh , and 3 = complex mesh of cell thickness >4 cells . The angiogenic score was then calculated as the product of the number of enclosed polygons and the complex mesh score . Assays were done in duplicate and were repeated in three independent experiments . The data represents the average of two technical duplicates and three independent experiments +/− SE . Graphing and statistical analyses were performed using GraphPad Prism software . All data are presented as mean +/− SEM of three independent experiments unless otherwise indicated . Paired parametric t-test was used for comparison between two groups . One-way repeated measures ANOVA was used for comparison between multiple groups where the mean of each group was compared to the mean of the control group . P values are displayed when p<0 . 05 . | We have only scratched the surface in understanding how viruses control host gene expression . Several viruses disrupt important sites of post-transcriptional control of gene expression known as processing bodies ( PBs ) , but underlying regulatory mechanisms and biological relevance remain poorly understood in most cases . Our study shows that the Kaposin B ( KapB ) protein of Kaposi's sarcoma ( KS ) -associated herpesvirus , known to block the degradation of a class of labile host mRNAs , does so by constitutively activating a signaling axis involving MK2 , hsp27 , p115RhoGEF and RhoA , thereby dispersing PBs . Thus , PB disruption may support the secretion of host pro-inflammatory cytokines and angiogenic factors that underlies KS tumor formation . Furthermore , by activating RhoA , KapB also causes cytoskeletal rearrangements , accelerated cell migration and angiogenesis in an endothelial cell model . Our findings position KapB as a key contributor to viral reprogramming of endothelial cells . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"viruses",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"medical",
"microbiology",
"viral",
"pathogens",
"microbial",
"pathogens",
"pathogens",
"biology",
"and",
"life",
"sciences",
"microbiology",
"kaposi's",
"sarcoma-associated"... | 2015 | Viral Activation of MK2-hsp27-p115RhoGEF-RhoA Signaling Axis Causes Cytoskeletal Rearrangements, P-body Disruption and ARE-mRNA Stabilization |
The finding that regular spatial patterns can emerge in nature from local interactions between organisms has prompted a search for the ecological importance of these patterns . Theoretical models have predicted that patterning may have positive emergent effects on fundamental ecosystem functions , such as productivity . We provide empirical support for this prediction . In dryland ecosystems , termite mounds are often hotspots of plant growth ( primary productivity ) . Using detailed observations and manipulative experiments in an African savanna , we show that these mounds are also local hotspots of animal abundance ( secondary and tertiary productivity ) : insect abundance and biomass decreased with distance from the nearest termite mound , as did the abundance , biomass , and reproductive output of insect-eating predators . Null-model analyses indicated that at the landscape scale , the evenly spaced distribution of termite mounds produced dramatically greater abundance , biomass , and reproductive output of consumers across trophic levels than would be obtained in landscapes with randomly distributed mounds . These emergent properties of spatial pattern arose because the average distance from an arbitrarily chosen point to the nearest feature in a landscape is minimized in landscapes where the features are hyper-dispersed ( i . e . , uniformly spaced ) . This suggests that the linkage between patterning and ecosystem functioning will be common to systems spanning the range of human management intensities . The centrality of spatial pattern to system-wide biomass accumulation underscores the need to conserve pattern-generating organisms and mechanisms , and to incorporate landscape patterning in efforts to restore degraded habitats and maximize the delivery of ecosystem services .
A succession of spatially explicit ecological models in the early 1990s indicated that large-scale regular spatial patterns could arise within homogeneous landscapes from local biotic interactions alone [1]–[3] , with potentially profound implications for the maintenance of biodiversity and ecological stability [4] , [5] . At first , large-scale ordered patterns were harder to find in natural systems than in systems of equations: the title of a 1997 review questioned whether ecological self-organization was “robust reality” or merely a theoretical set of “pretty patterns” [6] . Over the past decade , however , multiple studies have shown that regular patterns are both common and persistent across a range of ecosystems [7]–[10] . But the crucial questions of whether and how these patterns influence ecosystem functioning remain unanswered [11] . Here , we show that the even spacing of subterranean termite mounds in an apparently homogeneous African savanna provides a template for parallel spatial patterning in tree-dwelling animal communities . We further show that the uniformity of this pattern at small spatial scales elevates the productivity of the entire landscape , providing support for models linking spatial pattern with ecosystem functioning [12]–[15] . Our study site in central Kenya ( 0°20′ N , 36°53′ E ) is a wooded grassland on level vertisol soils . The high clay concentration ( 40%–60% ) of these soils reduces water infiltration and causes shrink-swell dynamics that can shear plant roots [16] . In this habitat , which is widespread in East Africa , a single Acacia species ( A . drepanolobium , an “ant plant” ) constitutes >97% of the canopy over a continuous understory dominated by five perennial bunchgrasses . Thus , the area appears strikingly homogeneous for a tropical terrestrial ecosystem ( Figure S1 ) . In addition to symbiotic ants ( 3 Crematogaster spp . , 1 Tetraponera sp . ) , A . drepanolobium canopies are inhabited by non-predatory insects , predatory insects and spiders , and insect-eating dwarf geckos ( Lygodactylus keniensis ) . Lygodactylus keniensis is diurnal and exclusively arboreal , and males are territorial; along with the arthropod arboreal predators in the system , it preys almost exclusively on tree-feeding insects [17] , excepting workers of the Acacia-ant species [18] . In this ecosystem , fungus-cultivating termites ( Macrotermitinae: Odontotermes ) nest within low , subterranean mounds ( 10–20 m diameter , <0 . 5 m tall ) ( Figure S1 ) . As in many other drylands worldwide , these mounds occur in regular , over-dispersed ( evenly spaced ) spatial patterns ( see Figure 1A ) [19]–[21] . These patterned arrays are apparently maintained in space by termite colonies' competitive partitioning of the habitat into non-overlapping foraging territories [16] , [21] , [22] and in time by cycles of colony extinction and mound re-colonization ( Text S1 ) [16] , [21] . The mounds themselves are typically treeless and dominated by the perennial bunchgrass Pennisetum stramineum [23] , and they are often centuries old [16] , [21] . Mound surfaces contain more sand than the surrounding soils , which , along with termite-created macrochannels , increases aeration and water infiltration [24] , decreases shrink-swell dynamics , and accelerates soil formation from bedrock [16] . Termite mounds are comparatively moist microenvironments in dry savannas [25] , and mound soils are enriched in nitrogen and phosphorus ( by 70% and 84% , respectively , relative to off-mound soils: [26] ) . This combination of physical and chemical properties results in greater production of grasses on termite mounds , which is clearly visible in multispectral satellite photographs using the near-infrared band ( Figure 1A ) . Woody productivity is also enhanced on mound edges: A . drepanolobium foliar nitrogen content is 19% greater within 5 m from edges [26] , new-shoot growth of trees is 60% greater within 10 m of edges [23] , and trees adjacent to mounds are ∼120% more likely to fruit in a given season [26] . Acacia trees in nitrogen-rich near-mound soils rely less heavily on fixed atmospheric N than trees in the inter-mound matrix [27] , providing one possible explanation for these patterns . We used field observations and manipulative experiments to show that , by enhancing primary productivity on and around their mounds , termites exert positive indirect effects upon multiple trophic levels of arboreal animals , from herbivorous insects to spiders and geckos . We further show that these indirect effects create spatial patterns in the abundance and reproductive output of these taxa that parallel the patterning of termite mounds . We then extrapolated these patterns to the landscape scale , showing that uniform spacing of termite mounds ( over-dispersion ) increases secondary and tertiary productivity relative to simulated landscapes in which mounds were randomly distributed .
We quantified the spatial pattern of termite mounds using Ripley's K [28] , showing that they exhibit significant over-dispersion at spatial scales<100 m ( Figure S2 ) . We then quantified consumer abundances at different distances from mounds to determine whether this pattern of high-productivity patches provides a template for the distribution of prey and predator communities . Aerial arthropods ( N = 3 , 277; 42% Hemiptera , 32% Diptera , 11% Coleoptera , 15% others ) were significantly more abundant in sticky traps at 10 m than at 30 m from termite mounds ( Figure 1B ) . Moreover , the sides of the traps facing the mounds captured nearly 40% more arthropods than the away-facing sides , and this discrepancy was more pronounced close to mounds . These results suggest that mounds are a local source of insects dispersing into the inter-mound matrix . For tree-dwelling arthropods , sampled by spraying with insecticide ( N = 1 , 503; 55% spiders , 23% Coleoptera , 6% Lepidoptera ) , the abundance and biomass of all arthropods and of predatory taxa only , and the abundance ( but not biomass ) of prey taxa , decreased significantly with distance from mound centers ( Figure 1C–D ) . More than 96% ( 824 of 858 ) predatory arthropods in our samples were spiders , so our conclusions about predatory arthropods in general are also true for spiders in particular . Only two of 4 , 780 ( 0 . 04% ) total arthropods were termites ( both alates captured in sticky traps ) , indicating that termites themselves were not driving the pattern in prey abundance or providing a prey base for arboreal insectivores ( Text S2 ) . To determine whether the gecko L . keniensis was more abundant near mounds , we exhaustively searched 60 randomly selected trees in hemispheres around each of three mounds where all trees had been mapped ( N = 180 trees total ) . On average , trees occupied by one or more geckos ( N = 72 ) were significantly closer to mounds ( median = 18 . 3 , interquartile range = 12 . 6–26 . 3 ) than were unoccupied trees ( N = 108 , median = 26 . 4 , interquartile range = 15 . 4–31 . 4; Wilcoxon Z = −3 . 6 , p = 0 . 0003 ) . We constructed a candidate set of 108 ordinal-logistic regression models to identify the factors influencing the number of geckos on trees . Ranking these models using the sample-size-corrected Akaike Information Criterion ( AICc ) [29] revealed that the number of geckos on a tree was principally a function of the tree's size ( estimated surface area of the main stem ) and its proximity to the nearest termite mound ( Table S1 ) . The best model achieved good correspondence between observed and predicted values ( Figure S3A ) and showed strong predictive power when applied to a larger dataset ( N = 477 trees ) collected 3 y after the model was parameterized ( August 2009; Figure S3B ) . Using the parameters of this model , we determined the mean probability of occupancy ( ≥ 1 gecko ) as a function of mound proximity for five percentiles of tree size ( Figure 2A ) , showing that whereas very large trees are nearly always occupied , occupancy of intermediate-sized trees hinges strongly on location relative to termite mounds . We then estimated the mean number of geckos expected on a tree of median size occurring anywhere in an actual landscape of mapped mounds within our study area , which revealed a strikingly uniform pattern in the spatial probability distribution of these predators ( Figure 2B ) . Two questions remain about the mechanisms causing this pattern . First , mean tree size decreased with distance from the nearest mound ( F1 , 475 = 19 . 3 , p<0 . 0001 ) , suggesting non-independence of these two effects in the regression models . Second , it is not clear what drives the “mound-proximity” effect in the gecko regressions . The decrease in arthropod abundance with increasing distance from mounds suggests—but does not demonstrate—that geckos might be responding to differences in prey availability . We addressed these issues experimentally using artificial “trees” consisting of wooden posts of two sizes ( “large” and “small” ) . These posts differed only in their size . At each of 12 mounds , we placed one post of each size at both 10 m ( “close” ) and 30 m ( “far” ) from the mound center , controlling for nearby tree density . From October 2006 to June 2007 , we surveyed all posts 12 times . Occupation frequency was greater on large and close posts than small and far ones ( Figure 3A ) , as our model predicted . Moreover , the mean snout-vent length and weight of territorial male geckos ( but not females ) was greater for close posts ( but did not vary by post size ) ( univariate effect test of mound proximity from two-way ANOVA: F1 , 32 = 7 . 8 , p = 0 . 009 for length; F1 , 32 = 4 . 5 , p = 0 . 04 for weight ) . Thus , tree size and mound proximity have independent effects on gecko occurrence . To test whether the effect of mound proximity indeed arose from variation in prey availability , we repeated the artificial-tree experiment with one modification: we affixed plastic cups to the base of every post and , each morning from August–December 2007 , we added 3–9 non-flying insects to cups at all far posts only . Consistent with the prey-availability hypothesis , prey supplementation almost equalized the overall mean occupation frequencies at 10 m ( 0 . 799±0 . 064 geckos/post ) and 30 m ( 0 . 771±0 . 046 geckos/post ) from termite mounds ( Figure 3A , Table S4 , Text S3 ) . We conclude that termites indirectly influence gecko distribution by increasing local densities of arthropod prey near mounds . That trees near mounds are on average larger , exhibit greater foliar N content [26] and growth rates [23] , and rely more heavily on soil N than atmospheric N [27] all strongly suggest an additional likely mechanism: termite activity increases mean tree size and thus , indirectly , occupancy of those trees . That adding prey increased occupation of far posts is consistent with a behavioral response to food availability but does not rule out a simultaneous numerical ( i . e . , reproductive ) response . It is difficult to measure variation in reproductive output for L . keniensis , which has a fixed clutch size . But we can easily measure fecundity for another group of arboreal predators , female spiders , which produce conspicuous egg masses of variable size . ( Multiple-regression and AICc analyses of occupancy patterns showed that the abundance of adult spiders , like that of geckos , rapidly decreased with distance from the nearest termite mound; Table S2 . ) We haphazardly collected 110 reproductive females of the most common arboreal web spider ( Araneidae: Cyclosa sp . ) from April–June 2008 at varying distances from 12 non-adjacent mounds ( ∼9 spiders per mound ) and reared their egg masses . Both the total number of spiderlings per female and the mean number of spiderlings per egg sac within each egg mass decreased with distance from mounds ( Figure 3B ) , indicating that spiders respond numerically to high-productivity mound areas . These results are a unique demonstration that subterranean termites indirectly enhance abundance and create spatial pattern across multiple trophic levels of tree-dwelling animals . We next tested the theoretical prediction [12]–[14] that the regularity of the spatial pattern should increase overall production at the landscape scale . We cannot easily conduct this test in the field: we lack “control” regions without patterned mounds , and the experimental elimination of termite colonies would not eliminate the gradient in production because the mound structures would likely persist for decades . We therefore employed a null-model approach . We first superimposed a grid of 5×5 m cells upon a ∼360 , 000 m2 mapped portion of the study site ( see Figure 1A ) . We then used best-fitting regression models ( selected from candidate sets using AICc , as for geckos above; Tables S1–S3 ) to estimate the value of the response variables at each of the 14 , 400 sample points defined by this grid . ( To isolate the effects of mound proximity , we set all other predictor variables equal to their observed median values . ) Next , for each variable , we averaged the values of the 14 , 400 sample points to yield a landscape-mean value . ( We refer to these values as “over-dispersed-landscape means” because each is based on the evenly spaced distribution of mapped mounds in the real landscape . ) Finally , we compared the over-dispersed-landscape mean for each variable with “random-landscape means” obtained from applying the same models to 1 , 000 simulated landscapes of randomly distributed mounds ( see Materials and Methods: spatial analysis of patterns in consumer abundance ) . For every variable , the real ( over-dispersed ) landscape was far more productive than were simulated landscapes with randomly distributed mounds . The estimated means for all response variables in the over-dispersed landscape were >99 . 9th percentile of the means obtained from the 1 , 000 randomly generated landscapes ( Figure 4 ) . Because the mean values of all variables were strongly negatively correlated with mean nearest-mound distance ( Figure S4 ) , the uniform spacing of mounds in the real landscape—which minimizes the mean distance from any arbitrarily chosen point in the landscape to the nearest mound—maximizes mean values of the response variables . This analysis assumes that multiple mounds would not have additive effects on the density or productivity of trees and tree fauna , which might lead to greater production under clumped scenarios than our models ( which were based only on nearest-mound distance ) predict . We tested this assumption . When we collected data on gecko abundances in 2009 to test the predictive power of our best-fitting ordinal regression model , we recorded the locations of the two mounds closest to each tree ( hereafter , nearest and second-nearest ) . Adding second-nearest mound distance as a predictor to our best model did not improve the model whatsoever ( −2×log-likelihood = 666 . 265641 for both models ) . Because nearest and second-nearest mound distances were very weakly correlated ( r = −0 . 08 ) , this result is not biased by collinearity of the predictors . We therefore conclude that the “mound effect” on production is adequately characterized by distance to the single nearest mound . Our simulation results also assume that trees are equally likely to occur anywhere in the landscape , so gradients in tree density might complicate our conclusions . In fact , the response of tree density to mound proximity is weak and inconsistent , and a separate set of landscape simulations in which we accounted for these effects ( see Materials and Methods ) produced qualitatively identical results ( Figure S5 ) . Collectively , our data show ( a ) that a regularly patterned array of termite mounds induces parallel patterning in the abundance and reproductive output of tree-dwelling fauna , ( b ) that these patterns arise via both consumptive ( i . e . , trophic ) and non-consumptive ( i . e . , engineering ) indirect interactions , and ( c ) that the uniformity of the pattern increases the total biomass of prey and predators in the landscape . This emergent effect of spatial pattern upon a fundamental ecosystem function ( productivity across trophic levels ) confirms theory predicting linkages between patterning and production . Our results further imply that the landscape-level effects of any set of features that induce local gradients in ecological processes are likely to hinge on the spatial patterning of these features , with highly uniform spacing often producing the strongest net outcomes . Future work should address how the landscape-level effects of different spatial patterns vary with the shape and slope of biotic distance-response functions , as well as with possible interactive effects among patterned features . Our study highlights the importance of conserving pattern-inducing taxa and processes—in this case , termites and their mound-building activities . In Africa's fields and pastures , termites are sometimes eradicated to protect crops and forage , and mounds are sometimes destroyed to redistribute the nutrients within them [20] , yet these actions may actually diminish overall landscape productivity . More generally , recent research shows that the influence of remnant trees in forest regeneration attenuates with distance [30] , which means that restoration efforts will be most effective if organisms—such as trees and corals intended as nucleating agents for forests and reefs—are added to landscapes in uniform , gridded patterns ( as theory suggests: [14] ) . Conversely , other desired ecological outcomes , such as the persistence of competitively inferior plant species , may be most effective if elements are arranged in aggregated distributions [31] . The uniform spacing of plants in plantations , and the ability to manipulate the spatial configuration of the agroscape , likewise provides opportunities to both study and apply the consequences of spatial patterning for the delivery of ecosystem services such as pest control and pollination [32] , [33] .
We conducted this study between June 2006 and August 2009 at the Mpala Research Centre ( 0°20′ N , 36°53′ E ) in central Kenya . Total rainfall during this period was 1 , 810 mm . The annual pattern was variable and tri-modal , with peaks in August ( 70 mm ) and November ( 93 mm ) of 2006; April ( 86 mm ) , June ( 152 mm ) , and September ( 98 mm ) of 2007; and May ( 99 . 6 mm ) , July ( 58 . 7 mm ) , and October ( 143 . 8 mm ) of 2008 , followed by drought . The study area is underlain by flat , heavy-clay vertisol ( “black cotton” ) soils of recent volcanic origin , which are characterized by impeded drainage , pronounced shrink-swell dynamics [34] , [35] , and species-poor plant communities [36] . These soils and associated vegetation occur in many parts of East Africa , including Nairobi National Park and the western extension of Serengeti National Park [37] . Each A . drepanolobium tree is inhabited by one of four species of symbiotic ants ( Crematogaster and Tetraponera spp . [38] ) . Trees inhabited by each ant species support robust communities of insects , predatory arthropods ( primarily spiders and mantids ) , and dwarf geckos ( Lygodactylus keniensis ) . Because worker ants do not appear to be frequent prey for any of the arboreal predators we studied , we did not include them in our samples or surveys . Adult male L . keniensis are distinguished by a chevron-shaped row of pre-anal pores and fiercely defend territories consisting of individual trees or adjacent trees with contiguous canopies , while several females and subadults can occur on the same tree [18] . Nests built by subterranean termites ( Macrotermitinae: Odontotermes ) occur in this and similar habitats throughout upland East Africa . As described above , various physical , chemical , and hydrological properties of mounds lead to greater productivity of both woody and herbaceous plants , revealed at our sites by both field measurements [23] , [26] and remotely sensed imagery ( Figure 1A; see also [39] ) . Similar effects of termite mounds on primary productivity occur in many systems [20] . Like all Macrotermitinae , Odontotermes spp . farm fungus in combs underground . Alates typically emerge with the first heavy rain of the wet season [21] , but workers and soldiers are virtually never exposed aboveground ( see Results ) , foraging instead within covered runways on the soil surface . Macrotermitinae mounds have long been known to occur with apparently even spacing in upland Kenya and other semi-arid regions throughout Africa ( Figure 1A; [16] , [19] , [21] ) . Such regular spacing ( 20–120 m between mounds ) arises from colonies' exhaustive partitioning of space into non-overlapping foraging areas ( Text S1 ) [20]–[22] . We quantitatively evaluated mound patterning at different spatial scales using Ripley's K function [28] . Using the near-infrared band from an orthorectified Quickbird satellite image ( June 20 , 2003 ) with 2 . 4 m resolution and ∼3 km2 extent , we visually identified circular areas of high productivity , corresponding to termite mounds . To verify accuracy of our visual photo-interpretation , we field-recorded the geographic coordinates of 50 mounds using a CMT March II GPS ( 1–5 m accuracy ) , which we overlaid as a shape file upon the satellite image , confirming that these ground-truthed points did indeed appear as mounds on the image . We then applied Ripley's K to the coordinates of these mounds using Programita [40] , establishing that the spacing of mounds is significantly uniform at spatial scales<100 m ( Figure S2 ) . We identified all arthropods to order and some spiders and beetles to family . For tree-dwelling arthropods , we analyzed predators and prey both separately and together . Because the ecology and taxonomy of the invertebrate fauna of this region is poorly characterized , we treated mantids and spiders as predators and assumed that all other insects represented “prey . ” Although this categorization slightly undercounts predators by excluding some predators from trophically mixed orders such as Coleoptera , a previously published stable-isotope analysis of these same samples [17] showed that such miscategorizations represented a small fraction ( ∼5% ) of all insects . We sampled aerial arthropods ( N = 3 , 277 ) from July 2007 to February 2008 using 10×13 cm yellow sticky traps ( Olson Products , Medina , Ohio , USA ) . Each month ( except August 2007 ) , we hung one sticky trap at chest height at both 10 m and 30 m from the center of each of 12 mounds . Trap locations were random with respect to prevailing wind direction , and we marked the side of each trap that was facing towards the mound . We collected all traps after 24 h and identified and counted all arthropods . We analyzed log-transformed data using repeated-measures MANOVA ( in JMP 8 . 01 ) with arthropod abundances in each month as the dependent variables . The between-subjects factors in this analysis were mound proximity , trap orientation ( i . e . , facing towards or away from mound ) , their interaction , and mound identity ( because individual termite mounds vary in size and primary productivity ) . The within-subjects factor was time . The number of known predators ( 182 spiders ) captured using this method was insufficient for separate analysis . We sampled arboreal arthropods ( N = 1 , 503 ) by spraying tree stems and canopies with 0 . 6% alphacypermethrin from a backpack sprayer [17] . Trees were selected randomly subject to the criteria that they were approximately 1–2 m tall ( mean ± SD: 1 . 73±0 . 25 m ) and occupied by the most common Acacia-ant symbiont , Crematogaster mimosae . Prior to spraying , we arranged white sheets beneath the canopies . On calm days , we sprayed each tree for 30 s and collected all arthropods falling onto the sheets during the subsequent 30 min . We sampled 10 trees at each of four mounds in July 2007 ( a wet period ) and an additional 10 trees at each of three mounds in February 2008 ( a dry period ) , for a total of 70 trees at seven mounds . We measured the distance from each tree to the nearest mound center . After identifying and counting all samples , we dried them to constant mass at 60°C and weighed them ( nearest 0 . 0001 g ) to obtain separate dry-biomass measures for prey and predators . We constructed candidate sets of multiple regressions and selected the best models for subsequent analyses ( see “Regression Modeling of Response Variables , ” below ) . Because we sampled only similarly sized trees , there were no statistically significant pairwise correlations between tree size and either mound proximity or the arthropod response variables ( all p≥0 . 25 ) , although tree size did appear as a predictor in the best-fitting ( as determined by AICc ) models of total-arthropod abundance and predatory-arthropod biomass ( Table S3 ) . In 2006 , as part of a concurrent study , Doak , Brody , and Palmer used a laser rangefinder ( accurate to within 10 cm ) to map and individually number all trees within ∼35-m-radius semicircles centered on each of six mounds . For this study , we selected three of these mounds and used a random-number generator to choose 60 trees ( >1 m tall ) for search . The mounds were several hundred meters apart . From July–August 2007 , Pringle and two assistants exhaustively searched all trees for geckos , using ladders to reach high branches and probing within any hollows . For all 180 trees , we recorded the number of geckos , mound proximity ( nearest 0 . 1 m ) , nearest-neighbor distance ( nearest 0 . 1 m ) , height ( nearest 0 . 1 m ) , basal diameter ( nearest 0 . 1 cm ) , and resident Acacia-ant species . In August 2009 , we repeated this process for an additional 477 trees at the same three termite mounds to obtain an independent dataset with which to test the predictive power of our best-fitting model of gecko abundance . Female spiders guarding egg masses were selected opportunistically and haphazardly . Upon collection , we preserved female spiders in ethanol , placed the egg masses in ventilated plastic cups in a common laboratory environment , and checked them periodically . When we were confident that all spiderlings had emerged from the egg sacs ( ∼14 d after first emergence ) , we froze the spiderlings and counted them using a dissecting microscope . It is extremely unlikely that cannibalism among spiderlings during this interval influenced our results; we are not aware of any reports of cannibalism among newly hatched juveniles in the Araneidae , and a bias would require that cannibalism was much more frequent among offspring of females far from termite mounds , which is improbable . Of 110 egg cases , 106 ( 96% ) hatched in the laboratory . We calculated two measures of reproductive output for each female: total number of spiderlings and mean number of spiderlings per egg sac per female ( each female's egg mass consisted of 1–12 individual egg sacs ) . Jocqué confirmed the genus identification for this as-yet-undescribed species and measured the width of the carapace and the combined length of the tibia and patella of leg I for each adult female . Measurements were made with an ocular graticule in a Leica M10 stereo microscope ( measurement unit = 0 . 0164 mm ) . We could not obtain reliable carapace-width measurements for four females , giving us a final sample size of 102 . Both measures of reproductive output were positively correlated with female carapace width ( r = 0 . 24 , F1 , 100 = 6 . 1 , p = 0 . 02 and r = 0 . 20 , F1 , 100 = 4 . 3 , p = 0 . 04 , respectively ) , while neither measure of reproductive output varied with tibia + patella length ( both p≥0 . 5 ) . Female carapace width was not significantly correlated with termite-mound proximity ( r = −0 . 11 , F1 , 100 = 1 . 3 , p = 0 . 3 ) . To determine the mechanisms ( especially the role of termite-mound proximity ) influencing tree-dwelling-arthropod abundance , gecko occurrence , and spider fecundity , we constructed sets of candidate regression models and ranked them using the AICc . Prior to constructing candidate sets , we visually examined the shape of the relationship between each response variable and mound proximity . In all candidate sets , we included both a raw mound-proximity term and one-or-more nonlinear transformations ( loge for gecko abundance; square-root for spider abundance; loge , square , and square-root for arthropod abundance/biomass and Cyclosa fecundity; Tables S1–S3 ) , as well as categorical mound-identity effects and ( for all variables except spider fecundity ) raw and transformed effects of tree size . Complete model sets and AICc results are available from Pringle on request . To explain variation in the number of geckos on trees , we employed ordinal logistic regression using the “Ordinal” routine in the Statbox 4 . 2 Toolbox for MATLAB ( http://www . statsci . org/matlab/statbox . html ) . The dependent variable—number of geckos per tree in our dataset of 180 trees at three mounds—took values 0 , 1 , 2 , or≥3 . Independent variables included combinations of mound proximity , tree size ( i . e . , estimated surface area of the main stem , using the equation for the area of the side of a cylinder , which we considered a more accurate representation of gecko habitat size than either height or diameter alone ) , distance to the nearest tree ≥1 m tall , and mound identity . We constructed 108 candidate models using combinations of these variables , their natural logarithms , and their first-order interactions . We then ranked these models using AICc ( Table S1 ) . Our notation and interpretation follow Burnham and Anderson [29] . Of the five most likely models , all contained terms for both tree size and mound proximity ( Table S1 ) . Examination of the complete model set revealed that the importance of variables decreased in the order: tree size > mound proximity > mound identity > nearest-tree distance . We evaluated the goodness-of-fit and predictive ability of our best model by comparing mean model predictions with mean observed results for 12 different categories of trees ( assigned based on which of three mounds and which of four 10 m distance intervals they belonged to ) . We performed this test using both the original 180-tree dataset from 2006 ( which reveals goodness-of-fit , Figure S3A ) and a novel 477-tree dataset from 2009 ( which reveals the substantial predictive power of our model: Figure S3B ) . We conducted multiple-regression analyses of the abundance of adult arboreal spiders ( based on our sample of 70 trees that we sprayed with insecticide ) that largely paralleled our ordinal-regression analyses of gecko abundance . Independent variables included combinations of mound proximity , estimated tree surface area , square-root transformations of these variables , their first-order interactions , and mound identity . We constructed 26 candidate models and ranked them using AICc ( Table S2 ) . Of the eight most likely models , all contained terms for mound proximity and mound identity ( which encompassed seasonal variations in abundance ) ; no model lacking a term for mound proximity received any empirical support . Examination of the complete model set revealed that variable importance decreased in the order: mound identity ≈ mound proximity > tree size . We analyzed arthropod abundance and biomass data ( log-transformed to meet parametric assumptions ) using multiple regression . Response variables included total arthropod abundance and biomass , prey-arthropod abundance and biomass , and predatory-arthropod abundance and biomass . We constructed 24 candidate models for each variable . Unlike for geckos and spiders alone , all models for arthropod abundance/biomass contained a mound-proximity term ( either raw or transformed , as described above ) , but none contained interactions . The other predictors included raw and log-transformed tree size ( estimated surface area , as described above ) and mound identity . The best models ( Table S3 ) explained between 2% ( for prey-arthropod biomass ) and 68% ( for predatory-arthropod abundance ) of the variation in the response variables . For spider fecundity , we constructed 16 candidate models using raw and transformed mound proximity , female carapace width , and mound identity as predictors . The best model ( Table S3 ) explained 23% of the variation in spider fecundity . For each response variable , we used the single best model for all spatial analyses ( see below ) and all tests of statistical significance for individual predictors . The log-transformed mound-proximity term was a better predictor of gecko abundance than the linear form . Square-transformed mound-proximity terms best approximated the responses of all arthropod variables except predator abundance , which was best approximated by a linear term , and prey biomass , which was best approximated ( albeit non-significantly ) by a log-transformed term . We extrapolated to the landscape scale for six response variables ( predatory-arthropod abundance/biomass , total arthropod abundance/biomass , gecko occurrence , and spider fecundity ) using a 600×600 m section of our study area , which included 62 termite mounds ( Figure 1A ) . We mapped this area using Quickbird satellite imagery and calibrated the map in the field with a laser rangefinder . We subdivided this area with a grid of 5×5 m cells that defined 14 , 400 sample points and computed the distance of each point to the nearest mound . We then applied the best-fitting regression model to the distance value for each point . This enabled us to produce the spatial probability distribution of gecko abundance in Figure 2B and also to compute the mean value across all points for each response variable . In making these estimates , we used the observed median value of all other predictor variables ( which , depending on the model in question , included tree size , spider-carapace width , and categorical mound-identity effects: Tables S1–S3 ) . In other words , although we refer to these estimates as “real-” or “over-dispersed-landscape” values , they actually estimate abundances in hypothetical landscapes in which the distribution of mounds corresponds to reality , but all trees , female spiders , etc . are assumed to be an identical , typical size , and tree density is assumed to be uniform throughout the landscape ( we provide support for this last and most-important assumption below ) . We then compared the estimated mean landscape value of each response variable from the actual , over-dispersed mound landscape with the corresponding distributions of values from simulated random landscapes that lacked the uniform spacing of real mounds ( Figure 4 ) . To do this , we generated 1 , 000 hypothetical landscapes that had the same number of mounds ( N = 62 ) as the real landscape , but a nearly Poisson ( independent and random ) distribution of the mounds . To generate these landscapes , we randomly picked sets of latitudes and longitudes to define the location of each mound center within the same size area as that actually surveyed . The only restriction on these randomly generated mound positions was that all mound centers be at least 10 m apart , since real mounds have radii of 5–10 m and cannot overlap . For each simulated mound landscape , we repeated the estimation procedure ( described above for the actual landscape ) to produce 1 , 000 hypothetical landscape-wide average values for each response variable . Comparing the mean estimated values of all sample points from the 1 , 000 hypothetical random landscapes with the mean values obtained from the actual , over-dispersed landscape generated the results shown in Figure 4 . As mentioned above , the most likely real-world complication that could influence the results of the randomization tests just described is variation in tree densities at different distances from mounds . For one mound in our study area , we mapped the positions of all trees out to ∼35 m in all directions; for five other mounds , we did the same for a ∼35 m radius semicircle . We used these data to determine whether and how the densities of trees >1 m tall vary with distance from mound centers . We used the following procedure to determine densities . First , for each mound , we used a MATLAB routine to construct Voronoi or Thiessen polygons [41] around each tree in the mapped area , which provides an estimate of local , tree-specific density . Next , we constructed a convex-hull line between the trees that defined the outermost boundaries of the mapped area . Because Voronoi polygons cannot be accurately estimated around these boundary trees , we eliminated these trees from further analyses . We then binned the remaining trees into either 5 or 10 m distance bins and divided the number of trees in each bin by their summed polygon areas to arrive at a distance-specific tree-density estimate . Using these estimates , we applied general linear models with distance and mound identity as independent variables and density as the dependent variable . For 5 m bins , mound ID is highly significant ( p<10−7 ) and distance is also significant ( p = 0 . 01 ) , with density increasing on average with distance , but with much variation between mounds ( no significant interaction effect ) . For 10 m bins , mound ID , distance , and their interaction are significant ( mound: p<10−7; distance: p = 0 . 03; interaction: p = 0 . 00003 ) . For these results , removal of the interaction effects makes the main effect of distance non-significant due to strongly varying patterns in density across mounds . These highly variable results make it unlikely that any consistent patterns in tree density would bias the conclusions of our simulations . However , to test for such effects , we used the results from the 5 m binned data to estimate changes in densities for the average mound effect ( density = 0 . 071+0 . 0011×distance ) . We used this relationship to estimate a weighted average of all sampled points in the real and simulated landscapes that accounted for the relative tree density at different distances . These results ( Figure S5 ) are qualitatively identical to our original results ( Figure 4 ) . Our experiments were designed to ( a ) isolate the effects of tree size and mound proximity on gecko occupation rates and ( b ) determine whether mound proximity truly represented a trophic effect . We created artificial gecko habitat using wooden posts of two different sizes . “Large” posts were 2 . 6±0 . 06 m tall and 10 . 3±0 . 58 cm in diameter ( means ± SD ) ( Figure S1B ) . “Small” posts were 2 . 0±0 . 03 m tall and 7 . 7±0 . 35 cm in diameter ( Figure S1D ) . All posts contained 12 1 . 5 cm diameter holes for refuge and a 1 m long horizontal crossbar to provide a perch . These posts were very similar to trees from the geckos' perspective , as we determined after the experiment by comparing occupancy of the 48 posts ( over the first 12 surveys ) with 48 real trees that matched in size ( mean estimated surface area = 0 . 67 m2 for both real and artificial trees; mean occupancy = 0 . 6±0 . 2 and 0 . 7±0 . 1 geckos/tree , respectively , means±95% CI ) . At each of 12 termite mounds , we placed one post of each size at both 10 m ( “close” ) and 30 m ( “far” ) from the mound center . We placed the large and small posts 5 m from one another at each distance . To control for any confounding influence of neighboring tree density , we situated each post 3 m from the nearest tree ≥2 m tall and ensured that the density of trees in the 20×20 m area surrounding the posts at each distance did not differ ( close density = 23 . 8±6 . 6 , far density = 24 . 2±6 . 0 , means ± SD ) . We completed the experimental setup on September 30 , 2006 and waited 1 mo prior to beginning surveys to allow geckos to adjust to the habitat perturbation and colonize the posts . Between October 28 , 2006 and June 20 , 2007 , we conducted 12 surveys of all posts . Because of the simplified architecture of the posts , we suspect that detection probability approached 100% . During five of these surveys , we captured geckos ( N = 134 ) , which we sexed , measured ( nearest 1 mm ) , weighed ( nearest 0 . 001 g ) , and replaced . To avoid pseudoreplication arising from multiple counts of the same individuals , we treated the posts as the experimental units: our response variables were mean adult gecko length and weight ( by sex ) and mean occupation frequency ( number of geckos observed on each post divided by 12 , the number of surveys ) of each post . ( Because up to three geckos sometimes occurred simultaneously on a single post , occupation frequency could take values >1 . ) Size and weight data were compared using two-way factorial ANOVA . To ascertain whether the effect of mound proximity on gecko occupation arose from variation in prey availability , we repeated this experiment in conjunction with daily food supplementation . Insects , which included mealworms , termite workers found in dried dung , and sweep-net contents ( all collected off site ) , were always added to the cups between 7:30 and 8:30 a . m . , immediately prior to the onset of peak gecko activity . We did not attempt to capture any geckos during this phase of the experiment ( Text S3 ) . As before , we conducted 12 surveys of all posts . Mean monthly rainfall did not differ between the pre- and post-prey-addition periods ( 63 . 2±46 . 1 mm and 52 . 4±34 . 9 mm , respectively; F1 , 12 = 0 . 2 , p = 0 . 7 ) . We analyzed the data from both runs of this experiment using a single repeated-measures MANOVA design ( in JMP 8 . 01 ) . The dependent variables were the mean occupation frequencies of each post during the 12 surveys prior to prey addition and the same mean frequencies for the 12 surveys conducted during daily prey addition to the far posts . The between-subject factors were post size ( large versus small ) , mound proximity ( close versus far ) , their interaction , and mound identity . The within-subject factor was time ( pre- versus post-prey addition ) . In this design , the significant time × mound proximity interaction ( Table S3 ) shows the equalizing effect of experimental food supplementation . | Local interactions between organisms in nature can scale up to produce strikingly regular patterns across entire landscapes . With improvements in satellite imagery , such patterns are increasingly reported in the ecological literature . It remains unclear , however , whether the existence of such patterns actually matters for key ecosystem processes such as productivity . In semi-arid East Africa , below-ground mounds built by Odontotermes termites frequently occur in uniform , “polka-dot” arrangements . We show that , due to the ways in which termites modify the soil , these mounds are hotspots of plant and animal productivity: close to termite mounds , plants grow more quickly , herbivorous and predatory animals are more abundant , and reproductive output is greater than is true farther away from mounds . Moreover , the evenly spaced distribution of termite mounds means that all points in the landscape are relatively close to the nearest mound—with the result that ecosystem-wide productivity is greater under the actual distribution of mounds than it would be if the same number of mounds were randomly situated . Thus , although subterranean termites may be less visible and charismatic than the large mammals of African savannas , they are nonetheless critically important engineers of structures and patterns that underpin ecosystem function . | [
"Abstract",
"Introduction",
"Results/Discussion",
"Materials",
"and",
"Methods"
] | [
"evolutionary",
"biology/evolutionary",
"ecology",
"ecology/conservation",
"and",
"restoration",
"ecology",
"ecology",
"evolutionary",
"biology/animal",
"behavior",
"plant",
"biology/plant-environment",
"interactions",
"ecology/community",
"ecology",
"and",
"biodiversity",
"ecolo... | 2010 | Spatial Pattern Enhances Ecosystem Functioning in an African Savanna |
Microbe- or host damage-derived patterns mediate activation of pattern-triggered immunity ( PTI ) in plants . Microbial virulence factor ( effector ) -triggered immunity ( ETI ) constitutes a second layer of plant protection against microbial attack . Various necrosis and ethylene-inducing peptide 1 ( Nep1 ) -like proteins ( NLPs ) produced by bacterial , oomycete and fungal microbes are phytotoxic virulence factors that exert immunogenic activities through phytotoxin-induced host cell damage . We here show that multiple cytotoxic NLPs also carry a pattern of 20 amino acid residues ( nlp20 ) that triggers immunity-associated plant defenses and immunity to microbial infection in Arabidopsis thaliana and related plant species with similar characteristics as the prototype pattern , bacterial flagellin . Characteristic differences in flagellin and nlp20 plant responses exist however , as nlp20s fail to trigger extracellular alkalinization in Arabidopsis cell suspensions and seedling growth inhibition . Immunogenic nlp20 peptide motifs are frequently found in bacterial , oomycete and fungal NLPs . Such an unusually broad taxonomic distribution within three phylogenetic kingdoms is unprecedented among microbe-derived triggers of immune responses in either metazoans or plants . Our findings suggest that cytotoxic NLPs carrying immunogenic nlp20 motifs trigger PTI in two ways as typical patterns and by inflicting host cell damage . We further propose that conserved structures within a microbial virulence factor might have driven the emergence of a plant pattern recognition system mediating PTI . As this is reminiscent of the evolution of immune receptors mediating ETI , our findings support the idea that there is a continuum between PTI and ETI .
Plants make use of a bipartite immune system to cope with microbial infection [1] . Microbial pattern recognition by host-encoded immune receptors is essential for the activation of plant antimicrobial defenses . Perception by pattern recognition receptors ( PRRs ) of pathogen-associated molecular patterns ( PAMPs ) is referred to as PAMP-triggered immunity ( PTI ) [2] , [3] . PTI is an ancient form of plant immunity that provides protection to host non-adapted pathogens , but limited or basal immunity to host-adapted microbes only . In addition , plant-derived damage-associated molecular patterns ( DAMPs ) are released either by the deleterious activities of secreted microbial enzymes or toxins that activate plant PTI in a PRR-dependent manner [2] , [4] . Host-adapted plant pathogens employ effectors to suppress PTI and to establish infection [5] . Co-evolution of hosts and host-adapted microbes has resulted in effector-triggered immunity ( ETI ) , which is dependent on immune receptors recognizing effectors directly or indirectly through sensing effector-mediated manipulations of host targets [1] , [3] , [6] . Plants recognize a wide range of proteinaceous , carbohydrate or lipophilic PAMPs [2] , [7] . In most cases , small epitopes within such patterns provide ligands for plasma membrane-localized PRRs [8] , [9] . These ligands are often broadly conserved among microbial species or genera and are not subject to frequent mutations likely because of their vital cellular functions [10] . Well-studied microbe-derived triggers of plant immunity comprise structurally conserved N-terminal regions of bacterial flagellin ( flg22 ) and elongation factor Tu ( EF-Tu , elf18 ) or oligomeric carbohydrate fragments of bacterial peptidoglycans , fungus-derived chitin or oomycete cell wall β-glucans [2] , [7] . Plant perception systems for flagellin , peptidoglycans or chitin are rather widespread among plant families , suggesting that these systems are evolutionarily ancient [11] . In contrast , EF-Tu or β-glucan receptors appear to have evolved more recently as perception systems are restricted to members of the Brassicaceae or Fabaceae families only [12] , [13] . Likewise , more recently identified Sclerotinia sclerotiorum-derived proteinaceous SSCF1 or Xanthomonas campestris-derived EMAX are recognized by Brassicaceae only [14] , [15] . Moreover , identification of a tomato flagellin perception system that recognizes flagellin epitopes different from flg22 [16] , or of a rice receptor that recognizes a central fragment of EF-Tu structurally unrelated to elf18 [17] , suggest substantial dynamics in PRR evolution . More systematic studies on PRR distribution patterns among Arabidopsis thaliana ecotypes have further revealed that individual pattern recognition specificities might also be lost during evolution [14] , [15] , [18] . In fact , such ecotype-specific differences in microbial pattern recognition are now increasingly being used to identify novel plant PRRs and to test their phytoprotective potential in crops [14] , [19] . Altogether , loss and gain of plant PRRs appears to be a characteristic of plant immunity that is also reminiscent of the dynamics underlying evolution of plant immune receptors mediating microbial ETI [20] . Plant pathogenic microbes produce multiple effector proteins that are secreted into the plant apoplastic space or that are translocated into host cells by means of specialized translocation systems , such as type III secretion systems of Gram-negative bacteria [5] . Major functions of these effectors comprise suppression of host immunity and microbial accomodation in host tissues . Plant immunity-stimulating activities of effectors are mediated by immune receptors recognizing effector structures or effector-mediated manipulations of host targets [3] , [5] . Likewise , phytopathogens preferring hemibiotrophic or necrotrophic lifestyles employ a wide range of structurally unrelated host-selective and host-nonselective toxins that are essential for establishment of infection [4] . As some effectors , some microbial toxins have been demonstrated to have dual functions in plant-microbe encounters as virulence factors and triggers of plant immunity [4] , [21] , [22] , [23] . Toxin-mediated host immune activation is thereby supposed to be the result of host target manipulation or host cellular damage . NLPs form a superfamily of proteins that are produced and secreted by bacterial , fungal and oomycete species [24] , [25] , [26] . NLPs have initially been discovered as cytotoxic proteins triggering leaf necrosis and plant defenses in dicotyledonous , but not in monocotyledonous plants [27] . 3D-structural analyses of Pythium aphanidermatum or Moniliophthora perniciosa NLPs , respectively , revealed substantial fold conservation with cytolytic , pore-forming actinoporins from marine organisms , suggesting that NLPs destabilize plant plasma membranes during infection thereby facilitating host cell death [4] , [28] . Indeed , cytotoxic NLPs from the necrotrophic or hemibiotrophic phytopathogens Pectobacterium carotovorum pv . carotovorum ( PccNLP ) , Pythium aphanidermatum ( PyaNLP ) or Phytophthora parasitica ( PpNLP ) were shown to be key virulence factors sharing identical fold requirements for NLP phytotoxin and virulence activities [4] . Notably , NLP-mediated phytotoxicity and plant immune marker gene expression also required the same structural features . This finding together with the fact that the native 3D structure of NLP is required for its immunogenic activity , strongly supports the assumption that NLP-mediated plant cell necrosis results in the release of immunogenic DAMPs [4] . This process is reminiscent of microbial toxin-triggered inflammasome activation in vertebrates [29] , [30] . There is accumulating evidence that NLP effectors have diversified in function [26] . The fungal pathogen Mycosphaerella graminicola produces MgNLP that is toxic on dicot plants , but not on its monocot host , wheat [31] . Moreover , knock-down of a cytotoxic NLP in Verticillium dahliae resulted not only in reduced virulence on host plants , but also in reduced vegetative growth and conidiospore formation , suggesting a role of this NLP in asexual reproduction in addition to its role in fungal pathogenicity [32] . The biotrophic oomycete Hyaloperonospora arabidopsidis was shown to produce up to 10 NLP proteins all of which failed to cause necrosis in dicot plants including the host Arabidopsis [33] . Likewise , 11 of 19 Phytophthora sojae NLPs tested lacked phytotoxic activities [24] . Functional diversification among the two NLP subfamilies in this hemibiotrophic oomycete was further supported by the fact that genes encoding non-cytotoxic NLPs were expressed predominantly during early ( biotrophic ) phases of infection whereas cytotoxic NLP genes were expressed only at the onset of necrotrophic growth [24] , [34] , [35] . In this study , we have investigated plant immunogenic activities of NLP virulence factors in greater detail . Mutations that rendered PccNLP inactive with respect to cytotoxicity , host virulence and plant immune activation , also abolished the cytotoxic activity of another NLP ( PpNLP ) , but surprisingly left intact its ability to trigger plant defenses . This suggested the presence of another , yet unidentified immunogenic activity of PpNLP . The elicitor activity of mutated PpNLP could be pinpointed to a peptide fragment ( nlp20 ) that triggered plant defenses in a manner comparable to that of bacterial flagellin . Importantly , immunogenic nlp20 fragments were found frequently in NLPs of bacterial , oomycete and fungal origin . In sum , we demonstrate that a common microbial effector harbors a PAMP motif that is found in both prokaryotic and eukaryotic microbes . Thus , its widespread occurrence is unique among microbial triggers of metazoan or plant innate immunity . In addition , the identification of two independent plant immunogenic mechanisms ( PAMP- and toxin-induced immunity ) within a particular microbial virulence factor is unprecedented and reveals an intricate complexity of microbial virulence and plant immune activation .
Typically , small epitopes within microbial patterns are sufficient for their immunogenic activities [2] , [11] , [19] . In search for such an immunogenic epitope within PpNLP , nested synthetic peptides covering the entire PpNLP protein sequence were produced and tested for their abilities to trigger ethylene production or PR1::GUS expression ( Figure 2 ) . Two overlapping peptides spanning residues G84 to V129 of PpNLP ( peptides c and j ) proved both to be able to elicit plant defense-associated responses . These peptides share residues G100-D113 ( GVYAIMYSWYFPKD , peptide 1 , Table 1 ) , suggesting that this fragment constitutes the core of the immunogenic activity of PpNLP . Another set of nested synthetic peptides spanning the peptide 1 sequence were analyzed for their abilities to trigger ethylene production in Arabidopsis leaf disks . The EC50 value determined for peptide 1 was 322 nM ( Table 1 ) . N-terminal deletion peptides lacking residues G100-Y106 ( GVYAIMY , peptide 4 ) or C-terminal deletion of residues K112 and D113 ( peptide 2 ) substantially reduced elicitor activity , suggesting that both motifs are important for the immunogenic potential of PpNLP ( Table 1 ) . In agreement with this , a peptide carrying an N-terminal extension , but lacking K112D113 ( peptide 2 ) or peptides with C-terminal extensions , but lacking residues Y102-Y106 ( peptides 4–6 , peptide 12 ) were all inactive . Substantial N-terminal extension ( peptide 8 ) did not increase elicitor activity of this peptide in comparison to peptide 1 , suggesting that no further sequence information N-terminal of the G100-Y106 motif is required for elicitor activity of PpNLP . To refine C-terminal sequence requirements for PpNLP elicitor activity , we further tested peptides containing the Y102-Y106 motif or a fragment thereof ( A103-Y106 ) and different C-terminal extensions beyond residues K112D113 ( Table 1 ) . These studies revealed two peptides with EC50 values of 14 ( peptide 9 ) or 1 , 5 nM ( peptide 13 ) , respectively , as the most elicitor-active peptides , which are both substantially more active than peptide 1 ( Table 1 ) . As both peptides lack residue Y102 we conclude that it is dispensable for elicitor activity . In contrast , C-terminal extensions gradually enhance elicitor activities of the respective peptides , and together with motifs A103-Y106 and K112D113 constitute major determinants of PpNLP immunogenic activity . Because of the origin of this motif from PpNLP protein and because of the number of residues building peptides 9 and 13 , these peptides were re-named nlp20 ( PpNLP ) and nlp24 ( PpNLP ) , respectively . To identify amino acids within both peptides that are essential for their elicitor activities , an alanine-scanning mutagenesis was conducted ( Table 1 ) . Individual exchange of each amino acid by alanine ( except A103W ) identified residues I104 , Y106 , W108 , and Y109 , of which replacement reduced immunogenic activities of mutant peptides more than 1 , 000-fold as compared to nlp24 ( PpNLP ) . All other exchanges had significantly less or no effect on the activities of the mutant peptides ( Table 1 ) . Importantly , all of these residues are part of or are in close proximity to the A103-Y106 motif , highlighting again the importance of this motif for PpNLP elicitor activity . Individual exchanges in the C-terminal regions of nlp20 ( PpNLP ) or nlp24 ( PpNLP ) , respectively , affected immunogenic activities of the mutant peptides in a rather moderate manner . To test whether the nlp20 ( PpNLP ) motif derived from cytotoxic NLP would retain both immunogenic and cell death-causing activities , leaf necrosis and plasma membrane permeabilization assays were performed using equimolar concentrations of intact PpNLP and of PpNLP-derived nlp20 ( PpNLP ) as well as 10-fold higher concentrations of the latter . As shown in Figure S1D–E , nlp20 ( PpNLP ) failed to trigger either response , suggesting strongly that its immunogenic activity is not linked to cell death or plasma membrane disintegration . This conclusion is further supported by our findings that heat treatment or mutations within intact PpNLP abolished its necrosis-inducing activity , but not its ability to trigger immunity-associated defenses ( Figure 1A ) . Likewise , low nanomolar concentrations of nlp20 ( PpNLP ) are required to trigger plant defenses ( Table 1 ) , which is in clear contrast to the failure of the peptide to trigger necrosis at 10 mikromolar concentrations ( Figure S1D ) , again disconnecting nlp20 ( PpNLP ) -induced defenses from the cytotoxic potential of intact PpNLP . NLPs are widespread microbial patterns that are found in bacteria , fungi and oomycete species [27] , [36] . Inspection of NLP protein sequences from the various lineages revealed the presence of an nlp20-motif in numerous cases . To test whether nlp20-like peptides of NLPs from different microbial origins harbor PAMP activity , synthetic peptides representing bacteria- ( Bacillus subtilis , Bacillus halodurans ) , fungus- ( Fusarium oxysporum , Botrytis cinerea ) or oomycete-derived ( Pythium aphanidermatum ) sequences orthologous to nlp20 ( PpNLP ) were analyzed for their immunogenic potential . As shown in Table 2 , all peptides tested exhibited the ability to trigger ethylene production , MAPK activation , production of reactive oxygen species ( oxidative burst ) , PR1::GUS expression , and callose apposition ( Figure S2A–E ) . For ethylene production , EC50 values were determined and found to be very similar for all nlp20 orthologs tested ( Table 2 ) . Arabidopsis seedling growth inhibition on agar plates containing flg22 , elf18 or AtPep1 is a hallmark plant response to those patterns that are recognized by LRR-RK-type pattern recognition receptors . Remarkably , reduced seedling size in the presence of PAMPs was only detectable in flg22 control treatments , but not in cases when nlp20 ( PpNLP ) or orthologous nlp20 peptides were tested ( Table 2 , Figure S2F ) . Likewise , Phytophthora parasitica-derived nlp20 failed to trigger an extracellular alkalinization response in Arabidopsis cell suspensions ( Figure S2G ) . As shown in Figure 1 , PccNLP mutants lacking cytotoxic activity also lacked immunogenic activity . In contrast , PpNLP mutants devoid of cytotoxic activity remained immunogenic due to the presence of the nlp20 motif . In agreement with the apparent absence of an immunogenic nlp20 motif in PccNLP , a synthetic peptide derived from the PccNLP sequence that corresponds to the nlp20 motif in PpNLP ( GSFYALYFLK DQILSGVNSGHR ) , proved largely inactive with respect to activating ethylene formation , MAPK activation and PR1::GUS expression ( Figure S3 ) . Residual ethylene-inducing activity was observed for this peptide ( EC50 5520 nM ) , which was approximately 400 times less active than nlp20 ( PpNLP ) ( EC50 14 nM ) ( Figure S3A ) . To analyze the relative distribution of nlp20 recognition systems among plants , we first tested whether other Brassicaceae species beside Arabidopsis thaliana responded to this peptide . As shown in Figure 3 , Arabis alpina , Thlaspi arvense , and Draba rigida mounted an ethylene response to nlp20 ( PpNLP ) treatment , suggesting that nlp20 recognition is widespread among the Brassicaceae family . Notably , another species from the genus Arabidopsis , Arabidopsis lyrata , did not respond to nlp20 ( PpNLP ) , but did so to the control treatment with flg22 . Although surprising in the first place , this finding might just reflect that PAMP responsiveness is often even not entirely conserved among ecotypes of the same species . Neither solanaceous plants ( tomato , potato , Nicotiana benthamiana ) nor parsley ( Petroselinum crispum , an Apiaceae ) or wheat ( Triticum aestivum , a monocotyledonous grass ) responded to nlp20 ( PpNLP ) ( Figure 3 ) . Failure to detect nlp20 ( PpNLP ) responses in parsley is in agreement with our previous studies showing that this plant species lacks the ability to recognize NLP peptide fragments [37] . In contrast , ethylene production was detectable after treatment of leaves of lettuce ( a member of the Asteraceae family ) with nlp20 ( PpNLP ) ( Figure S4 ) . As lettuce did not respond to an nlp20 ( PpNLP ) derivative lacking PAMP activity in Arabidopsis thaliana ( Figure S4 ) we conclude that nlp20 ( PpNLP ) perception systems in both plants exhibit similar ligand specificities . Whether nlp20 recognition is even more widespread among plant families requires comprehensive , systematic surveys of its immunogenic activity . PAMP treatment results in enhanced plant immunity to subsequent microbial infection [2] , [11] , [19] . For example , treatment with flg22 of Arabidopsis plants prior to infection with virulent Pseudomonas syringae pv . tomato DC3000 reduced bacterial growth by about 100-fold within three days post infection when compared to bacterial growth rates on mock-treated plants ( Figure 4 ) . Likewise , nlp20 ( PpNLP ) treatment limited bacterial growth rates on ecotype Col-0 to a similar extent as did flg22 treatment , suggesting that both patterns have an immunogenic activity ( Figure 4 ) . Nlp20 ( PpNLP ) also reduced bacterial growth on an fls2 efr genotype ( Figure S5 ) , thus ruling out flg22 contamination issues here that have raised concerns about recent studies on plant PRRs [38] . In contrast , pre-treatment with immunogenically inactive nlp20 ( PccNLP ) ( Figure S3A ) or peptide 20 ( Table 1 ) did not result in reduced bacterial growth , which documents the ligand specificity of the observed biological phenomenon ( Figure S5B ) . As further shown in Figure 4 , nlp20 ( PpNLP ) treatment also primed Arabidopsis plants for enhanced immunity to infection by the fungal phytopathogen Botrytis cinerea . Lesion sizes in plants pretreated were significantly smaller than those observed in mock-treated plants . Likewise , pre-treatment of lettuce with an nlp24 peptide derived from H . arabidopsidis nlp24 ( HaNLP3 ) enhanced resistance to infection with Bremia lactucae ( Figure S6 ) . Altogether , our findings demonstrate that nlp20 recognition contributes to plant immune activation and to reduced symptom development and microbial growth rates on infected plants .
Cytotoxic NLPs are microbial virulence factors facilitating both microbial infection and activation of plant immunity-associated responses . Toxin-mediated release of diffusible DAMPs from lyzed plant cells and subsequent PRR-mediated plant immune activation in neighboring cell layers or local systemic tissues has been proposed as the likely molecular mechanism underlying immunogenic activity of , for example , PccNLP [4] . Experimental findings in support of this model comprise ( i ) identical fold requirements for microbial virulence and immune activation , ( ii ) requirement of natively folded cytotoxic NLPs for immune activation and ( iii ) the apparent lack of NLP enzyme activity ( no primary sequence or 3D-structure similarity to known enzymes ) . Other examples for microbial toxins as triggers of plant defenses include Fusarium spp . -derived fumonisin , Phomopsis amygdali-derived fusicoccin or Cochliobolus victoriae-derived , victorin . Toxin-induced immunity is thus considered a hallmark of innate immunity not only in metazoans , but also in plants [19] , [29] . To our surprise , we have been able to unveil a second molecular mechanism by which cytotoxic NLPs are able to evoke plant immunity . This discovery was spurred by findings that mutations that rendered PccNLP non-cytotoxic and non-immunogenic failed to have the same effect in other cytotoxic NLPs , such as PpNLP . We have now been able to identify a peptide motif ( nlp20 motif ) within PpNLP and other NLPs that is missing in PccNLP . This strongly suggests that cytotoxic NLPs carrying the nlp20 motif are potentially capable of evoking plant immunity by two different mechanistic modes , by toxin action and by a classical PAMP motif . To our knowledge , this is an unprecendented finding as microbial patterns with dual immunogenic activities are currently unknown in both metazoan and plant immunity . These results shed light on how intricately complex and mechanistically diverse microbe sensing in individual plant microbe encounters might be . In support of this notion , Arabidopsis thaliana alone is capable of recognizing at least seven structurally different patterns derived of pseudomonads [10] . This and substantial diversification and expansion of gene families encoding plant PRRs strongly suggests that many more immunogenic patterns than those currently known might exist [14] , [19] . Thus , the number of microbial patterns recognized in particular plant-microbe interactions together with different immunogenic modes of individual microbial patterns appears to represent an immunogenic potential of microbial surfaces of which complexity is most likely much larger than anticipated previously . PAMPs triggering immunity in metazoans or plants are supposed to be widespread among microbial species [2] , [39] . Bacteria-derived flagellin , peptidoglycans or lipopolysaccharides are patterns that are found across taxonomical orders . Likewise , fungus-derived chitin or oomycete-derived ß-glucan structures are extremely common among these organisms . The immunogenic nlp20 motif is unique , however , in that it is found conserved not only in NLPs of bacterial origin , but also in fungal and oomycete genera . To our knowledge , none of the currently known triggers of metazoan or plant innate immunity shows a comparably wide distribution pattern among eukaryotic and prokaryotic microbes . By using synthetic nlp20 peptides derived from two bacterial , oomycete and fungal organisms , respectively , we could demonstrate PAMP activity associated with NLPs from all three lineages . Currently , 1 , 091 NLP sequences can be retrieved from databases using the PpNLP1 sequence as query ( 221 , 558 , 312 sequences of bacterial , fungal , oomycete origin , respectively ) . Preliminary inspection of these sequences for the presence of the nlp20 motif and of those residues that are crucial for its PAMP activity ( I104 , Y106 , W108 , Y109 ) revealed that a remarkably low number of bacterial sequences ( 20 out of 221 ) , but a majority of fungal and virtually all oomycete NLPs likely contain an elicitor-active nlp20 motif . In sum , this motif is a predominant feature within a vast number of NLP sequences particularly in eukaryotic NLP-producing microorganisms . Importantly , in comparison to the relatively small numbers of NLP-encoding genes in fungal genomes , the number of NLP genes has expanded significantly in oomycete species . For example , the P . sojae genome harbors 33 NLP genes 20 of which have been shown to be expressed , whereas H . arabidopsidis encodes 12 NLP genes 8 of which are expressed early during plant infection [24] , [33] . Clustering of these sequences in species-specific groups and the occurrence of non-cytotoxic NLPs indicates rapid expansion and functional diversification within these gene families without an apparent deleterious effect on the nlp20 motif . Predictions whether bacterial NLPs have largely lost this motif during evolution ( such as phytopathogenic P . carotovorum ) or whether nlp20 motif-containing NLPs have been acquired from eukaryotic species via horizontal gene transfer are difficult to make as of now . The nlp20 motif exhibits molecular features similar to that of the prototype immunogenic pattern , bacterial flagellin ( flg22 ) [2] . It is active at low nanomolar concentrations , it triggers several immunity-associated plant responses including broad spectrum immunity to bacterial and fungal infection , and it is evolutionarily conserved within NLPs . Although flagellin and nlp20 patterns trigger a set of overlapping plant responses , substantial differences are apparent , too . For example , flg22 evokes extracellular alkalinization in Arabidopsis cell suspensions and Arabidopsis seedling growth retardation [40] , [41] , whereas nlp20 does not trigger these responses . Whether or not these differences in the immunogenic activities of both patterns reflect recognition by different receptor types remains to be seen . In summary , we here report the identification of a common immunogenic pattern within a microbial virulence factor . The nlp20 motif of bacterial , fungal or oomycete NLPs possesses the ability to trigger plant immune responses in a manner comparable to bacterial flagellin . Unique features of this pattern comprise ( i ) its presence in both prokaryotic and eukaryotic microbes and ( ii ) the fact that it constitutes a second immunogenic principle within cytotoxic NLPs . Further , we suggest that a microbial effector might have driven the emergence of plant pattern recognition systems mediating PTI . This is important as it is reminiscent of the evolution of immune receptors mediating recognition of pathogen race-specific microbial effectors and activation of ETI [3] , [5] , [6] . In this respect , our findings support the concept of an evolutionary and functional continuum between plant PTI and ETI [20] .
Arabidopsis Col-0 and efr fls2 plants were grown in soil at 22°C , 8 h light and used for the experiments at an age of 5–6 weeks . Plants used for infection assays were grown under translucent cover . 5–6 weeks old Arabidopsis thaliana Col-0 plants were primed 24 hours before bacterial or fungal infection by leaf infiltration of nlp20 ( PpNLP ) , flg22 , C6 ( 1 µM peptide solution ) or mock-treatment , respectively . To assess bacterial growth rates , Pseudomonas syringae pv . tomato DC3000 ( Pst DC3000 ) strain was used . The strain was maintained at 28°C on King's B medium ( 20 g l−1 glycerol , 40 g l−1 proteose pepton , 15 g l−1 agar ) containing rifampicin and cycloheximide ( 50 µg ml−1 ) . Overnight cultures were centrifuged , washed twice in 10 mM MgCl2 and adjusted to a bacterial density of 104 cfu ml−1 . Primed leaves were pressure-infiltrated with the bacterial solution and the plants were kept under high humidity . Leaves were harvested and surface sterilized in 70% EtOH and ddH20 for 1 minute each . Two leaf discs per plant were stamped out , ground in 10 mM MgCl2 , diluted serially 1∶10 and plated on LB plates containing the appropriate antibiotics . After 2 days of incubation , colony-forming units were counted . For fungal infection , primed Arabidopsis leaves were drop-inoculated with 5 µl droplets of Botrytis cinerea isolate BO-10 containing 5×106 spores ml−1 in PDB ( potato dextrose broth , Sigma ) and kept under high humidity . Photographs were taken 2 days after infection and lesion sizes were determined using the Photoshop CS6 Lasso tool . Selected pixels were counted and the lesion size in cm2 was calculated using a 0 , 5 cm2 standard . For oomycete Bremia lactucae infection , L . sativa leaf discs were vacuum-infiltrated with 1 µM nlp24 ( HaNLP3 ) and 24 hours later treated with a 20 µl droplet spore suspension ( 120 spores/µl ) . Sporulation was assessed 8 days post inoculation . For functional studies , secretory expression of NLPs was performed either in Pichia pastoris GS115 ( secretory expression plasmid pPIC9K , Multi-Copy Pichia Expression Kit Instructions , Invitrogen ) or in the NLP-deficient Pectobacterium carotovorum subsp . carotovorum SCC3200 strain ( Pcc nlp− ) . Isolation of PccNLP proteins from the periplasmic space of transgenic Pcc nlp− was performed by osmotic shock as described [42] . Purification of PpNLPs from P . pastoris culture medium or from P . carotovorum subsp . carotovorum SCC3200 periplasmic protein solution was achieved by ion exchange chromatography followed by gel filtration ( GE Healthcare ) . As ion exchanger either HiTrap Q FF ( equilibrated in 20 mM Tris-HCl pH 8 . 5: PpNLP ) or HiTrap SP FF ( equilibrated in 50 mM MES pH 5 . 7: PccNLP ) was used . Following elution ( 0–500 mM KCl in equilibration buffer ) , NLP containing fractions were pooled and subjected to HiLoad™ 16/60 Superdex 75 , equilibrated in 150 mM KCl in the corresponding buffer . NLP containing fractions were finally pooled and dialyzed against H2O . Protein concentrations were calculated by UV spectroscopy ( wavelength λ280 ) using the protparam tool ( http://web . expasy . org/protparam ) to determine protein-specific extinction coefficients ε280 for each protein . Determinations were verified by SDS-PAGE using a standard protein solution . Peptides were purchased from Genscript Inc . , prepared as 10 mM stock solutions in 100% DMSO , and diluted in water prior to use . DMSO concentrations corresponding to those in peptide solutions used in this study did not trigger themselves any of the responses shown here . For MAPK activity assays , infiltrated plant material was harvested after 15 minutes and frozen in liquid nitrogen before used for protein extraction in 20 mM Tris-HCl pH 7 . 5 , 150 mM NaCl , 1 mM EDTA , 1% Triton X-100 , 0 , 1% SDS , 5 mM DTT , Complete Protease Inhibitor Mini , EDTA-free ( Roche , Mannheim ) , PhosStop Phosphatase Inhibitor Cocktail ( Roche , Mannheim ) . After pelleting the cell debris ( 10 min , 16000 g , 4°C ) , the supernatant ( 30 µg protein ) was separated on a 10% SDS-PAGE and transferred to a nitrocellulose membrane and activated MAPK6 , 3 and 4 were detected by western blotting using the anti phospho p44/42-MAPK antibody from rabbit ( Cell Signaling Technology , The Netherlands ) . For ROS burst measurements two leaf pieces , floated on ddH2O overnight , were placed in one well of a 96-well plate , containing 100 µl of a 20 µM L-012 and 0 . 5 µg ml−1 peroxidase solution . Background was measured shortly in a 96-well Luminometer , ( Mithras LB 940 , Berthold Technologies ) before elicitation with a peptide solution or control treatment respectively . The detection of ethylene was performed as described [40] . Leaf pieces were incubated in 20 mM MES buffer , pH 5 . 7 . To visualize callose apposition , leaves were treated as described [40] and harvested 24 hours after infiltration of a peptide solution . Quantification of callose was performed by counting selected pixels and calculated in % relative to the respective image section of the leaf surface . Pictures were analyzed using Photoshop CS6 Magic tool , hereby removing background and leaf-veins within a certain color range . ( Use: white , Mode: normal , Opacity: 100% ) . Medium alkalinization in suspension-cultured Arabidopsis cells and detection of GUS enzyme activity in PR1::GUS transgenic Arabidopsis plants were performed as described previously [40] , [43] . Surface-sterilized Arabidopsis Col-0 seeds were grown in ½ MS liquid medium supplemented with 1 µM of nlp20 ( PpNLP ) peptide or its orthologs respectively , and flg22 or H2O serving as controls . Root length of two weeks-old seedlings was determined upon transfer onto agar plates . Arabidopsis leaves were infiltrated with 300 nM PpNLP or PccNLP , heat-denatured ( 1 . 5 hours , 95°C ) proteins or mutant versions ( H121A D124A ) , respectively . RNA was isolated using the RNeasy Plant MiniKit ( Qiagen ) and synthesis of cDNA was performed by means of the RevertAidTM MuLV reverse transcriptase ( Fermentas ) . Quantitative real-time PCR amplification was carried out in the presence of SYBR Green ( Bio-Rad ) with an iQ5 iCycler ( Bio-Rad ) . Amplification of EF1-α served as internal standard . Data were analyzed according to the 2−ΔΔCT-method [44] . Gene induction ( fold change ) by NLPs was presented as the average of 3 determinations plus or minus standard deviation relative to the expression level of H2O infiltration . Calcein release from intact Arabidopsis plasma membrane vesicles was performed as described [4] . | Eukaryotic host immunity to microbial infection requires recognition systems sensing the presence of potential invaders . Microbial surface structures ( patterns ) or host breakdown products generated during microbial attack serve as ligands for host immune receptors ( pattern recognition receptors ) mediating activation of immune responses . Microbial pathogens employ , however , host-targeting effector proteins to establish infection , and the efficiencies of microbial pathogen attack and host defense mechanisms determine the outcome of microbe-host interactions . Necrosis and ethylene-inducing peptide 1 ( Nep1 ) -like proteins ( NLPs ) from bacteria , oomycetes and fungi are cytotoxic virulence factors ( effectors ) that trigger plant immunity through toxin-induced host cell damage . Here we show that , in addition , numerous NLPs harbor a characteristic 20-mer sequence motif ( nlp20 ) that is recognized by Brassicacae plant species and perception of which confers immunity to infection by bacterial , oomycete and fungal pathogens . Our findings provide evidence that cytotoxic NLPs are virulence factors that trigger plant immunity by pattern recognition and by inflicting host cell damage . We further conclude that NLPs from prokaryotic and eukaryotic microorganisms and from three organismal kingdoms evoke plant defense . Such an exceptionally wide taxonomic distribution of microbe-derived triggers of immunity has neither been reported before from metazoans nor from plants . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"plant",
"science",
"biology",
"and",
"life",
"sciences",
"immunology",
"microbiology",
"species",
"interactions",
"agriculture"
] | 2014 | A Conserved Peptide Pattern from a Widespread Microbial Virulence Factor Triggers Pattern-Induced Immunity in Arabidopsis |
Mapping expression quantitative trait loci ( eQTLs ) has been shown as a powerful tool to uncover the genetic underpinnings of many complex traits at molecular level . In this paper , we present an integrative analysis approach that leverages eQTL data collected from multiple population groups . In particular , our approach effectively identifies multiple independent cis-eQTL signals that are consistent across populations , accounting for population heterogeneity in allele frequencies and linkage disequilibrium patterns . Furthermore , by integrating genomic annotations , our analysis framework enables high-resolution functional analysis of eQTLs . We applied our statistical approach to analyze the GEUVADIS data consisting of samples from five population groups . From this analysis , we concluded that i ) jointly analysis across population groups greatly improves the power of eQTL discovery and the resolution of fine mapping of causal eQTL ii ) many genes harbor multiple independent eQTLs in their cis regions iii ) genetic variants that disrupt transcription factor binding are significantly enriched in eQTLs ( p-value = 4 . 93 × 10-22 ) .
Expression quantitative trait loci , or eQTLs , are genetic variants that are associated with gene expression levels . Mapping eQTLs can help in dissecting the molecular mechanisms by which genetic variants impact organismal phenotypes . Recent studies [1–3] have revealed that there are substantial overlaps between eQTLs and genetic variants identified from genome-wide association studies ( GWAS ) of disease phenotypes . In addition , eQTL mapping provides a powerful tool for investigating the regulatory machinery in different tissues [4 , 5] or cellular environments [6–8] . In this paper , we jointly address three outstanding issues in eQTL mapping . First , due to the high experimental cost , most available eQTL data sets typically have limited sample sizes . To improve power of eQTL discovery , it becomes necessary to aggregate evidence across multiple data sets . Second , because a gene is typically regulated by many regulatory elements , it is highly likely that there exist multiple independent eQTLs in its proximity ( i . e . , cis region ) . In this scenario , a multi-SNP analysis is required to uncover all relevant cis acting genetic factors involved in the gene regulation process [9] . Third , the availability of extensive functional annotations [10–12] now enables integration of functional genomic information into eQTL analysis , which can be useful to dissect the functional basis of eQTLsf . Linking genomic annotations to eQTLs goes beyond genetic association analysis , and helps gain a better understanding of the underlying biological processes . Individually , some of these three issues have been discussed by previous works . For example , [3 , 9 , 13–16] discussed single SNP analysis of eQTLs jointly from different studies , populations or tissues . But these methods do not naturally extend to multi-SNP analysis . [17–20] examined the enrichment of selected genomic features in cis-eQTLs , mostly based on single SNP association results . To the best of our knowledge , there is no existing approach that jointly addresses all three issues in a systematic way . In this paper , we demonstrate an integrative analysis framework to perform multi-SNP fine mapping analysis of eQTLs using cross-population samples . Our statistical methods stem from an established Bayesian framework proposed by [14 , 15 , 21] , which has been successfully applied in mapping eQTLs from multiple tissues . We apply our statistical framework to analyze the data from the GEUVADIS project [20] , where the expression-genotype data are collected from five population groups . In GWAS , trans-ethnic meta-analysis of genetic association data from diverse populations has been shown to be a powerful tool in detecting novel complex trait loci and improving resolution of fine mapping of causal variants by leveraging population heterogeneity in local patterns of linkage disequilibrium ( LD ) and allele frequencies [22 , 23] . This approach , to the extent of our knowledge , has not been applied to eQTL analysis . Utilizing cross-population expression-genotype data , we are interested in identifying eQTL signals that show consistent effects in all populations . Furthermore , we aim to examine whether we have sufficient statistical power to identify multiple independent cis-eQTL signals with the available aggregated sample size . Finally , we set out to investigate whether the genetic variants that disrupt transcription factor ( TF ) binding are enriched in eQTLs , and we integrate such annotations to further improve fine mapping analysis of eQTLs . Our three main aims are also inherently inter-related and reinforce each other . With higher power through sample aggregation in mapping eQTLs , our method can identify at a high resolution potentially multiple genomic regions that harbor casual eQTLs . Consequently , these efforts improve the statistical power and precision of localization for our functional analysis .
We start with a brief description of our statistical framework and general strategy for multi-SNP fine mapping analysis in a meta-analytic setting across multiple populations . We performed a series of simulation studies to evaluate the proposed Bayesian multi-SNP fine mapping procedures . Specifically , we simulated under the setting where eQTL data are collected from multiple population groups . This section briefly summarizes the main simulation results , and we leave the relevant details in the Materials and Methods section . First , we examined the power of the proposed Bayesian multi-SNP analysis approach in identifying multiple independent association signals using cross-population samples . In particular , we compared its performance with the commonly applied single SNP meta-analysis procedure and a ( step-wise ) conditional meta-analysis procedure [25 , 26] , treating each population group as a participating study in a meta-analysis . To simulate the genotype-expression data in multiple population groups , we used the real genotypes of 2 , 500 SNPs from 100 randomly selected and distinct genomic regions ( i . e . , 25 consecutive SNPs per region ) in the GEUVADIS data . As a consequence , there are generally modest to high levels of LD within each region , but unremarkable levels of LD between regions . ( In other words , we assembled 100 relatively independent LD blocks for each gene . ) It is worth noting that within each region , the variation in patterns of LD between populations is maintained . We randomly assigned one to four eQTL SNPs for each gene and simulated their expression levels according to a system of linear models ( see Material and Methods section and Section S . 5 of S1 Text ) . For evaluation , we assessed the ability of each procedure in correctly identifying the 25-SNP regions ( i . e . , the LD blocks ) that harbor the true eQTLs . The simulation results indicate that the proposed Bayesian multi-SNP analysis approach is the most powerful in identifying independent eQTL signals among the three methods compared ( Fig 1 ) . Further investigation reveals that all three methods achieve similar power when genes contain only a single eQTL; however when a gene harbors multiple independent cis-eQTLs , the proposed Bayesian fine mapping approach becomes highly advantageous ( S1 Fig ) . In addition , the simulation results show that the reported PIPs from the proposed Bayesian approach are well-calibrated ( S2 Fig ) . We also performed additional simulations to evaluate the proposed computational procedures for enrichment testing and fine mapping analysis integrating genomic features . In brief , we found that the proposed procedures have greatly improved power in testing enrichment of genomic features when compared with standard approaches based on single SNP testing results ( S2 Table ) . In addition , we observed that the proposed computational procedure provides accurate estimates of the enrichment parameters ( S3 Fig ) . In this paper , we focused on analyzing the expression and genotype data collected from the GEUVADIS project [20] . More specifically , the data set consists of RNA-seq data on lymphoblastoid cell line ( LCL ) samples from five populations: the Yoruba ( YRI ) , CEPH ( CEU ) , Toscani ( TSI ) , British ( GBR ) and Finns ( FIN ) . In our analysis , we selected 420 samples which were densely genotyped in the 1000 Genomes Phase I data release [27] and 11 , 838 protein coding genes and lincRNAs that are deemed expressed in all five population groups . Throughout , our analysis focused on the SNPs that locate within a 200kb genomic region centered at the transcription start site ( TSS ) of each gene . In contrast to the original eQTL mapping approach discussed in [20] , we treated each population as a single group and performed cis-eQTL analysis jointly across all five groups . We have made the full analysis results available on the website http://www-personal . umich . edu/~xwen/geuvadis/ .
We have presented an integrative analysis framework to perform fine mapping analysis jointly from cross-population samples while incorporating functional annotations . Our core statistical methods are built on and naturally extend the established Bayesian framework for association analysis of genetic data from heterogeneous groups [14 , 15 , 21] . The most notable features of our statistical framework are its ability of multi-SNP analysis and integration of SNP-level genomic features . Through simulation studies , we have demonstrated the power and efficiency of the proposed approaches . For the first time , we have applied this framework for multiple cis-eQTL fine mapping analysis in a cross-population meta-analytic setting using the GEUVADIS data . Our analysis reveals that with an aggregated sample size of around 400 , multiple independent cis-eQTL signals can be confidently identified in many genes , which makes evident the necessity of considering multiple association signals in eQTL studies . The commonly applied strategy for mapping additional association signals is to conduct conditional analysis , which can be viewed as a step-wise variable selection algorithm . Besides lack of power , one additional disadvantage of conditional analysis is that it only reports a single “best” model in the end , and completely ignores its uncertainty . From many of our examples shown in this paper , it is clear that in most cases there is typically a great deal of uncertainty in determining the truly causal SNPs due to LD , and the posterior probabilities of the “best” models are often unimpressive . Consequently , it is generally inappropriate to solely rely on the information from the “best” model in the downstream analysis ( as we have illustrated in the power simulations of enrichment testing ) . In comparison , in addition to gaining power in identifying multiple independent cis-eQTL signals , our Bayesian approach provides much more comprehensive information that fully conveys the uncertainty of the inference result , and the quantified uncertainty information is naturally propagated in our integrative analysis of genomic features . In this paper , we have employed a two-step procedure that first screens eGenes by performing gene-level hypothesis testing , and then carries out multi-SNP analysis for the identified eGenes . This procedure is analogous to the fine mapping procedure that is commonly used in GWAS , where interesting loci are ranked and selected by single SNP association testing before an in-depth analysis focusing on each flagged high priority locus . We find this procedure not only yields considerable computational savings , but also provides a sound argument to specify the prior inclusion probability Pr ( γj ) . It should be noted that our proposed analysis procedures are completely applicable for fine mapping of general QTLs in either a meta-analytic or single study setting . They can be further extended to the applications of fine mapping analysis where subgroups of eQTL data are formed by different tissues [4 , 14] or cellular conditions [6 , 8] . Comparing to the meta-analytic setting considered in this paper , the parameter space of {γj} in those applications is more complicated ( which includes 2s potential values , where s is the number of subgroups/tissues ) . Nevertheless , [14] has provided a principled way to “learn” the priors on possible values that γj can take by pooling information across genes through a hierarchical model . Our fine mapping results of eQTLs also demonstrate the benefit of utilizing cross-population samples in genetic association studies . Most importantly , the population heterogeneity of local LD patterns serves as an efficient filter that narrows down the regions harboring casual eQTLs . Nevertheless , varying LD patterns can cause some SNPs to display large degree of heterogeneity across populations in their estimated effect sizes from single SNP analysis: in the extreme cases , a SNP may appear to possess strong “population specific” effects . As we acknowledge that genuine population specific eQTLs are certainly interesting phenomena and very much likely exist , we suggest interpreting highly heterogeneous eQTL signals from single SNP analysis with caution . In our view , it may be necessary to carry out multi-SNP analysis , as we have demonstrated in this paper , to simply rule out the possibility that the seemingly population specific effects are artifacts due to varying LD patterns . Our Bayesian inference framework naturally incorporates functional annotations in fine mapping eQTLs across population groups . This feature allows us to quantitatively assess the enrichment of certain functional features in eQTLs , and in turn to use the quantified enrichment information to prioritize annotated SNPs for fine mapping analysis . Overall , our model for integrative eQTL mapping analysis is similar to those presented in [17 , 18] . However , these previous approaches make simplifying assumptions to restrict at most one cis-eQTL per gene , such that single SNP association results can be directly used . Our method relaxes this assumption and is fully integrated into our multi-SNP analysis procedure . In addition , our use of CENTIPEDE annotation to examine the relative enrichment of binding variants and footprint SNPs is also novel . Although it is largely expected that binding variants are enriched in eQTLs , it is important to note that the level of enrichment for footprint SNPs is much lower than that for binding variants . Interestingly , this finding seems concurring with the results reported by [30] where the relative enrichment of binding variants vs . footprint SNPs in other cellular and organismal phenotype QTLs is examined . Last but not least , our multi-SNP fine mapping analysis of the GEUVADIS data has created a comprehensive resource for the community to gain better understanding of the genetic basis of gene regulation . With proper uncertainty assessments , our results enable follow-up experimental validations and functional studies of causal genetic variants that alter gene regulation . They also provide a unique and powerful resource to study population genetics of expression traits . For example , we found that the distribution of Fst of eQTL signals consistent across populations has a shorter tail than the distributions of random SNPs ( Kolmogorov-Smirnov test p-value = 0 . 001 ) . Nevertheless , when examining the Fst distributions between primary and secondary cis-eQTL signals ( stratified by the association strength in single SNP analysis with respect to their target genes ) , we did not find statistical evidence to differentiate the two ( Kolmogorov-Smirnov test p-value = 0 . 77 ) . There are many interesting aspects to be explored with our analysis results , and we will leave the more in-depth investigations for our future work .
The genotype and RNA-seq data were downloaded from the GEUVADIS project website . We selected 420 samples whose genotypes are directly measured in the 1000 Genomes project . The samples are evenly distributed in the five population groups , and the detailed breakdown of samples by population is shown in Table 1 . For the RNA-seq data , we used a slightly more stringent threshold than the original analysis [20] to select genes that are expressed in all five populations . Specifically , for each selected gene , we required > 90% individuals in each population group to have RPKM ≥ 0 . 1 . From the 17 , 361 Ensembl genes that passed this filter , following the original analysis , we selected a subset of 11 , 838 genes consisting of annotated protein-coding genes and lincRNAs according to GENCODE [31] release 17 . We log transformed the RPKM values , and used the pipeline employed in [29 , 32] to remove the effect of GC content on expression measurements . We then followed the same strategy as described in [20] to remove latent confounding factors using the software package PEER [33] . However , unlike the original analysis , we ran PEER for each population group separately . In the end , we removed 15 , 13 , 15 , 20 and 20 PEER factors for samples from YRI , CEU , TSI , GBR and FIN , respectively . Finally , the expression levels of each gene were quantile normalized across individuals separately in each population group . For genotype data , we filtered out SNPs whose sample allele frequencies < 0 . 03 in the overall samples across population groups . Note , we did not apply the allele frequency filter in each population group . The SNPs passing this filter must have sample allele frequencies ≥ 0 . 03 in at least one population group . In general , as discussed in the Results section , rare SNPs do not impose any statistical or computational problems for our analysis . Nevertheless , removing SNPs that are not likely informative in any population group helps improve computational efficiency . Following [17 , 32] , we defined a 200kb cis region for each gene centered at its TSS . In total , the final data set contains 6 . 7 million gene-SNP pairs . For each gene , we tested the null hypothesis that asserts no cis-eQTLs . Specifically , we adopted the Bayesian hypothesis testing procedure discussed in [14] . Essentially , [14] assumes a Bayesian model that is mostly similar to ( Eq 1 ) , except for an additional simplifying assumption , “at most one cis-eQTL per gene” . Given a gene with p cis-SNPs , this additional assumption reduces the possible alternative scenarios into p single SNP association models , for which a gene-level Bayes factor can be easily computed analytically by Bayesian model averaging . In the context of eQTL mapping in multiple tissues , [14] considered all possible configurations , i . e . , γj values , for each assumed associated SNP , whereas in our analysis of the GEUVADIS data we only allowed γj ∈ {0 , 1} to focus on discovering population consistent cis-eQTLs . Briefly , for the alternative model where the j-th SNP is assumed the lone eQTL , the linear model ( Eq 1 ) is simplified to y i = μ i 1 + β i , j g i , j + e i , e i ∼ N ( 0 , σ i 2 I ) , i = 1 , … , s . ( 7 ) We modeled the correlation of genetic effects , βi , j’s across population groups through the following prior specification , β i , j ∼ N ( β ¯ j , ϕ 2 ) , β ¯ j ∼ N ( 0 , ω 2 ) . ( 8 ) Equivalently , the above prior can be represented by a multivariate normal distribution by integrating out the average effect parameter β ‾ j , i . e . , β j = ( β 1 , j ⋮ β s , j ) ∼ N ( 0 , W ) , ( 9 ) where the s × s variance-covariance matrix W is given by W = ( ϕ 2 + ω 2 ω 2 ⋯ ω 2 ω 2 ϕ 2 + ω 2 ⋯ ω 2 ⋮ ⋮ ⋱ ⋮ ω 2 ω 2 ⋯ ϕ 2 + ω 2 ) . The parameter ϕ2 + ω2 characterizes the overall genetic effects of SNP j , and ϕ2/ ( ϕ2 + ω2 ) represents the degree of heterogeneity across population groups . Following [14 , 15] we considered the values of ϕ2 + ω2 from a set E = {ϕ2 + ω2 : 0 . 12 , 0 . 22 , 0 . 42 , 0 . 82 , 1 . 62} which covers a wide range of plausible magnitude of genetic effects . We allowed limited degree of heterogeneity by taking ϕ2/ ( ϕ2 + ω2 ) values from the set H = {ϕ2/ ( ϕ2 + ω2 ) : 0 , 0 . 1} which reflects our prior belief that effects of genuine eQTL signals should be highly consistent across population groups . Overall , we considered a combination of |E| × |H| grid for ( ϕ2 , ω2 ) values for each alternative model . Given this model , a single SNP Bayes factor , BFj , can be analytically evaluated following [14 , 15] , and the corresponding gene-level Bayes factor is obtained by Bayesian model averaging as BF gene = 1 p ∑ j = 1 p BF j . ( 10 ) Upon obtaining the gene-level Bayes factors , we used the methods implemented in the software package eQTLBMA [14] to select eGenes at FDR 5% levels . eQTLBMA implements two types of FDR control procedures: one is a permutation based procedure which converts a gene-level Bayes factor to a corresponding p-value and control FDR using Storey’s procedure; the other procedure is based on the EBF procedure discussed in [34] which directly works with gene-level Bayes factors and avoids any permutations . We found that the latter approach is much more computationally efficient , however slightly conservative . The results reported in the Results section are based on the EBF procedure . The power gain of the Bayesian gene-level testing procedure in identifying consistent association signals across multiple populations/subgroups and its comparison with commonly applied frequentist approaches have been fully demonstrated in [14] and [21] through simulations and real data analysis . Here , we examined a few known factors that may potentially affect the power of the gene-level analysis in the GUEVADIS data . Using the hierarchical model proposed in [14] , we performed the analysis described in [21] to estimate the effect size heterogeneity of eQTL signals across populations: our choice of the heterogeneity parameters is mostly informed by this analysis . More specially , we used a full spectrum of grid values for the heterogeneity parameter ϕ2/ ( ϕ2 + ω2 ) from a comprehensive set H = {ϕ2/ ( ϕ2 + ω2 ) : 0 . 0 , 0 . 1 , 0 . 2 , … , 0 . 9 , 1 . 0} . We then estimated the probability weight on each grid value by pooling information across all genes using the EM algorithm implemented in eQTLBMA . The estimates show that the majority of the probability mass is concentrated in very low heterogeneity levels ( S1 Table ) , and the mean heterogeneity E[ϕ2/ ( ϕ2 + ω2 ) |Y , G] = 0 . 060 , i . e . , on average , the eQTL effects are very consistent across populations . This result is also highly concordant to the findings in [21] where eQTL data from microarray experiments across European , African and Asian populations were examined , and further justifies our selection of heterogeneity parameter values in the analysis . Among other factors impacting eGene discoveries , we noted that the number of cis-SNPs is negatively correlated with the gene-level local false discovery rate ( lfdr ) . That is , on average , genes harboring more cis-SNPs tend to exhibit stronger gene-level association signals . The Spearman’s rank correlation between the two quantities is modest ( -0 . 20 ) , nonetheless statistically highly significant ( p-value < 2 . 2 × 10−16 ) . In our multi-SNP analysis , we no longer assumed “one cis-eQTL per gene” , and considered the full range of alternative scenarios described by model ( Eq 1 ) . To make joint inference with respect to Γ = {γi , … , γp} , we further specified effect size distribution , β j | γ j = 1 , ϕ 2 , ω 2 ∼ N ( 0 , W ) , ( 11 ) where W is constructed in the same way as in the gene-level analysis . In particular , we used the same set of grid values of ( ϕ2 , ω2 ) in the gene-level analysis . Unconditional on γj , the prior on βj is a type of “spike-and-slab” , where the “slab” is represented by a mixture of multivariate normal distribution , and the vast majority of the prior probability mass , 1 − 1 p , is assigned to the “spike” ( i . e . , a point mass at 0 ) . The described linear model system including the prior specification ( i . e . , SSLR ) , is a special case of the general system considered by [15] . Given a specified Γ value , a Bayes factor , BF ( Γ ) , contrasting to the trivial null model , Γ ≡ 0 , can be analytically approximated by applying the result discussed therein . With this result , the posterior probability , Pr ( Γ|Y , G ) , can be computed up to an unknown normalizing constant , i . e . , Pr ( Γ|Y , G ) ∝ Pr ( Γ ) BF ( Γ ) . We implemented a Metropolis-Hastings algorithm , similar to what is discussed in [15] for multivariate linear regression model ( MVLR ) , to efficiently traverse the space of 2p possible Γ values . In particular , we designed a novel proposal distribution that utilizes marginal and conditional analysis results to prioritize SNPs with strong marginal or conditional association signals . In practice , we observed that the resulting MCMC algorithm achieves fast mixing . The details of the algorithm is provided in S1 Text ( section S . 1 ) . In the end , the MCMC algorithm yields a set of Γ samples from the posterior distribution , from which we computed the PIP for each SNP by marginalization . The posterior expected number of independent cis-eQTLs and its variance for each gene are obtained by computing the sample mean and variance of the number of non-zero γj’s in each posterior model . Equivalently , the posterior expected number of cis-eQTLs can be computed by the sum of PIPs , i . e . , E [ ∑ j = 1 p 1 ( γ j ≠ 0 ) | Y , G ] = ∑ j Pr ( γ j = 1 | Y , G ) . ( 12 ) For the fine mapping analysis of the GEUVADIS data , we applied the MCMC algorithm for each identified eGene individually . We carried out 25 , 000 burnin steps and 50 , 000 repeats for each MCMC run . Taking advantage of parallel processing , we performed multiple-SNP analysis for multiple genes simultaneously in a distributed computing environment , which greatly reduced the overall computational time . The MCMC run for the eGene with the most cis-SNPs , HLA-DRB1 ( Ensembl ID: ENSG00000196126 , with 11 , 400 cis-SNPs ) , took 30 minutes on a computer with a single Intel Xeon 2 . 13GHz 8-core CPU . For average eGenes with ∼ 2 , 000 cis-SNPs the running time is approximately 3 to 4 minutes on the same machine . We ran our fine mapping analysis for GEUVADIS data with 8 parallel threads , the overall computation took about 30 hours . Although the MCMC output conveys the full information of the fine mapping results , parsing the information manually to identify independent eQTL signal clusters can still be non-trivial . To tackle this problem , we designed and implemented a simple hierarchical clustering based algorithm that automatically parses the MCMC output and aides identifying independent eQTL signal clusters . The details of the algorithm are described in the S1 Text ( section S . 4 ) . To test the enrichment of a particular genomic annotation in eQTL signals , we examine the associations between SNP level annotations of interest and the PIPs from our multi-SNP fine mapping approach computed under the null hypothesis of no enrichment . Without loss of generality , we consider a single annotation and use δ j g to denote the annotated value of SNP j in gene g ( the additional super-script for gene emphasizes the annotation is specific to each gene-SNP pair ) . Our testing procedure starts by computing PIPs for all gene-SNP pairs using the proposed multi-SNP cis-eQTL mapping method with the exchangeable prior specification ( Eq 4 ) . Under the null hypothesis , the resulting PIPs should be independent of the annotation information . We then fit a logistic regression model logit [ Pr ( γ j g | Y , G ) ] = α 0 + α 1 δ j g , ( 13 ) and perform the Wald test with respect to α1 . The commonly applied enrichment testing approach typically classifies the latent association status of each SNP ( γ j g ) into 0 or 1 depending on the single SNP association test results ( usually p-values ) , and then test their associations with the annotations of interest . In comparison , the use of PIPs in our proposed method presents at least two obvious advantages: first , it naturally accounts for potentially multiple association signals within a single gene; and second , it fully accounts for the uncertainty in determining γ j g which is ignored by the binary classification approach . These advantages are reflected by the gain of powers in enrichment testing . To incorporate genomic annotations into the fine mapping analysis of cis-eQTLs , we specify the prior distribution of Pr ( γj ) by the parametric function ( Eq 6 ) , whereas the other parts of the Bayesian multi-SNP analysis model remain intact . For the g-th gene and its j-th SNP , we re-write ( Eq 6 ) using an equivalent vector form as follows logit [ Pr ( γ j g = 1 ) ] = α ′ δ j g , ( 14 ) where the enrichment parameter α is assumed to be shared across all gene-SNP pairs . Let Dg denote the collection of the annotation data for gene g . For a total of q genes , it follows from the Bayes rule that Pr ( α , Γ 1 , … , Γ q | { Y 1 , G 1 , D 1 } … { Y q , G q , D q } ) ∝ P ( α ) ∏ g = 1 q [ Pr ( Γ g | D g , α ) BF ( Γ g ) ] . ( 15 ) Given a prior specification P ( α ) ( e . g . a flat prior ) , the quantities on the right hand side are individually straightforward to compute analytically . It is conceptually easy to modify our MCMC algorithm for the multi-SNP analysis to jointly sample ( α , Γ1 , … , Γq ) . Nevertheless , due to the extremely high dimensionality of the target space ( which is approximately the number of gene-SNP pairs , in the case of the GEUVADIS data the number is ∼ 6 . 7 million ) , the convergence of the MCMC within a reasonable time frame may be in doubt . Furthermore , the computational resources , especially the memory usage , demanded by the MCMC algorithm may be too high to afford in a practical setting . Alternatively , we consider an empirical Bayes approach which employs an EM algorithm to find the MLE of α by treating ( Γ1 , … , Γq ) as missing data , and perform a final round of multi-SNP analysis ( possibly only for eGenes ) conditional on the resulting MLE α ^ . The derivation of the EM algorithm is mostly straightforward , we give the relevant details in S1 Text ( section S . 2 ) . Briefly at ( t+1 ) -th iteration , in the Expectation step ( E-step ) , we evaluate E [ γ j g | Y g , G g , α ( t ) ] for each gene-SNP pair given the current estimate α ( t ) . It should be noted that this conditional expectation is exactly the PIP of each gene-SNP pair which can be obtained by running the multi-SNP analysis for each gene separately given the hyperparameter α ( t ) . In the Maximization step ( M-step ) , we find α ( t+1 ) by fitting a logistic regression model relating PIP of each gene-SNP pair to the genomic annotations of the corresponding SNP . Overall , we describe the complete algorithm as an “MCMC-within-EM” algorithm , which is initiated at some arbitrary α ( 0 ) value , and iteratively performs multiple cis-eQTL mapping using the MCMC algorithm and maximization by fitting logistic regression models until convergence . The main computational benefit of the “MCMC-within-EM” algorithm is that the E-steps involving MCMC runs can be processed in parallel on a distributed computing system , hence the memory requirement is much relaxed . To illustrate the procedure for estimating α , we simulated a set of eQTL data using the same scheme in the enrichment testing . In particular , we used a single binary annotation and set α1 = 1 . 50 . We ran the MCMC-within-EM algorithm on this simulated data set with the starting point α1 = 0 , and plotted the estimated α1 values in each iteration in S3 Fig . The figure indicates that the estimates of α1 quickly converge to the close neighborhood of the true value after 6 to 7 iterations . To reduce computation , in practice we analyzed the GEUVADIS data using only a single iteration of EM where all enrichment parameters are initialized at 0 . As a consequence , our reported point estimates of the enrichment parameters are likely to be lower bounds of the true values . Nevertheless , as we have shown in the analysis of the GEUVADIS data , this simplified version of the algorithm still improves the resolutions of the fine mapping analysis by up-weighting the SNPs with relevant annotations . | Expression quantitative trait loci ( eQTLs ) are genetic variants associated with gene expression phenotypes . Mapping eQTLs enables us to study the genetic basis of gene expression variation across individuals . In this study , we introduce a statistical framework for analyzing genotype-expression data collected from multiple population groups . We show that our approach is particularly effective in identifying multiple independent eQTL signals that are consistently presented across populations in the proximity of a gene . In addition , our analysis framework allows effective integration of genomic annotations into eQTL analysis , which is helpful in dissecting the functional basis of eQTLs . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2015 | Cross-Population Joint Analysis of eQTLs: Fine Mapping and Functional Annotation |
Vascularisation is a key feature of cancer growth , invasion and metastasis . To better understand the governing biophysical processes and their relative importance , it is instructive to develop physiologically representative mathematical models with which to compare to experimental data . Previous studies have successfully applied this approach to test the effect of various biochemical factors on tumour growth and angiogenesis . However , these models do not account for the experimentally observed dependency of angiogenic network evolution on growth-induced solid stresses . This work introduces two novel features: the effects of hapto- and mechanotaxis on vessel sprouting , and mechano-sensitive dynamic vascular remodelling . The proposed three-dimensional , multiscale , in-silico model of dynamically coupled angiogenic tumour growth is specified to in-vivo and in-vitro data , chosen , where possible , to provide a physiologically consistent description . The model is then validated against in-vivo data from murine mammary carcinomas , with particular focus placed on identifying the influence of mechanical factors . Crucially , we find that it is necessary to include hapto- and mechanotaxis to recapitulate observed time-varying spatial distributions of angiogenic vasculature .
The role of angiogenesis—the process whereby existing blood vessels produce new vasculature—in cancerous growth , invasion and metastasis has been extensively studied over the past five decades . Starting with the assertion of Folkman [1] that angiogenesis is a necessary component for neoplasmic growth , the current paradigm is that tumours induce neo-vascularisation upon reaching an avascular limit [2] . This limit represents a critical tumour size that can be supported by oxygen diffusion from the existing vasculature alone , beyond which substrate gradients produce internal regions of oxygen deprivation , i . e . hypoxia . To avoid necrosis , cells in the hypoxic regions secrete diffusible chemical signals , termed tumour angiogenic growth factors ( TAFs ) . Upon reaching the existing vasculature , the TAFs stimulate endothelial cells ( ECs ) to degrade their basement membrane and extracellular matrix ( ECM ) via the secretion of matrix metalloproteases ( MMPs ) [1] . Motile ECs then migrate from the vessel lining up the TAF gradient field towards the TAF source , forming tubes with sprout-tips at the leading edge of a new vascular lumen . These tubes can form networks in a process termed anastomosis and penetrate into the tumour , depending on how the ECs respond to mechano-chemical factors . As blood flows through the vessels and remodels their structure through its response to fluid shear stresses and vascular pressure , the tumour is provided with a direct supply of nutrients and oxygen , enabling further expansion into the surrounding tissue . The neo-vasculature , however , is pathological and during tumour growth its structural integrity can be compromised . Indeed solid stresses in the tumour are elevated as a consequence of rapid growth into the confined space of the host tissue , which can compress and ultimately collapse intra-tumoural blood vessels , rendering tumours hypo-vascular and hypo-perfused [3 , 4] . Hypo-perfusion , in turn , has been shown to inhibit the delivery of chemotherapy , reducing drastically treatment efficacy [5 , 6] . Furthermore , a key factor in angiogenic tumour growth is cell response to mechano-chemical cues . Of particular interest is their directed motion along chemical gradients , termed chemotaxis and haptotaxis for soluble ( e . g . oxygen ) and insoluble ( e . g . proteoglycan ) substrates , respectively , and mechanical gradients ( e . g . solid stresses ) , termed mechanotaxis . This produces a biophysical system with multiple components interacting at multiple scales . In order to study the effect of a given component and characterise its physical origins , it is instructive to construct physiologically-representative mathematical models . Here we focus on continuum models of tumour growth coupled with angiogenesis; for more detail on angiogenesis modelling alone , see the recent review by Scianna et al . [7] . Whereas previous studies have explored the chemical , i . e . solute-driven , underpinning of both tumour growth and angiogenesis , here we focus on the interplay between angiogenic network evolution and growth-induced solid stress generation . Broadly speaking , the tumour and vasculature can each be described by continuum , discrete or hybrid models , with their coupling either static or dynamic . One approach is to characterise the angiogenic response in terms of the blood vessel density , with the dynamics and chemical factors obeying continuum conservation laws . Prominent early examples are [8 , 9] , which extended the seminal work of Balding and McElwain [10] to define a set of coupled integro-differential equations that characterise tumour-induced neo-vascularisation and network formation . More recently in [11] , the authors utilised a continuum approach to investigate the role of feedback regulation processes on sprout inception . However , the blood vessel density paradigm is unable to account for vascular morphology and its explicit impact on blood flow heterogeneity . In order to characterise network morphology and blood flow , it is necessary to model the vasculature discretely in terms of line segments , curves or lattices . This approach was adopted by Zheng et al . [12] , who employed a hybrid model to describe the vasculature [13] coupled with a nonlinear continuum description of the tumour mass [14] . The model simulated physiologically-realistic tumour morphologies as a result of a static coupling with angiogenic growth , but was limited in its description of tumour-environment interactions and mechanical factors . This model was later extended by Macklin et al . [15] to include dynamic angiogenesis , allowing for an explicit description of vascular remodelling and blood flow , and further extended by Wu et al . [16] to model the effects of interstitial fluid pressure . A contrasting approach was proposed by [17] , who presented a solid mechanics description of the tumour and its environment . Their model featured a dynamic coupling of angiogenesis and vascular remodelling with a growing domain , and accounted for deformation due to growth , response to hypoxia and blood flow . However , there was no account of the effect of solid stress on vascular development and integrity inside the tumour . To account for the effect of haptotactic ( i . e . insoluble ) vascular endothelial growth factors ( VEGF ) , Milde et al . [18] presented a deterministic , hybrid model of sprouting angiogenesis . The matrix-bound VEGF was cleaved by MMPs produced at the endothelial tip cells , and the response characterised in terms of the resulting vessel geometries . Furthermore , they included the effect of ECM fibre density and structure on the sprout tip migration velocity . They found that high density matrices produced shorter capillary networks , and observed an increase in the number of branches with the density of matrix-bound VEGF . Similar results were obtained by Bauer et al . [19] , who developed a cell-based model of tumour-induced angiogenesis . In both models , however , there was no account of mechanotaxis or the effects of vessel wall collapse due to applied solid stresses . Previous studies by Breward et al . [20] and Bartha et al . [21] have presented models describing the effect of pressure on vessel integrity , but no attempt is made to explicitly model vessel compression owing to intratumoural forces . To our knowledge there is no existing model that dynamically couples capillary growth , morphology and structure with mechanochemically-regulated blood flow and growth-induced solid stresses . Numerical simulations of angiogenic tumour growth are performed for spheroidal geometries , which we use to gain insight into the relationship between mechano-chemical factors and tumour development . Key measures of vascular development , defined in the following section , are validated using data from in-vivo murine mammary carcinomas [22] . The hypotheses we test are: One part of the novelty of the mathematical formulation lies in the introduction of a mechanotactic term to the function defining the orientation of capillary tip elongation . The second part is the use of a phenomenological description of the tip extension velocity . The final part is a set of constitutive terms defining the capillary wall remodelling as a function of mechanical factors . These features furnish the model with a complete , if simplified , description of the biophysical factors influencing tumour-induced angiogenesis . Specifically , a discrete three-dimensional model of vascular sprouting is employed to describe the angiogenic response to TAF secretion , where the sprout-tips are represented as point masses in a continuum substratum [23] . The secretion of TAFs and MMPs by the tumour and vasculature , respectively , are described by coupled reaction-diffusion equations [24] . Tumour growth is modelled according to a Gompertz-type relation derived in previous work [25] , and the quasi-static linear momentum equation is solved at the macroscopic scale assuming hyperelastic material properties . Capillary elongation , branching and remodelling are made dependent upon mechanical factors , such as traction and magnitude of wall shear stress . Finally , similar to Stylianopoulos and Jain [26] , intra- , trans- and extravascular fluid flow is described by Poiseuille’s , Starling’s and Darcy’s laws respectively . To our knowledge , we present here for the first time a validated , three-dimensional , tumour-induced angiogenesis model using published in-vivo data . These data include the vascular density and structure obtained from image analysis of MCaIV carcinomas . To constrain the model , ex-vivo measurements of MCaIV carcinoma material properties were used as input parameters . Where possible all other input parameters were specified according to either in-vivo or in-vitro data from the literature .
One of the first steps in angiogenesis is the production of diffusible angiogenic growth factors ( such as VEGF , PDGF , etc . ) by tumour cells , and their subsequent binding to corresponding receptors of nearby blood vessels [38] . For model simplicity , we focus on a single growth factor—referred here as the tumour-angiogenic factor ( TAF ) —as a homogenised chemical modulator of capillary sprouting and elongation . It would be straightforward to extend this model to account for multiple interacting growth factors ( e . g . VEGF binding receptors and inhibitors ) to represent the complex patho-physiology of tumour-induced angiogenesis . However , given that it is well-documented that interstitial fluid velocities are weak compared to vascular flow and diffusion [39] , we have not included advection in the tissue biochemical transport model . Transport of TAF ( with normalised concentration denoted τ ) is described by a reaction-diffusion equation [15] that accounts for random spatial diffusion , TAF production as a function of the normalised local oxygen concentration , ξ , in the tumour ( defined below ) , and its natural decay and finally loss due to cellular consumption , d τ d t = ∂ ∂ X · D τ ∂ τ ∂ X + Q - δ τ τ , ∀ X ∈ Ω , ( 12 ) where Dτ is the isotropic diffusion coefficient for TAF ( given in m2 day-1 ) , δτ represents the aggregate loss due both decay and cellular consumption , the production rate Q ( given in day-1 ) is defined by Q ( ξ ) = λ τ exp - 2 ξ / ξ ¯ ; if X ∈ Ω T 0 ; elsewhere , where ξ ¯ is a scaling parameter that modulates the oxygen level at which cancer cells release TAF , and λτ is a TAF-production rate parameter . Parameter values are provided in S3 Table , including references to the relevant literature . The microvascular network provides a uniform source of oxygen , which diffuses into the interstitial space and is consumed by the cells [28] , such that d ξ d t = ∂ ∂ X · D ξ ∂ ξ ∂ X + λ ξ ρ v - δ ξ ξ , ∀ X ∈ Ω , ( 13 ) where Dξ is the isotropic diffusion coefficient for oxygen ( given in m2 day-1 ) , λξ is the oxygen production rate due to supply from the microvascular network and δξ is the consumption rate of the species by the cancerous cells ( both expressed in day-1 ) . Here , the dimensionless parameter ρv represents the average vascular density in a tissue FE; if fully-functional , well-perfused blood vessels are present in a particular element then ρv = 1 , whereas if the vessels are hypo-perfused or non-functional ( i . e . collapsed ) or absent then ρv = 0; see Fig 1B and 1C for a visual guide . Definition and differentiation of a well-perfused from a hypo-perfused node of the vascular network is explicitly provided in the Capillary wall remodelling subsection , while definition of a collapsed vascular node is given further in the Interaction between the tumour–host biomechanics and the capillaries subsection . Finally , the concentration of the matrix-degrading enzymes: MMPs , μ , in the extracellular space of the host tissue is modelled through [15 , 40] d μ d t = ∂ ∂ X · D μ ∂ μ ∂ X + F - δ μ μ , ∀ X ∈ Ω H . ( 14 ) Here Dμ ( m2 day-1 ) is the isotropic diffusion coefficient for MMPs , while δμ ( day-1 ) represents the natural decay of μ . The production rate function F of MMPs ( day-1 ) is taken as the superposition of contributions of production from the proliferating cancer cells and the tip-endothelial cells , defined as: F = λ μ -c + λ μ -v ρ t ; if X ∈ Ω T λ μ -v ρ t ; elsewhere . In a similar fashion to ρv , the dimensionless variable ρt represents the density of tip vascular nodes in a finite element of the tissue domain , which is zero when no newly-formed sprouts are present and increases proportionally with the number of branches present ( see Fig 1 ) . Evidently , both ρv and ρt vary with respect to space since not all finite elements contain vascular segments or tip vessels , and also with respect to time due to the temporal evolution of the network structure as it undergoes angiogenesis . As is evidenced by Eqs ( 12 ) , ( 13 ) and ( 14 ) , the diffusion coefficients are assumed to be constant and homogeneous everywhere ( see S3 Table for the parameter values ) . The mathematical model of tumour-induced angiogenesis hence assumes linear , Fickian diffusion of the chemical species . A description of the dynamic angiogenesis model is presented here . The model is decomposed into to two primary components: a model of capillary sprouting , and a model that couples capillary wall mechanics with growth-induced solid and fluid mechanical loads . Using Fig 1A as reference , the solid solver boundary conditions are defined as follows: traction-free ( S ⋅ n = 0 , where n is the outward surface normal ) on the outer surfaces without inlets or outlets , ΓV; zero displacements ( u = 0 ) on the outer surfaces with inlets or outlets , ΓT ( to avoid rigid body motion ) ; continuity of stress and displacement on the tumour–host interface , ΓI . No residual stresses are considered in the present analysis , hence the initial tissue deformation is zero ( Fe = Fg = I ) . In order to replicate an initially healthy host environment ( i . e . prior to angiogenic tumour development ) , a uniform distribution of capillaries having approximately uniform diameter , thickness and pore size is imposed initially ( see S3 Table ) , whereas the tumour environment is initially avascular with the mean inter-capillary distance being 0 . 6 millimetres approximately . The assumption of initial-state parallel , unbranched vessels is based on previous pertinent work in tumour-induced angiogenesis [13 , 15] . For the fluid solver , 0 . 1 mm-Hg interstitial fluid pressure was assigned on the tissue boundary ΓV according to [53] , and the interstitial fluid flux was assumed continuous at the tumour–host interface . The vascular pressure , pvsc , is prescribed at the inlet and outlet vascular nodes of the initial network as 25 mm-Hg and 10 mm-Hg , respectively [54] . Furthermore , throughout the analysis , pvsc = 0 is enforced at every node belonging to a collapsed vessel , in order to effectively model the obstruction of the natural flow of erythrocytes and plasma at that part of the vascular network . For the biochemical solver , zero-flux boundary conditions of all species are applied at the outer boundary of the tissue region , Γ ( and note the domain is chosen to be sufficiently large that this does not impact the solution in the main body of the tissue ) . Continuity of concentration and its flux are assumed at the tumour–host interface . The initial conditions are τ = ξ = μ = 0 everywhere except for the healthy tissue domain , where ξ = 1 . The initial condition for the ordinary differential Eq ( 1 ) describing the ECM density is ϵ = 1 in the entire region of the analysis domain . The problem under consideration is transient by nature and the modelling framework , as described in the previous section , consists of four interconnected core components , namely the Vascular Network Module , the Biochemical Solver Module , the Solid Solver Module and the Fluid Solver Module . Fig 3 illustrates the building blocks of the proposed modelling framework in a flow diagram and the interaction amongst them . The numerical procedure of the coupled tumour-growth and tumour-induced angiogenesis multiscale solver is outlined in the following section . The mathematical model identifies three different time-discretisation scales , one for each of the first three solver modules: In order to ensure that the explicit time integration scheme of the biochemical equations produces stable solutions , the Courant-Friedrichs-Lewy condition needs to be satisfied; given the smallest-element size in the 3D mesh , a very short time step has been adopted of the order of a few seconds . Similarly , to ensure that the nonlinear solid mechanics solver converges after a reasonable number of Newton-Raphson iterations ( e . g . four or five ) , the corresponding time increment has been set equal to approximately one hour . Finally , after parametrically investigating the results produced by the vascular network update module , time increment Δtv has been set to six hours . It is important to note here that using large values of Δtv can generate long , straight vascular segments , hence influencing the prediction of the vascular tree “shape” and organisation; whereas using very small Δtv can increase the computational cost substantially , since the generation of very small vascular segments will result in a rapid increase in the number of degrees-of-freedom in the final system of the fluid mechanics solver . In light of the model variables ρv and ρt , to ensure that a newly-formed vascular segments passes through ( and not crosses ) any tissue element ( see Fig 1 ) , the 3D mesh density of the analysis domain must be properly selected , as follows . The maximum elongation length from Eq ( 17 ) gives: ( vv-max Δtv ) = 0 . 0625 mm ( see material parameter S4 Table ) . Therefore , the minimum element size of the 3D mesh must be greater or equal to maximum elongation length—in the present analysis the minimum edge length ( at the centre of the analysis domain; see S1 Fig ) is 0 . 07 mm approximately . The algorithmic structure is as follows: The above procedure is repeated until the termination of tumour growth . Details about the FE implementation of the proposed tumour-induced angiogenesis and growth model are provided in the S1 File .
Here we present the validation of the model in terms of vascular and interstitial fluid pressure and velocity , and vascular density predictions against published data . To qualitatively validate the fluid solver predictions , Fig 4A presents the average interstitial fluid pressure , pint , distribution in the tumour and the peri-tumoural area for various time instants , where the centre of the tumour is at zero radial distance . Complementary to this plot , the significant variability of the IFP is shown in Fig 5 as a function of radial distance from the tumour centre for a number of time points . Interestingly , a rapid increase in the IFP is observed within one day , as a result of the formation and elongation of new and hyper-permeable capillary sprouts . This reflects the patho-physiological nature of such sprouts , which typically have a discontinuous endothelial lining and no basal membrane , rendering them hyper-permeable . Boucher and colleagues [56] measured the IFP for two rat tissue tumour types ( mammary adenocarcinoma and Walker 256 carcinoma ) using micro-pipettes . Comparing the numerical predictions , shown in Fig 4A , with the experimental results reported by Boucher et al . ( see Fig 3 in [56] ) , a strong qualitative agreement is observed in the drop of the IFP away from the tumour–host interface . The numerical predictions also show that the IFP reaches a maximum plateau throughout the tumour , which , despite the significant increase of tumour volume—from 19 . 06 mm3 in day-10 to 141 . 37 mm3 in day-40—remains relatively stable . The mean plateau IFP predicted by the model is approximately 8 . 3 mm-Hg . This falls within the experimentally measured pressure range reported previously for tissue-isolated ( 9 . 1±3 . 9 mm-Hg; see Table 2 in [56] ) and subcutaneous ( 7 . 8±3 . 8 mm-Hg; see Table 2 in [56] ) small-size tumours ( <1 g ) , while it is also in qualitative agreement with the IFP measurements [37] for MCaIV-type murine mammary carcinomas ( 5 . 6±1 . 2 mm-Hg; see Fig 3F in [37] ) . Further qualitative validation of the fluid solver was performed by examining the interstitial fluid velocity ( IFV ) , calculated from Darcy’s law ( see Eq ( 11 ) ) . Assuming that the tissue hydraulic conductivity is constant , the IFV is driven only by gradients of the IFP local to the tumour–host tissue . However , as illustrated in Fig 5 , the IFP remains relatively flat in the tumour region and decreases rapidly when moving from the tumour to the healthy tissue . Therefore , the IFV in the tumour centre is negligible compared to the corresponding pronounced IFV measured near the tumour boundary . This effect is visualised in S10 and S11 Videos: the IFV gradually increases in the tumour centre during angiogenesis , whereas abrupt IFV elevation is observed at the tumour periphery . The latter observation is made more evident in S12 and S13 Videos , where the IFV peak values are predicted primarily in the vicinity of collapsing blood vessels of the original network . Thus , drastic changes in the functionality of the tumour vasculature lead to loss of balance in the microvascular pressure distribution and the intra- and extravasation flux of plasma/proteins in the capillaries . This drives dynamic changes in the IFP , which subsequently result in larger IFP gradients and hence a larger IFV . As such , the above predictions confirm the hypothesis that there is significant interstitial hypertension in the tumour . The maximum estimated value of the fluid velocity in the interstitium was approximately 0 . 15 μm/s , which is in qualitative agreement with the experimentally measured values ( 0 . 6±0 . 2 μm/s ) reported in the early paper of Chary and Jain [58] . In Fig 4B , the magnitude of the mechanical forces in the stroma—i . e . the tissue hydrostatic pressure ( THP ) which is equal to one third of the trace of the stress tensor , also referred as the mean solid stress—is shown as a function of the radial distance from the centre of the tumour at several time points . Also Fig 6 reports the spatial distribution of THP at various time instants , where both the average value of the THP and the standard deviation ( in error bars ) is shown . In contrast to the fluid-pressure distributions in the interstitium shown in the Fig 5 , THP shows substantial spatio-temporal variability primarily at the interface of the tumour and the host tissue ( vertical red bar in Fig 6 ) . This can be explained by the anisotropic growth driven by the irregular angiogenic vasculature , which provides a non-uniform supply of oxygen to the tumour . These results support the idea of a heterogeneous force environment at the tumour periphery , consistent with the observations reported in the past [4 , 59] . Numerous features of our evolving vascular structures recapitulate experimental observations . First of all , the mean inter-capillary distance increased from around 0 . 557 mm to 0 . 771 mm at the end of simulation time . This is consistent with experimental measurements , for example in stage IIb and III carcinomas of the cervix ( measured using histochemistry , see [60] ) , where the average inter-capillary distance in the cancerous tissue was around 304±30 μm higher than that in healthy tissue . Secondly , our predictions of tumour vascular density ( defined as blood vessel surface area divided by the tissue volume , normalised against the corresponding healthy tissue value ) are reported in Fig 7 , for mechano- and haptotactic stimuli switched on or off . The normalised vascular density increases in time as the tumour grows , reaching around 3 . 4 and 3 . 0 with/without the mechano-/haptotactic stimuli by the end of the simulation . This is in good agreement with reported measurements [61–63] , which lie in the range 3 . 3 to 5 . 0 , with the inclusion of mechanical stimuli improving the agreement with measured observations . Vascular density and inter-capillary distance alone do not assess the functionality of the three-dimensional vasculature in its entirety . In particular , solute ( such as oxygen and drugs ) delivery is determined by diffusion distances from the vessels , and influenced by the spatial architecture and organisation . We consider the averaged power spectrum of the distance map computed on the vascular network , which was originally proposed by Risser et al . [64] in brain tumours ( and further used by Baish et al . [65] ) for assessing the diffusive capacity of drugs in normal tissue and tumours . We introduce two scaling parameters λv and δv-max , following the approach of Baish [65]; the first measures the shape of the space between blood vessels , whereas the second estimates the shortest distance between a tissue point and vessel . We compare here the model predictions of parameters λv and δv-max against the experimentally measured values , extracted by quantifying the vascular structure in murine breast carcinomas using in-vivo imaging . The images of the tumour vasculature were obtained in a previous research study using the optical frequency domain imaging [22] . However , brief description of the present analysis of the in-vivo images can be found in S2 File . The comparison of the two scaling parameters is summarised in Fig 8 . Model predictions of λv agree very well with the in-vivo data , and this agreement is improved with the mechano-/haptotactic stimuli switched on compared to off . This is an extremely promising validation step of our in-silico cancer model , which provides strong evidence that our inclusion of mechano- and haptotaxis is highly relevant to predicting and testing delivery of diffusible agents to vascular tumours . The analysis of the parameter δv-max is less conclusive . The trend offered by the experimental data points here is less well-defined ( see Fig 8B ) ; our model certainly predicts differing behaviours with the mechano- and haptotactic terms switched on/off , and there are currently insufficient experimental data points to draw a conclusion on which model prediction is correct . Non-isotropic growth is driven by the non-uniform spatial distribution of capillaries at the tumour periphery and hence—as described by Eq ( 3 ) and the relation defining Fg—the resulting oxygen distribution . In Fig 9 ( see also S1 video ) , the tumour is illustrated to grow due to the oxygen transcending from the vasculature and diffusing in the interstitial space . The visualisation also demonstrates dynamic network remodelling: not only do the vessels lack a regular hierarchy and structure , but they also display a higher degree of tortuosity than the initial ( healthy ) vascular network . In S2 Fig , the numerically predicted increase of the cancer mass as a function of time ( in days ) is illustrated . The in-silico cancer model enables a quantitative investigation of blood flow in the evolving vascular network . S5 fig compares the mean blood flow velocity with vessel diameter ( considering functional i . e . non-collapsed vessels only ) at different time points . Such information is extremely hard to measure accurately using experimental methods , given the requirement to image flow in individual microvessels of micron-level diameters; a validated in-silico framework is therefore highly valuable in providing insight into microvessel functional behaviours . Fig 10 shows the perfusion state of the vasculature as a function of time . The blood vessels are distinguished into functional and non-functional ( i . e . collapsed ) vessels , with the former being categorised according to their mean blood velocity ( MBV ) as hypo-perfused ( BMV <0 . 1 mm/s ) , perfused ( BMV in the range 0 . 1–0 . 5 mm/s ) and well-perfused ( BMV >0 . 5 mm/s ) ( this follows the convention proposed by Kamun et al . [48] ) . The proportion of hypo-perfused vessels increases quickly over the first 10 to 15 days , primarily driven by the unregulated vessel sprouting . After this time , the increase slows as a consequence of anastomosis and branching of vessels . To date , mathematical models of angiogenesis have mostly focused on chemotactic ( e . g . VEGF ) and haptotactic tip cell migration as determinants of vascular sprouting and branching ( key example papers [18 , 23 , 66–68] ) . These studies do not incorporate the impact of mechanotaxis in determining vascular topology , and both vascular and tumour growth . Here we investigate the hypothesis that vascular structure during tumour growth is influenced by chemo- , mechano- and haptotactic stimuli in combination . Specifically , we carry out simulations when chemotaxis alone determines the vessel tip extension rate and direction , and compare them against those where all three stimuli are included ( see Eq ( 15 ) ) . In Fig 7 we compare the normalised vascular density ( as defined previously ) as a function of time with chemotaxis in isolation , with the case of combined chemo- , hapto- and mechanotaxis . When only chemotaxis is active , the pathway of the tumour vessels is dictated by the gradients of TAF , where TAF—owing to the isotropic diffusion of these chemical cues in the ECM—is distributed spherically . Therefore , the elongation direction vector of the newly-formed sprouts points towards the tumour centre . However , owing to the growing tumour mass , solid stresses increase dramatically at the peritumoural stroma ( especially at day 10 and onwards; see Fig 6 ) , inducing collapse of the infiltrating nascent vessels when mechanical cues for vessel growth are included . Therefore , when chemo- , hapto- and mechanotaxis are combined , the pattern of angiogenesis changes significantly . Careful inspection of the simulation results reveals that the elevated mechanical forces at the peritumoural stroma re-direct the vessels to elongate circumferentially and , thus , increase the likelihood for the formation of anastomoses and vascular shunts . Notably , anastomoses are evident after day 10 owing to the growing population of microvessels and the significant increase of branches adjacent to the tumour . Summarising the above , and as shown in Fig 7 , our in-silico model predicts an increase of the normalised vascular density in the combined taxis case of angiogenesis . Also , Fig 8 compares the numerically predicted scalar parameters λv and δv-max in the two cases of taxis . The combined mode model produces higher normalised vascular densities , more consistent with experimental measurements , and also provides an excellent prediction of λv , as discussed in the previous section . This analysis indicates that hapto- and mechanotaxis may play an important role in determining the density and three-dimensional spatial arrangement of angiogenic vessels in tumours; in turn , these structural features of the vasculature are key in predicting diffusion of solutes ( e . g . oxygen , drugs ) into the interstitial space , and thus drug penetration and efficacy . Our model incorporates these mechanical stimuli , and could be used in the future to optimise tumour drug delivery and dosage . Furthermore , the model has predictive capability to characterise tumour solid stresses , and their interplay with tumour growth . Fig 11 illustrates in snapshots the in-silico predictions of tumour-induced angiogenesis when chemo- , mechano- and haptotaxis is taken into account ( images on the left of each column ) , and when chemotaxis applies only . Notably , from day 30 onwards , the tumour vessels follow a rather radial extension pattern in the pure chemotaxis case . However , when mechano- and haptotactic cues are also considered , the growing vessels are observed to encapsulate ( rather than penetrate ) the tumour . This “tumour framing” effect becomes more striking when the minimum TAF threshold required for angiogenesis , τ* , is lowered ( see following subsection ) . Also , particularly at days 35 and 40 , the formation of vascular shunts is also more pronounced when all three taxes ( referred to as the ‘combined mode’ ) are included , vessel tortuosity is increased , and the presence of anastomoses is also more frequent ( see also S14 Video ) . These are characteristic features of tumour vasculature , and it is highly promising that they are promoted by an in-silico model that includes mechanotaxis . Finally , the dependency of angiogenic vascular growth on various model parameters was examined . A sensitivity analysis of all parameters was performed and those with the largest influence on the normalised vascular density were identified , namely the TAF concentration and vascular wall stiffness . Fig 12A shows the normalised vascular density as a function of time for various values of the TAF threshold , ( τ* ) , above which angiogenesis is permitted ( all other parameters were kept constant ) . A non-linear dependency of the vascular density on τ* is observed , with a value of τ* = 0 . 02 producing a monotonic increase . This reflects the inherent non-linearity of TAF-induced vessel production , and suggests that the TAF threshold can quash other limiting factors—such as collapse from solid stress—which cause a decrease in vascular density above day 35 for the lower threshold values . This establishes TAF as a dominant factor in angiogenesis , which is in agreement with previous findings [69] . To test the response on vascular mechanics , Fig 12B shows the normalised vascular density as a function of time for various values of the maximum wall stiffness , Ew-max; all other parameters were kept constant . Specifically , in the baseline test 1 . 3 ≤ Ew-max ≤ 5 . 22 , which—as described in the previous section , and given that the critical radial strain a capillary can sustain is approximately 0 . 92—is equivalent to the critical pressure for capillary wall collapse 1 . 6 ≤ pc ≤ 3 . 4 . In the first sensitivity test the stiffness was increased by a factor of 10 , and in the second test by a further factor of 10 . This produced a highly nonlinear response , with a monotonic increase in the vascular density in time observed for both sensitivity tests . Considering Eq ( 21 ) , a capillary segment is assumed collapsed when the sum of the interstitial fluid pressure and the tissue hydrostatic pressure is at least pc-times larger than the microvascular pressure at this segment . As such , these results suggest that increasing pc by a factor of 10 renders the vessels essentially immune to collapse by external solid and fluid pressure . Fig 13 illustrates the impact of the microvascular wall stiffness in the predictions of the tumour growth and angiogenesis simulator ( see also the animation of S15 Video ) . It is evident from this figure that enabling vessel wall more resilient ( rightmost column of images ) , nascent vessels can resist and withstand the elevated mechanical forces at the tumour periphery . Thus , they can penetrate the tumour whose growth is speeded up and is slightly pronounced—provided that functional and well-perfused vessels supply the core of the cancer mass with vital nutrients and oxygen—as opposed to for example the baseline case ( leftmost column of images ) . In summary , varying the wall stiffness in our coupled model produces a similar effect on the vascular density predictions to that when varying the TAF threshold triggering angiogenesis . This work presents a validated three-dimensional mathematical and computational framework that encompasses solid tumour growth and tumour induced angiogenesis . The in-silico cancer model has been implemented in our in-house , open-source numerical platform FEB3 ( see for details S1 file ) . The proposed multiscale approach is capable of modelling the mechanical interactions between healthy and cancerous tissues , and associated vasculature , with the solid ( tissue ) and the fluid ( blood ) part of the tumour environment modelled separately . The important novel contributions of our model are: ( i ) the dynamic remodelling of the vascular network under mechanical stress during tumour growth , ( ii ) the incorporation of mechanotaxis ( alongside chemo- and haptotaxis ) in the determination of vessel sprouting orientation and speed , ( iii ) the collapse of tumour blood vessels as a consequence of solid stress produced by the surrounding tissue . The model recapitulates experimental observations of fluid and solid pressure ( and fluid velocity ) distributions , vascular density and three-dimensional spatial arrangement , with the improvement between experimentally measured and theoretically predicted vascular measurements increasing upon the inclusion of hapto- and mechano-tactic ( alongside chemotactic ) stimuli in the model . This supports our hypothesis that hapto- and mechanotaxis play an important role in determining the density and 3D spatial arrangement of tumour induced vasculature . As further data that quantify tumour vascular structures become available , we will continue to test the applicability of the metrics δv-max , λv in capturing intra-tumour heterogeneity , as well as tumour-to-tumour variations , alongside the ability of our mechanotactic model in predicting these changes . Despite the complexity of the proposed framework , our model remains subject to some limitations; particularly , we do not model space- and time-varying haematocrit and relative blood viscosity [50 , 54 , 70] , or the impact of the lymphatic system . The focus of this study was the investigation of the coupling between tumour vasculature , growth and the generation of solid stresses , and therefore these features were ignored at this stage , for the sake of simplicity . Also , we envisage to incorporate in a future version of the tumour angiogenesis module an elaborate model of vascular wall-remodelling and biomechanics . For example , the biomechanical properties of the microvascular wall ( i . e . lumen size , thickness and pore size ) could explicitly be described with respect to local gradients of chemical cues affecting the endothelium integrity ( i . e . stiffness , permeability ) . As such , we will model the spatio-temporal biomechanics of the microvasculature in a more physiologically realistic manner; this will , for example , overcome the simplification of isotropic mechanical compression . In addition , we plan to extend the present in-silico tumour model to account for polyclonal cell populations and , thus , investigate the importance of the angiogenesis in the interplay between different cancer cell phenotypes ( i . e . proliferative versus invasive phenotype ) under different vascularisation regimes [71] , as well for various tumour-related cell types ( e . g . cancer-associated fibroblasts , cancer-associated macrophages , etc . ) . However , the increased sophistication of our computational framework results in a large number of model parameters presented in S1–S4 Tables . Most of the model parameters were determined independently from the others based on previous studies and experimental data . Other parameters—not found in the literature—were defined in the current study so that model predictions to be physiologically relevant . The good agreement of model predictions with experiments that has been demonstrated in this work validates the choice of model parameters . Furthermore , we performed a parametric analysis of the parameters that play a key role in the angiogenesis procedure and in the compression of the blood vessels ( i . e . mechano-/haptotaxis , TAF threshold for angiogenesis , vessel wall stiffness ) , which are the two factors of tumour progression that our work has been focused on . Variation of other model parameters is expected to change the results only quantitatively , while qualitatively the model predictions and the conclusions of this study will remain the same . We also note that the model framework we present is deterministic; a natural next step would be to investigate variability in the model predictions induced by stochastic features , for example , of the initial vasculature ( length , diameter and separation of vessels ) , informed by distributions of these data as measured in practice . We leave this to future work , but note its important in understanding tissue heterogeneity , both within an individual tumour , and between tumours of the same/different type . Finally , our model can be further used as a platform for the study of the delivery of drugs to solid tumours by adding equations for the transport of the therapeutic agent into the vascular network , across the vessel walls and into the tumour interstitial space [4 , 26 , 72] . Additionally , the efficacy of strategies that target the tumour microenvironment to enhance the delivery of drugs—such as vascular normalization and stress alleviation treatments—can be studied in more detail than by existing models . | Angiogenesis is a hallmark of tumour growth and a key feature for invasion and metastasis . Thus , to elucidate the governing biological processes of cancer development , it is instructive to develop physiologically representative mathematical and computational ( in-silico ) models with which to compare to experimental data . We present a validated , three-dimensional , in-silico model of tumour-induced angiogenesis and growth—unique in the field of mathematical cancer modelling in its biophysical prescription and predictive accuracy . Our novel modelling approach encompasses the effects of both biochemical and biomechanical processes on the development of angiogenic vasculature and solid tumour growth . Unlike other works—which typically focus on biochemical processes—our model depicts a dynamic and coupled biophysical system of the tumour-host microenvironment . In concert with in-vivo experimental data , we have a physiologically-representative in-silico model that enables us to classify and rank the governing biological processes . This yielded the remarkable discovery that it is necessary to include biomechanical factors to accurately describe angiogenic topology . Our model not only provides insight into fundamental cancer biophysics , but is also a powerful tool to aid in our overall objective of re-engineering the tumour microenvironment to optimise the delivery of therapeutic drugs . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"and",
"Discussion"
] | [
"medicine",
"and",
"health",
"sciences",
"body",
"fluids",
"classical",
"mechanics",
"cardiovascular",
"physiology",
"cardiovascular",
"anatomy",
"oxygen",
"basic",
"cancer",
"research",
"capillaries",
"angiogenesis",
"oncology",
"developmental",
"biology",
"tumor",
"phys... | 2017 | A Validated Multiscale In-Silico Model for Mechano-sensitive Tumour Angiogenesis and Growth |
Nodamura Virus ( NoV ) is a nodavirus originally isolated from insects that can replicate in a wide variety of hosts , including mammals . Because of their simplicity and ability to replicate in many diverse hosts , NoV , and the Nodaviridae in general , provide a unique window into the evolution of viruses and host-virus interactions . Here we show that the C-terminus of the viral polymerase exhibits extreme structural and evolutionary flexibility . Indeed , fewer than 10 positively charged residues from the 110 amino acid-long C-terminal region of protein A are required to support RNA1 replication . Strikingly , this region can be replaced by completely unrelated protein sequences , yet still produce a functional replicase . Structure predictions , as well as evolutionary and mutational analyses , indicate that the C-terminal region is structurally disordered and evolves faster than the rest of the viral proteome . Thus , the function of an intrinsically unstructured protein region can be independent of most of its primary sequence , conferring both functional robustness and sequence plasticity on the protein . Our results provide an experimental explanation for rapid evolution of unstructured regions , which enables an effective exploration of the sequence space , and likely function space , available to the virus .
Nodamura virus ( NoV ) is the founding member of the family Nodaviridae . Viruses of this family combine several remarkable features . First , they are capable of replicating in a great variety of diverse hosts . While their natural hosts are insects ( in the case of alphanodaviruses ) and fish ( for betanodaviruses ) , RNA from NoV and Flock House virus ( FHV ) is capable of replication in plant , yeast , and mammalian cells [1]–[4] . Furthermore , their genomes are among the smallest known ( NoV genome is 4 . 5 kb long ) , and are split between two segments , called RNA1 and RNA2 . RNA1's ORF A encodes protein A , which contains the viral RNA-dependent RNA polymerase ( RDRP ) , but is likely to possess other activities as well , such as capping of viral RNAs [5] . RNA2 encodes the capsid protein alpha . RNA2 is not required for viral RNA replication . Indeed , RNA1 can replicate autonomously when introduced into cells [2] . Thus , at 3 . 2 kilobases in length , RNA1 represents one of the smallest animal virus replicons which encodes its own polymerase . During replication , RNA1 gives rise to a subgenomic RNA , called RNA3 ( Fig . 1A ) . This 473-nucleotide sequence is identical to the 3′ end of RNA1 , and can translate two different products from two overlapping ORFs . B1 is a protein produced from the same frame as protein A , and therefore represents the C-terminus of protein A . B2 , on the other hand , is encoded by an overlapping , +1 frameshifted ORF [6] . Remarkably , the reading frames run alongside each other for the entire length of B2 ( see Fig . 1A ) . B2 is required for nodavirus replication for at least two reasons: first , it enables RNA2 translation [7] , and second , it blocks the antiviral RNAi response in insect cells [8] , [9] . RNA1 replication in interferon-deficient mammalian cells , however , does not require B2; thus , its ORF can be removed from RNA1-based replicons with minimal consequences for replication [10] . In summary , three features of NoV combine to set it apart: ( 1 ) The compactness of its genome , ( 2 ) the self-contained replication apparatus with minimal demands on the host , and ( 3 ) the ability to replicate its RNA in the absence of several viral proteins . These features make NoV an ideal platform for understanding critical requirements for replication of eukaryotic viral RNA . In addition , NoV represents a simple and attractive model for studying virus biology , for assessing host responses to a viral pathogen , and for engineering simple expression vectors . We therefore set out to develop NoV RNA1-based replicons which can express foreign genes . In the process of engineering NoV replicons , we identified an RNA element mapping around the stop codon of ORF A , which is required for efficient RNA replication . We also found that the nodavirus-specific C-terminus of NoV proteins A and B1 , which we call AC-TERM ( depicted in magenta in Fig . 1A ) , and which has not been characterized , is essential for replication of viral RNA . Strikingly , AC-TERM can be replaced by completely unrelated amino acid sequences as long as they contain a certain arrangement and a minimum number of positive charges . This terminal region of the polymerase is predicted to be disordered ( [5] , also see below ) . Disordered regions in many proteins are segments that do not fold stably into 3-dimensional domains but rather remain unstructured and are highly flexible , exerting their effects via short peptide motifs [11]–[13] . These regions are widespread , enriched in certain viral proteins [14]–[17] , and play important roles in mediating regulatory protein-protein [18]–[21] and protein-nucleic acid interactions [22] , [23] . While disordered regions are important in regulating cellular [24] , [25] and viral functions [11]–[13] , [26] , the relaxed sequence and structural requirements placed on them are likely to make them amenable to rapid evolutionary adaptation . The relatively high evolutionary rate [27]–[29] associated with these regions can thus facilitate the rise of novel functions . Here , we demonstrate that the C-terminus of nodaviral protein A is highly variable and provide experimental evidence that only a few positively charged residues within an unstructured region can preserve its essential function in virus replication . Strikingly , the sequence of this region evolves rapidly , incorporating changes that may be neutral or result in novel adaptive functions . Our observations support the concept that disordered regions within essential viral proteins expand the sequence and function space accessible to the virus . We propose that these regions can rapidly gain new functions ( for example , form new protein-protein interactions ) during viral adaptation to a changing environment .
We used NoV virus stocks to construct cDNA clones derived from RNA1 and RNA2 . RNA1 was cloned into a plasmid ( pNodBall ) such that it was driven by the SV40 promoter and trailed by an HDV ribozyme ( Fig . 1B ) . This approach , similar to the one previously used in FHV [30] allowed production of replication-competent RNA1 transcript without the need for T7 RNA polymerase-expressing cell lines used so far with NoV [10] . ( For details on NoV cDNA derivation , please see Supporting Information ) . In order to examine replication of RNA1 launched from plasmid DNA , we transfected BSR hamster kidney cells with pNodBall , isolated total RNA at 20 hours post-transfection , and analyzed it by Northern blotting ( Fig . 1C ) . NodBall RNA replicon accumulated over time , while transfection with control NodBall-GAA ( Fig . 1C , right lane ) in which the “GDD” motif of the polymerase was replaced with the inactive amino acid sequence GAA [31] , did not generate detectable replicon RNA . We conclude that wildtype RNA1 replicated efficiently and , as expected , also gave rise to RNA3 ( Fig . 1C , left lane ) . We next introduced several reporter genes into NodBall RNA1s . Protein B2 is required for viral RNA replication in insects as it suppresses antiviral RNAi [8] . It is , however , dispensable for viral RNA replication in yeast and interferon-deficient mammalian cells , though replication is slightly decreased [10] . To simplify the replicons we either deleted or truncated B2 . We either fused the reporter gene to the C-terminus of protein A ( “protein A fusions” ) or to the C-terminus of a truncated protein B2 in the other frame ( “protein B2 fusions” ) . Replacement of B2 portions with reporter genes which truncated protein A did not result in replication-competent replicons ( not shown ) . In contrast , attaching GFP to the C-terminus of protein A was previously reported in Flock House virus [32]; we therefore chose this strategy to generate NoV replicons carrying foreign genes fused to the C-terminus of protein A . In our design , we additionally made use of the foot-and-mouth disease virus ( FMDV ) 2A peptide , which induces co-translational “self-cleavage” to release the foreign protein from the rest of protein A ( [33] , also see below ) . Initially , we created a series of constructs with modified 3′ end of RNA1 . These constructs are schematized in Fig . 2A . First , we generated a B2-defective replicon pNodaB2 ( - ) containing two premature stop codons in the B2 ORF shortly after the second AUG . This modification does not alter the amino acid sequence of protein A , but eliminates B2 production . In agreement with the literature [10] , we find that B2 is not necessary for RNA1 replication in mammalian cells , as assessed by quantitative RT-PCR ( Fig . 2B ) . We next inserted a BsiWI cloning site immediately downstream of ORF A , producing pNoda-bsiw . This minor alteration nevertheless lowered replication efficiency by 2–3 fold ( Fig . 2B ) . When the 2A sequence derived from FMDV was fused to the C-terminus of protein A , we observed a further reduction in RNA1 ( Fig . 2B , pNoda-Pol2A ) . The inserted 2A peptide is 24 amino acids long - it extends protein A by 23 C-terminal amino acids ( and adds an N-terminal proline to any C-terminal transgene ) . RNA1 levels dropped even more when we introduced GFPbsd , a fusion of GFP and blasticidin resistance gene ORFs [34] ( Fig . 2B , pNodaPol2A-GFPbsd ) , downstream of 2A . Because these progressive changes altered both RNA and protein sequence , we hypothesized that both RNA1 3′ end and protein A C-terminus are important for efficient RNA1 replication . To facilitate assessment and quantification of replicon efficiency , we generated new replicons , called pNoda-Pol2A-GFP and pNoda-Pol2A-Luc . These constructs are identical to pNoda-Pol2A-GFPbsd ones , with the exception of the GFPbsd reporter gene , which was replaced with GFP or the firefly luciferase gene , respectively . Next , we sought to improve polymerase activity of protein A fusion constructs . Since FMDV 2A ( “F2A” ) peptide-containing replicon was inefficient ( Fig . 2B ) , we tested several other 2A sequences from other viruses [35] . We found that the Thosea asigna virus 2A ( “T2A” ) sequence allows for a modestly improved protein A expression ( Fig . S1 in S1 Text ) . T2A sequence was shown to self-cleave very efficiently [36] . We verified that the transgene ( GFP ) is cleanly excised from the Pol2A-GFP replicons ( Fig . 3 ) . Pol2A-GFP expression plasmids demonstrated that T2A is indeed cleaved more efficiently than F2A ( Fig . 3B ) . Consequently , T2A sequence was included in all subsequent constructs that employed protein A fusions . We next examined the role of RNA1 3′ end sequence in replication efficiency . A 9-nucleotide insertion directly downstream of ORF A stop codon is sufficient to reduce replication levels of Noda-bsiw 2–3 fold ( Fig . 2B ) . It is well established that the 3′ end of nodaviral RNA1 forms a structure termed 3′ Replication Element ( 3′RE ) , which is required for RNA1 replication [37] , [38] . However , this structure has not been characterized . Our data shows that the 3′RE likely extends upstream into ORF A , and that splitting the ORF-encoded ( blue bars in Fig . 4A ) and 3′UTR-encoded ( red bars in Fig . 4A ) parts of 3′RE lowers replication efficiency . We used MFold [39] to examine potential secondary structures which can form at the 3′ end of RNA1 . MFold predicts a 13-base pair stem-loop within the ORF , directly adjacent to the stop codon ( “ORF stem-loop” , Fig . 4A ) . It also indicates the presence of RNA secondary structures in the 3′ UTR , such as the 3′UTR stem-loop I ( Fig . 4A ) . Thus , 3′RE may consist of 3′UTR and ORF-encoded modules . Since our data predicts a proximity-dependent interaction between such modules , we wondered whether restoring the ORF-encoded module to the 3′ end of the replicon could reestablish efficient replication . Therefore , we systematically introduced different length fragments of RNA1 derived from the 3′end of the ORF back into Pol-2A-Luc 3′UTR ( Fig . 4A , Noda-Pol2A-Luc ( 30 ) or -Luc ( 75 ) , or -Luc ( 140 ) constructs ) . This design leads to the duplication of the 3′RE ORF module: the 5′ repeat at the end of ORF A is translated , while the 3′ repeat ( hatched box ) is not , as it is downstream of the luciferase stop codon . We found that 75- and 140-nt inserts increase Luc expression approximately 6-fold with respect to control; in contrast , a 30-nt long insert was not sufficient ( Fig . 4C ) . Therefore , the 75 nucleotide genomic segment likely forms an RNA structure that needs to adjoin the 3′ UTR of RNA1 . Notably , the 75-nt segment encompasses the predicted ORF stem-loop , while the 30-nt segment does not . This RNA segment could act at either the level of mature RNA1 ( increase translation , replication , or stability of the replicon RNA ) , or during the production or processing of the initial RNA1 transcript from the pNodaPol2A plasmid ( i . e . increasing transcription or nucleocytoplasmic transport ) . To distinguish between these possibilities , we produced Pol-2A-Luc replicons by in vitro transcription using T7 RNA polymerase , and electroporated them into BSR cells ( Fig . 4D ) . Luciferase measurements showed that replicons containing the 75 nt-long RNA segment exhibited higher Luc activity than replicons missing these sequences . Interestingly , the GAA polymerase mutant replicon also expressed significantly higher levels of luciferase at early time points if it contained the 75-nt extension . This observation suggests that this RNA element , positioned next to the translational stop of the A and B2 ORFs , may enhance translation or stability of the viral RNA ( see Discussion ) . The Nodavirus replicons described here efficiently express a transgenic GFP reporter . These modifications included incorporation of the ORF RNA element at the 3′end of RNA1 and an optimized choice of an efficient 2A cleavage site . Median fluorescence intensity of GFP in replicon-transfected cells was found to be 17 . 5 times higher than in cells transfected with the non-replicative GAA control , due to the presence of a population of very brightly fluorescent cells ( Fig . 5A , pNoda-PolT2A-GFP ) . Flock House Virus ( FHV ) , a close relative of NoV , has also been engineered as a replicon using a different strategy . In this case , the bulk of ORF B2 was replaced with GFP [40] . In examining the design of that original FHV-GFP construct , we found that protein A was not truncated at the point where the abbreviated protein B2 ORF was fused to GFP . Instead , the overlapping protein A ORF continued uninterrupted throughout the GFP ORF in an alternative reading frame . Given that the FHV replicon could replicate successfully , we constructed a similar NoV replicon , pNoda-B2GFP , with GFP ORF fused to Glu24 of B2 . In this construct , the C-terminal 114 amino acids of B2 were replaced with 239 amino acids of GFP . This replacement in ORF B2 resulted in a concomitant replacement of the protein A C-terminus sequence , where the C-terminal 112 amino acids in ORF A were replaced with a completely different 250 amino acid sequence which we termed “FPG” ( Frameshifted Protein derived from GFP ) ( see Fig . 5 for a scheme of pNoda-B2GFP and Fig . S2 in S1 Text for the sequence relationships between different ORFs ) . This construct , pNoda-B2GFP , replicated efficiently , judging by the very high levels of GFP expression ( Fig . 5B , lower panel ) . Furthermore , the GFP background in non-replicative GAA replicon is greatly reduced because the transgenic GFP reporter is only produced by the replicated subgenomic RNA3 ( Fig . 5B ) . We were intrigued by the unexpected ability of FPG – a completely different protein sequence with no previously known functions – to functionally substitute for the native C-terminus of protein A . The C-terminus of protein A and FPG share no significant sequence similarity ( as determined by a BLAST comparison [41] ) , and there are no shorter segments that can result in a significant alignment between the two ( as determined independently using the alignment program MUSCLE [42] ) . We compared the amino acid composition of the wildtype C-terminus of NoV protein A with that of NoV protein B2 , FPG , and GFP ( Fig . 6A ) , and found that four amino acids are significantly enriched in both A C-TERM and FPG . Indeed , the combined percentage of prolines , arginines , alanines and glycines ( % P/R/A/G ) reached almost 60% in both polypeptides . This was in stark contrast to B2 and GFP , where the P/R/A/G proportion was significantly lower ( Fig . 6A ) . Three of these amino acids ( proline , arginine , glycine ) are known to be enriched in disordered regions in polypeptides [18] , [43]–[46] . Indeed , when analyzing the profile of these two proteins , we observe that unlike structured proteins such as GFP and B2 , both nodavirus AC-TERM and FPG are predicted to be intrinsically disordered – that is , their sequences are not expected to fold into a 3-dimensional structure ( Fig . 6B ) . Another striking feature of both sequences is their relatively high net positive charge ( +14 in the 112 amino acids of AC-TERM and +45 in the 250 amino acids of FPG ) . We speculate that some of these features , which are serendipitously shared by the two unrelated sequences , mediate the essential function/s required for RNA1 replication . When two viral ORFs overlap , often at least one of them encodes a disordered polypeptide region within the overlapping frame [47] . When we analyzed the AC-TERM region of various other nodaviruses , we observed that , while they have low sequence conservation and different chain lengths , they are all predicted to be highly disordered ( using IUpred [48] ) . The overlapping frame's protein B2 is known to fold into a simple structure [49]–[51] . B2 is more conserved than AC-TERM ( Fig . 6C ) , consistent with the idea that disordered regions tend to evolve faster than structured regions [27] , [29] , [52] . Thus , while B2 is a structured and relatively conserved protein , AC-TERM is highly disordered and poorly conserved , to the point that almost no similarity exists among AC-TERM proteins of various Nodaviridiae family members . Our observation that two vastly different sequences , which are both disordered , can support replication , as well as the lack of conservation among AC-TERM sequences across the nodavirus family , led us to further explore the essential features within this region . We created a new set of replicons where the C-terminal region of protein A was replaced by various engineered sequences which lack alternative ( protein B2-derived ) ORFs ( Fig . 7A ) . All inserts were encoded by nucleotide sequences unrelated to the original viral RNA sequence in order to eliminate any RNA1 structure effects on replication . RNA1 levels were determined by qPCR , as described above . As expected , deletion of the protein A C-terminus prevented RNA1 replication ( Fig . 7C , compare Noda-B1-Truncated with the negative control replicon Noda-B1-FL/GAA ) . A wt AC-TERM fragment ( full-length B1: Noda-B1-FL ) , however , produced RNA levels 3 orders of magnitude above those of the GAA negative control replicon . We note that AC-TERM is required for RNA1 replication at the protein rather than at the RNA level because replication proceeded despite the fact that wt RNA sequence of AC-TERM was replaced with synonymous codons . We next substituted AC-TERM with three fragments derived from the AC-TERM sequence ( B1x , B1y , B1z , Fig . 7B ) . Each of these fragments is about 35 amino acids in length; they correspond to the N-terminal , middle and C-terminal thirds of AC-TERM , respectively . To our surprise , two of these segments restored replication to wildtype levels ( B1y and B1z , Fig . 7C ) , as we would anticipate that only one , if any , of the 3 segments would incorporate the region needed for replication . This is despite the lack of significant similarity between any of the segments ( as determined by BLAST ) . GFP and its frameshift ( FPG ) inserts , used in the same experiments , confirm that a mere extension of protein A ( by GFP ) is insufficient for activity , while the FPG insert stimulates RNA1 replication well above that of truncated or GFP controls ( Fig . 7C ) . Thus , several dissimilar , yet all structurally disordered , sequences can support RNA1 replication . While AC-TERM is required for amplification of RNA1 , its activity may not be influencing the process of RNA replication as such . It is possible that its importance lies in enhancing translation of RNA1 or stability of protein A . We thus further examined the role of AC-TERM by modifying all the replicons in Fig . 7A via addition of the HA tag between the bulk of protein A and AC-TERM ( magenta line in Fig . 7A ) . Protein A expression from HA-containing constructs mirrored RNA replication of non-HA constructs , in a manner dependent on the exact AC-TERM present ( Fig . 7D , right panel ) . However , protein A levels in non-replicative ( GAA mutant ) HA constructs did not correlate with the replication efficiency of the corresponding replication-proficient replicons ( compare the left and the right panels in Fig . 7D ) . This supports the idea that AC-TERM plays a direct role in RNA replication , rather than in translation or stability of protein A itself . We next examined whether the disordered nature of the AC-TERM sequence , by itself , is sufficient to support replication . To this end , we replaced AC-TERM in the wildtype protein A replicon ( pNoda-B1-FL ) with artificial 35-amino acid tails . We designed these fragments to approximate a random disordered region , and based them on either the average composition of protein B1 of various nodaviruses ( “N sequences” , Fig . 8A ) , or the total amino acid content of all known disordered regions in the entire Uniprot database [53] ( “U sequences” , Fig . 8A ) . Three different amino acid sequences were cloned for each of these two amino acid compositions ( N1 , N2 , N3 and U1 , U2 , U3 ) , and in each case the order of residues within the constructs was randomly assigned . Thus , the N sequences contained the same set of amino acids differing between the three variants in their exact positions; and the U sequences contained a different set of amino acids which was similarly shuffled from one U sequence to the next ( see table S3 in S1 Text for the relative fractions of each amino acid in these sequences , and Fig . 8A for their precise sequences ) . As shown in Fig . 8A , only one of the six constructs – N1 – was active . This suggests that the disordered nature of AC-TERM is likely to be required but is insufficient for the RNA replication activity of protein A , and that certain uncharacterized features within these disordered regions are needed to support replication . In addition , these results exemplify the rapid evolvability of the functions encoded in the disordered C-terminus , as one in six randomly designed disordered regions is indeed able to support replication . To further characterize the features of AC-TERM , we performed a systematic mutagenesis of the amino acids in the B1z construct . A given type of amino acid was mutated in a concerted fashion to a different amino acid throughout B1z . Thus , arginines were mutated to lysines ( Mutant 1 ) , to glutamates ( Mutant 2 ) , or to alanines ( Mutant 3 ) ; Mutant 2 also contained the lysine-to-aspartate , and Mutant 3 – lysine-to-alanine substitutions to remove all positive charge . In Mutant 4 , glycines and alanines were mutated to prolines; in Mutant 5 , prolines were mutated to glycines , and , in Mutant 6 , serines and threonines were replaced with alanines . Finally , several mutations were built into the tail to remove the few amino acids which are not proline , arginine , alanine , glycine and serine/threonine ( the most common residues in the original B1z sequence; Mutant 7 ) . Amino acid sequences and replication levels for all of these constructs are shown in Fig . 8B . Strikingly , all of the mutants , with the exception of Mutant 2 and Mutant 3 , in which positive charges were removed , replicated efficiently . Therefore , we conclude that two major features characterize the C-terminus of protein A: structural disorder and the requirement for positive charges . Additionally , it appears that the positively charged amino acids may need to be arrayed in some particular order , since not every sequence carrying the same charge can support replication ( Fig . 8A , see constructs N2 and N3 ) .
While attempting to explain decreased replication of some initial replicon constructs , we found an RNA element straddling the stop codon of the ORF A . It seems likely that this RNA element is a part of the previously observed 3′ Replication Element of RNA1 ( 3′RE ) in FHV , which was found to cover almost the whole RNA3 region in FHV [38] . Here we find that some portions of this region are in fact dispensable for RNA1 replication , at least in NoV ( nucleotides 2816–3078 of NoV RNA1 ) . Our results suggest that the 5′ border of this element probably lies in the coding region between nucleotides 3079 and 3124 of NoV RNA1 . Our findings resemble those of Albarino et al [37] , who found that the in trans replication of FHV RNA1 requires only the 108 3′-most nucleotides in 3′RE – and that this region extends 5′ of the stop codons of the A and B2 ORFs . Of note , our analysis and Albarino′s study used mammalian cells , while the study by Lindenbach et al . was conducted in yeast . RNA1 replication requirements in these systems may be somewhat different . The exact role of the 75 nucleotides at the 3′ end of the protein A ORF is not yet clear . It seems that the 3′RE can be split into two parts , the 3′UTR module and the 75-nt ORF module ( red and blue bars , respectively , in Fig . 4A ) . We note that RNA1 derivatives with the full 3′RE produce more luciferase upon electroporation into BSR cells than those whose ORF module does not adjoin the 3′UTR module ( Fig . 4D ) , even when these RNA1s encode inactive polymerase . Improved stability can be imparted on nodaviral RNA by protein A [54] , and it is possible that the 75-nt RNA region provides an additional binding site for the RDRP . However , it appears that RNA remains relatively stable with or without full 3′RE , as luciferase expression does not decrease appreciably over the first 8 hours . Thus , although our study is not conclusive with respect to the mechanistic role of this RNA structure , it is mostly consistent with the idea that this RNA element ensures efficient translation , rather than improved stability , of RNA1 . Additionally , we cannot exclude a direct role for the 75nt module in replication . In silico RNA structure prediction suggests the presence of two separate stem-loops ( Fig . 4B ) , one formed by nucleotides 3113–3142 ( the ORF stem-loop ) , and the other , by nucleotides 3162–3177 ( the 3′UTR stem-loop I ) . However , since efficient replication depends on the proximity of the ORF module and the 3′ UTR module ( Figs . 2B and 4C , D ) , we speculate that these RNA structures may function as a single RNA element . We also note that the loop of the RNA1 3′UTR stem-loop I proposed here ( Fig . 4B ) contains the same hexanucleotide sequence CCAUCU that forms the loop of the recently characterized 3′SL replication element in NoV RNA2 [55] . This sequence , perched at the end of a stem-loop , may serve as a common binding site for the polymerase , or a site for an RNA-RNA interaction . A replicon's utility is fully realized in its ability to encode a foreign gene . Currently , several designs have been published which allow one to express a reporter gene in a nodaviral system . Some schemes require the use of two co-transfected plasmids [55] , [56] . A simpler scheme requiring only one RNA1-based plasmid would be desirable . Such systems have been established for FHV . In one design ( the protein A-fusion ) , GFP fused to the C-terminus of the FHV polymerase resulted in a functional replicon [32] . In a different design ( protein B2-fusion ) , GFP was used to replace most of the B2 ORF [40] . Here , we describe replicons which can efficiently express transgenes from NoV RNA1 ( Fig . 3 ) . Of the two possible transgene designs , protein A fusions ( Fig . 5 , panel A ) can express any transgene which tolerates an N-terminal proline in mammalian cells . The alternative , protein B2-fusion approach ( Fig . 5 , panel B ) , has proven much more surprising . While GFP is expressed comparably from both protein A and protein B2 fusions , the background fluorescence is lower in the latter . This is expected , since the B2-GFP ORF can be translated only from replicating RNA3 , but not from the primary unreplicated transcripts . However , protein B2-GFP construct design cannot be extended to most other genes , as it relies on a very particular alternative ORF of GFP , which we named FPG . In contrast , the replicons described here allow for robust expression of an ectopic protein that is not fused to protein A or to B2 . The use of this type of replicons may need to be carefully optimized since NoV lyses infected cells . We wondered if , unlike the wild-type virus , capsid-deficient replicons described here may maintain long-term replication in cells . However , our initial observations indicated that robust replication of these replicons may also lead to cell lysis . Several functional domains have been mapped in protein A . Its central domain contains the catalytic core of the viral polymerase , and its N-terminal third is hypothesized to encode an RNA capping enzyme [5] . The extreme N-terminus of A localizes the protein to mitochondrial membranes [57] , [58]; additionally , multimerization motifs are present throughout the protein [59] . However , the function and requirements of its C-terminus ( AC-TERM ) have remained unclear . The finding that a completely unrelated protein sequence , FPG , can replace the original AC-TERM , was quite unexpected . We hypothesized that its disordered nature , and/or amino acid composition , may mimic the features of the natural AC-TERM . Strikingly , AC-TERM could be shortened to less than one third of its size ( 35 amino acids ) without losing function . Furthermore , either the middle third or the C-terminal third of the sequence could fulfill the requirements for RNA1 replication ( Fig . 7 ) despite the lack of any significant similarity between the two sequences . We found that even a randomly shuffled disordered sequence could support the replication function of AC-TERM ( Fig . 8A ) – a unique situation rarely encountered in proteins . However , other disordered sequences with similar compositions had no activity ( Figs . 7B and 8 ) , indicating that a pattern , and not merely a set , of amino acids was required for function . Further examination revealed that positive charges , preferably arginines , but not other specific amino acids , are necessary for the function of AC-TERM ( Fig . 8B ) . Thus , successful RNA1 replication requires a pattern of positive charges in the context of a disordered region . Positively charged amino acids , and arginines in particular , are known to facilitate binding to RNA [60] , [61]; thus the C-terminus of protein A may be required for interactions with viral RNA . Interestingly , intrinsically disordered regions are thought to be important in regulation of nucleic-acid binding [22] , [23] , [62] . Alternatively , AC-TERM could function much like a positively charged amino acid cluster in FHV coat protein , known to direct the protein to mitochondria [63] . It is noteworthy that membrane localization of protein A is not entirely dependent on the N-terminus [57] , [58]; therefore , a contribution from the C-terminus may also be required for subcellular localization of the replication complex . Our evolutionary analysis ( Fig . 4 ) demonstrates that an exceptionally diverse array of sequences serve as C-termini for the polymerases of the Nodaviridae . This analysis strongly supports the concept that the sequence requirements placed on the C-terminus of protein A are remarkably relaxed . It is important to note that , while we demonstrate a very limited set of requirements imposed on AC-TERM itself , there are other constraints on its sequence . Since ORF B2 overlaps ORF A , there must be a strict limit to the diversity allowed at AC-TERM . Indeed , it has been shown that in general , overlapping reading frames tend to be relatively conserved [64] . Furthermore , as we show above , the region encoding the last 20 or so amino acids in AC-TERM contains an RNA element important for RNA amplification . In light of this , it is especially remarkable that AC-TERM is the most variable protein region in the virus ( Fig . 6C ) . In part , this may be due to the fact that B2 can also accommodate a lot of variation ( Fig . 6C ) . We propose that because AC-TERM is disordered , it can tolerate higher sequence diversity without compromising its overall structure or stability . As a consequence , the disordered region may allow for a more thorough sequence space exploration , which may result in the generation of novel , adaptive functions ( see below ) . Taken together , our analysis indicates that the extreme sequence plasticity of the C-terminus of Protein A has been exploited during evolution , leading to the diversification of protein sequence in this region . In this study , we integrated two different and complementing analyses . First , by using structure predictions and evolutionary comparisons , we demonstrated that protein A C-terminus is highly disordered and that its similarity across different Nodaviruses is extremely low in comparison to other nodaviral proteins . Second , we showed that replacing the original C-terminus sequence with other disordered regions – some with no apparent similarity – still allows replication . We thus discovered that a rather extended protein region , which is required for a core viral activity exhibits a remarkable robustness of its function in the face of numerous and drastic sequence changes . Many studies have presented extensive computational analyses of disordered proteins and their evolution [28] , [29] , [65] . They are in broad agreement that disordered protein regions evolve rapidly; however , the exact extent to which these regions can be mutated and yet maintain their function has not been explored . Here , we have marshalled both computational and experimental evidence to show a remarkable capacity of structurally disordered regions to evolve fast while maintaining essential functions such as virus replication . Intrinsically disordered protein regions are now appreciated as an extremely versatile part of the proteome . Their conformational flexibility and potential for mediating multiple intermolecular interactions is thought to allow them to rapidly acquire novel functions in the course of evolution . Comparative studies of orthologous sequences in various eukaryotes have revealed that disordered regions indeed evolve exceptionally fast [27] , [28] . NoV AC-TERM represents a notable illustration of this rapid evolution . Viral genome size is constrained . In principle any “extra” genomic sequence space in a virus would be lost without a selective pressure to maintain it , yet a rather extended AC-TERM persists throughout Nodaviridae . While it is possible that maintaining the entire length of this segment is selectively neutral , we hypothesize that mechanisms to maintain it can exist and provide a selective advantage . Recent studies suggest that viruses exploit short peptide motifs residing in disordered regions to evolve many of their interactions with host proteins [12] , [13] . Here , we demonstrate that the disordered C-terminus of NoV protein A , which is essential for viral RNA1 replication , exhibits extremely lax sequence requirements for function , and thus has the potential to incorporate new motifs and functions during evolution . From the point of view of viral evolution , then , the relaxed nature of the AC-TERM sequence effectively allows the diversification of the sequence in this region without a significant loss of fitness . This in turn may provide a rich reservoir of novel sequences and functions . Additionally , the variability of the AC-TERM sequence can be the result of neutral selection where replication function is maintained and the rest of the sequence changes under no selective pressure . At the moment we do not know whether or not the AC-TERM sequence indeed carries out additional functions . Future experimentation will be necessary to determine the specific role of this region in a given host or environment . In this regard , it is notable that protein B1 , which is colinear with AC-TERM , is expressed from RNA3 [6] . B1 expression is dispensable for RNA replication [6] , but B1 may harbor additional functions . Whereas a limited number of amino acids are required for RNA replication at AC-TERM , a different subset of amino acids in this region can fulfill a different function as a part of a ( free ) B1 protein or as the C-terminus of protein A . This sequence and functional flexibility underscores the remarkable fact that this region of the viral genome can maintain function through the conservation of a few dispersed residues while allowing evolution to produce highly diversified intervening sequences .
Freeze-dried Nodamura virus , strain Mag115 , was obtained from ATCC ( Cat . No . VR-679 ) and resuspended in 1 ml of water . 250 ul of this suspension was processed with 750 ul of TRIzol LS according to the manufacturer's protocol . Following cDNA synthesis using random hexamers and SuperScriptIII ( Invitrogen ) , several primer pairs were used to amplify , TA-clone and sequence the amplicons ( noV1Fpst2: 5′-ccac ctgca gtattgaatccaaaactcaaaatgctgaac-3′ with noda1654R: 5′-GAT CAC GGA ATG CCA GCG TAT AGC TGG AAA ACC G-3′; noda1383F: 5′-caaggtccactggccagcgcacgtcgaag-3′ with NoV1-RT: 5′-ACC ACT GGC ATA AGC CTA GTT C-3′ were used for the initial cloning and sequencing . For amplifying the 5′-3′ end junctions , primers nov1-2938F ( 5′-catcaaaccgcgagtcgcag-3′ ) and nov1-171R ( 5′-CGT GCG TCG ATG CAC GAT-3′ ) were used ) . 5′ genomic and 3′ genomic clones were then merged and cloned behind the SV40 minimal early promoter ( amplified with sv40promFsac: 5′-ccac ttataa gcgatcgc gagctc tgcatctcaattagtcagcaacc-3′ and sv40pRpst: 5′-CTGCAG CGG CCT CGG CCT-3′ ) . Standard cloning techniques were employed for construction of these and other plasmids; PCR was conducted with Phusion proofreading polymerase ( New England Biolabs ) . Protein A expression plasmids did not contain 5′ and 3′ UTRs of RNA1; furthermore , a number of synonymous mutations were introduced into the 3′ third of protein A ORF ( by gene synthesis: BioBasic , Inc . ) in order to remove any RNA replication elements . GFP used in our replicons is a brighter variant of GFP with a very similar nucleotide sequence , called Venus [66] . Plasmid sequences and details of construction are available upon request . BSR cells ( which do not express T7 polymerase ) were obtained from the Matthias Schnell laboratory and cultured in the DMEM-high glucose medium ( UCSF Cell Culture Facility ) supplemented with 10% FBS ( Sigma ) , glutamine , penicillin-streptomycin and non-essential amino acids ( UCSF CCF ) . Transfections of the NoV replicon-expressing plasmids were done in 24-well plates ( using 20–50% confluent cells ) with 1 ul of LipofectAMINE2000 and 0 . 5 ug of plasmid ( s ) according to the manufacturer's protocol . For qPCRs , RNA was subsequently collected using 0 . 3 ml Trizol ( Invitrogen ) per well , precipitated and resuspended in 25–60 ul of water . 1 . 5–2 . 5 ug total RNA was treated with DNaseI ( Promega ) and , following DNase inactivation at 75°C , 200–500 ng was used in a SuperScriptIII ( Invitrogen ) reverse transcription . First strand cDNA was used in the qPCR run with the SYBR FAST Universal 2× qPCR Master Mix ( Kapa Biosciences ) on a CFX-Connect cycler ( Biorad ) . Primers used in the PCR were: noda1219F ( 5′-gccataaatcccaaggtccactg -3′ ) and noda1325R ( 5′-GGC ATC ATA TTT TCG TCA GAT ACC AAC G -3′ ) to amplify NoV RNA1 , hamsteRplF ( 5′-AGC CCG TGA CTG TCC ATT C -3′ ) and hamsteRplR ( 5′-GGC AGT ACC CTT CCG CT -3′ ) to amplify the hamster Rpl19 message , in a final volume of 10 ul . Cycling conditions were: 95°C for 10 seconds , 62°C for 20 seconds , and 72°C for 30 seconds , for 40 cycles . 1∶5 or 1∶200 dilutions of the cDNAs were used as the starting material . “No RT” controls were always run in parallel to ensure that the signal did not originate from the DNase-undigested plasmid . Relative amount of NoV replicons was obtained by referencing the Noda signal to the Rpl19 signal ( delta-delta Ct method ) . For in vitro transcriptions , SV40 promoter was substituted with the T7 promoter in the relevant plasmids , and the plasmids were linearized at the XbaI restriction site 3′ of the HDV ribozyme . Transcriptions were conducted with mMessage mMachine reagents from Ambion , as suggested by the manufacturer , using the cap analog∶GTP ratio of 4∶1 . Northern blotting was done according to the standard protcols [67] , after running RNA in formaldehyde gels and transfer to nitrocellulose membranes . The Northern probe was labeled by random priming of the PCR product , the 3′-most 1167 nucleotides of RNA1 , derived by amplifying pNodBall ( see text ) with noda2038F ( 5′-caatattgctccattcgaatgacacacccagagc-3′ ) and NoV1-RT . Hybridization at 42°C in UltraHyb buffer ( Ambion ) was followed by washes at 50 and 55°C in 2×SSC/0 . 1% SDS and 0 . 2×SSC/0 . 1% SDS , respectively . Dual Firefly/Renilla luciferase assays were run using Promega's Dual Luciferase Assay system according to the manufacturer's protocol . Cells were lysed in 50 ul of Passive Cell Lysis buffer per one well of the 24-well plate . 10 ul of the cleared lysate was added to 70 ul of the Firefly assay reagent , followed by 70 ul of the Renilla assay reagent . Measurements were conducted in Tecan's UltraEvolution 96-well plate luminescence reader . Similar protocol was followed when using single firefly , Luciferase Assay System ( Promega ) . For flow cytometry , transfected cells were trypsinized , washed and resuspended in PBS with 2% FBS , then analyzed on FACScalibur ( Becton-Dickinson ) for GFP expression; FlowJo software was used for data processing . For Western analysis of 2A-mediated cleavage , cells were lysed 48 h after transfection of 30%-confluent BSR in a well of a 6-well plate , in 75 ul of cytoplasmic lysis buffer ( 150 mM KCl , 2 mM MgCl2 , 30 mM Hepes pH 7 . 4 , 0 . 5% NP40 , protease inhibitors ( Roche ) ) . 20 ul of the resulting sample was run in a 4–12% NuPAGE denaturing polyacrylamide gel ( Invitrogen ) . Samples were then transferred onto a PVDF membrane and probed with an anti-GFP rabbit polyclonal ( sc-8334 , Santa Cruz Biotechnology ) and an HRP-conjugated donkey anti-rabbit IgG-F ( ab ) 2 ( GE Healthcare ) , both at 1∶2 , 500 . ECL system ( Pierce ) was used for signal detection . Western blot of the HA expression replicons was done by lysing BSR cells 24 hours after transfection ( as above ) with a mixture of the replicon ( 1 . 5 ug ) and pEF6-GFPflag internal normalization control plasmid . 40 ul of the RIPA buffer with protease inhibitors ( Roche ) was applied per well of a 12-well plate . 10 ul of each sample was loaded per lane of a 4–20% Biorad denaturing polyacrylamide gel . After transfer to a PVDF membrane , anti-HA mouse monoclonal ( 6E2 , Cell Signaling #2367 ) was used at 1∶1000 , or anti-Flag mouse monoclonal M2 ( F3165 , Sigma ) was used at 1∶25 , 000 . Secondary antibody ( sheep anti-mouse HRP-conjugated IgG-F ( ab ) 2 ) was bought from GE Healthcare ( NA9310V ) and used at 1∶2 , 500 to detect 6E2 and at 1∶10 , 000 to detect M2 . Signal was developed using the ECL system ( Pierce ) . For RNA structure predictions , we accessed an MFold server ( http://mfold . rna . albany . edu/ ? q=mfold/RNA-Folding-Form ) and submitted queries using the Web interface . Standard conditions were chosen , and 5 resulting structures were examined . We assembled a set of 7 viruses belonging to the alpha and beta branches of the nodavirus family . For each of the 7 viruses , we collected the capsid polyprotein , the RDRP ( protein A ) , and protein B2 from the NCBI online database ( see links in table S2 in S1 Text ) . We inferred the sequence of protein B1 – the C-terminus tail of protein A , when it was not available online , based on sequence similarity to other family members or based on the existence of a methionine which presumably acts as its initiation site . Using BLAST [41] , we calculated the similarity score of each of the protein ( A , B1 , B2 ) with respect to the nodavirus orthologs , by finding the region that is most significantly similar to the nodavirus protein in the orthologous protein , and computing a normalized similarity score as: ( the length of the similar region ) × ( % similar residues that are similar in the alignment in this region ) / ( total length of the orthologous protein ) . The results of the normalized similarity scores for protein A , B1 and B2 appear in Fig . 4 , with a phylogenetic tree that was constructed based on the capsid proteins . The tree was formed by creating an alignment using the MUSCLE program [42] and by using the PhyML program with default parameters [68] . The tree figure was created using the FigTree program ( http://tree . bio . ed . ac . uk/software/figtree/ ) . We predicted the disorder profile of each of the protein A's sequences using the disorder predictor IUpred [48] , using the default parameters and the ‘long’ version ( which is optimized to search for long stretches of disordered regions , such as the C-terminus tail of protein A ) . IUpred was shown to give similar predictions to other methods and to be in a strong agreement with experimental data [69] . A residue was considered to be ‘disordered’ if its predicted disorder value was 0 . 4 or higher ( in the scale of 0 to 1 ) . We then computed the average AA content of the 7 B1 proteins , so that each of the 20 AAs has a fraction of occurrence in the set of B1 proteins ( e . g . – on average , 6 . 3% of the residues in the 7 B1 proteins are lysine residues ) . See Table S3 in Text S1 for the composition of disordered regions in the Nodaviridae B1 proteins . We downloaded the entire uniprot dataset [53] ( version 2013_1 ) , and predicted the disordered regions of each of the proteins as described above , and calculated the average disordered content of the entire proteome ( using all the residues in the uniprot dataset that are predicted to be disordered; e , g . – on average , 6 . 5% of the residues in disordered regions of the entire uniprot set are Lysine residues ) . See Table S3 in Text S1 for the composition of disordered regions in the entire uniprot database . We predicted the disorder profile of GFP and its frame-shifted version ( “FPG” ) using the same parameters . Using BLAST , we searched for sequence similarity between “FPG” and the nodavirus protein A . Using the fraction of occurrence of residues in disordered regions in the nodavirus B1 proteins , we created three constructs with this amino acid content ordered at random . In each of these constructs we replaced the nodavirus B1 protein with a sequence of 35 residues , that the order of the residues was determined by random using the uShuffle program [70] , and the propensity of each residue was similar to its fraction in the B1 proteins . Similarly , we created a set of three 35-AAs long constructs , with amino acid propensities based on the disordered content of the entire uniprot database , with a random order along the construct . The engineered region of each of these constructs sequences was predicted to be disordered according to the procedure described above . | Proteins often contain regions with defined structures that enable their function . While important for maintaining the overall architecture of the protein , structural conservation adds constraints on the ability of the protein to mutate , and thus evolve . Viruses of eukaryotes , however , often encode for proteins with unstructured regions . As these regions are less constrained , they are more likely to accumulate mutations , which in turn can facilitate the appearance of novel functions during the evolution of the virus . Even though it has been known that such “disordered protein regions” have been particularly malleable in evolution , their functions and their ability to withstand extensive mutations have not been explored in detail . Here , we discovered that a disordered part of the Nodamura Virus polymerase is both required for replication of the viral genome , and extremely variable among different nodaviruses . We examined the tolerance of this protein region to mutations and found an unexpected ability to accommodate very diverse protein sequences . We propose that disordered protein regions can be a reservoir for evolutionary innovation that can play important roles in virus adaptation to new environments . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"biochemistry",
"biology",
"and",
"life",
"sciences",
"species",
"interactions",
"microbiology",
"evolutionary",
"biology"
] | 2014 | Rapid Evolution of Virus Sequences in Intrinsically Disordered Protein Regions |
The pentatricopeptide repeat ( PPR ) is a helical repeat motif found in an exceptionally large family of RNA–binding proteins that functions in mitochondrial and chloroplast gene expression . PPR proteins harbor between 2 and 30 repeats and typically bind single-stranded RNA in a sequence-specific fashion . However , the basis for sequence-specific RNA recognition by PPR tracts has been unknown . We used computational methods to infer a code for nucleotide recognition involving two amino acids in each repeat , and we validated this model by recoding a PPR protein to bind novel RNA sequences in vitro . Our results show that PPR tracts bind RNA via a modular recognition mechanism that differs from previously described RNA–protein recognition modes and that underpins a natural library of specific protein/RNA partners of unprecedented size and diversity . These findings provide a significant step toward the prediction of native binding sites of the enormous number of PPR proteins found in nature . Furthermore , the extraordinary evolutionary plasticity of the PPR family suggests that the PPR scaffold will be particularly amenable to redesign for new sequence specificities and functions .
Much of modern biology deals with understanding and predicting macromolecular interactions . The biotechnological possibilities inherent in being able to predict , design and manipulate macromolecular interactions are immense . The well-understood Watson-Crick pairing between nucleic acid strands facilitates the design of nucleic acids that can interact with specific DNA or RNA sequences , and this ability underlies a huge swathe of modern research and biotechnology . Given the greater functional potentialities of proteins compared to nucleic acids and the ability to target proteins to different intracellular compartments , new opportunities would emerge from the ability to design proteins to bind specific RNA or DNA sequences . Unfortunately , most protein-nucleic acid interactions are idiosyncratic , and lack the predictability necessary to engineer specific interactions . Recently , a great deal of excitement has accompanied the characterization of Transcription-Activator-Like Effectors ( TALEs ) , a set of modular repeat proteins that bind via a predictable code to specific double-stranded DNA sequences [1] , [2] . TALEs belong to the alpha-solenoid superfamily comprising proteins that consist of degenerate repeats of 30–40 amino acids , each of which forms two or three alpha-helices . This superfamily includes only one well characterized member that binds RNA: the Puf domain family . Puf domains consist of eight tandem repeats of a triple-helix motif that bind 8–9 nucleotide sites ( reviewed in [3] ) . The residues within each motif that dictate sequence specificity have been identified , and experiments to manipulate binding specificity and protein function by exploiting this modular recognition have been successful [3] , [4] , [5] . This study focuses on a second class of helical repeat motif that binds RNA , the pentatricopeptide repeat ( PPR ) . PPR proteins harbor degenerate ∼35 amino acid repeats that are related to tetratricopeptide ( TPR ) motifs [6] . PPR proteins localize primarily to mitochondria and chloroplasts where they influence various aspects of RNA metabolism [7] . Many PPR proteins are essential for photosynthesis or respiration , and mutations in PPR-encoding genes are associated with genetic diseases in humans ( e . g . [8] ) . Although less widely known than Pufs and TALEs , PPR proteins are much more prevalent in nature . Protist , fungal and metazoan genomes encode roughly 5–50 PPR proteins , but the family has expanded to >400 members in plants ( reviewed in [9] ) . The products of evolution illustrate the apparent ease with which PPR tracts can be modified to bind diverse sequences and mediate diverse functions: PPR proteins harbor between 2 and ∼30 repeats and they influence the processing , editing , splicing , stability or translation of specific organellar RNAs [7] . The remarkable evolutionary plasticity of PPR proteins is highlighted by their natural exploitation to silence rapidly evolving mitochondrial open reading frames that confer cytoplasmic male sterility in plants [10] . Results presented here demonstrate that PPR tracts bind RNA via a modular mechanism that conceptually resembles Puf-RNA recognition . However , the details of nucleotide recognition by PPR motifs differ from those for Puf repeats , revealing a diversity of independently evolved RNA recognition modes by alpha solenoid repeats . These insights provide a significant step toward the prediction of binding sites and functions for the large number of PPR proteins found in nature . Additionally , the evolutionary malleability of the PPR family implies that PPR binding specificities can be engineered to match a wide variety of desired targets .
Recombinant PPR10 ( rPPR10 ) elutes from a gel filtration column at a position corresponding to a globular homodimer [11] , as does HCF152 , which likewise consists almost entirely of PPR motifs [13] . Models for PPR-RNA interaction would need to incorporate homodimerization , should this be physiologically relevant . To clarify this point , we analyzed rPPR10 by sedimentation velocity analytical ultracentrifugation ( SV-AUC ) . rPPR10 was found predominantly in two forms whose ratio changed in a concentration-dependent fashion ( Figure 1A ) . At 3 µM , the major species sedimented at ∼5 S and had an estimated molecular weight of 84 . 9 kDa , close to rPPR10's monomeric molecular weight of 82 . 6 kDa . A two-fold increase in rPPR10 concentration shifted the distribution toward a larger species ( ∼6 . 5 S ) , which predominated when protein concentration was further increased to 12 µM . These results strongly suggest the ∼5 S and 6 . 5 S species to be monomers and dimers , respectively . Thus , rPPR10 can dimerize , but only at very high concentrations . To determine which form of PPR10 binds RNA , rPPR10 was analyzed by SV-AUC in the presence of its 17-nt minimal RNA ligand . This RNA is small in comparison with rPPR10 ( 5 kDa versus 84 kDa ) and does not contribute significant signal with the interference optical system used for these experiments . With rPPR10 at 3 µM and RNA at half that concentration , PPR10 monomers partitioned into two species of similar abundance with an S value near 5 S ( Figure 1B ) . The concentration , sedimentation rate , and RNA-dependence of the second ∼5S species strongly suggest it to be a PPR10 monomer bound to RNA . The pair of species near 5S collapsed into a single ∼5 S species when the RNA concentration was increased to be equimolar with PPR10 ( 3 µM ) . As this concentration is much higher than the Kd for the PPR10-RNA interaction ( <1 nM ) [12] , it is predicted that essentially all of the protein was bound to RNA , assuming a 1∶1 stoichiometry . Taken together , these results provide strong evidence that PPR10 binds RNA in its monomeric form , and that each PPR10 monomer binds one RNA molecule . Under conditions of saturating RNA , PPR10 dimers were not detected . Thus , RNA binding appears to preclude protein dimerization , suggesting that PPR10's RNA binding and dimerization surfaces overlap . The minimal PPR10 binding site in the atpH 5′-UTR spans 17-nt and PPR10 leaves a ribonuclease-resistant footprint spanning ∼24 nucleotides [12] ( Figure 2A ) . To identify specificity determining amino acids , we sought correlations between the amino acid residues at each position of PPR10's PPR motifs and the bases within its footprint . We modeled the RNA in parallel to the protein ( i . e . 5′-end aligned with N-terminus ) due to the organization of PPR proteins that specify sites of RNA editing: such proteins have an N-terminal PPR tract and a C-terminal domain that is required for editing , and they bind cis-elements that are 5′ of the edited sites ( reviewed in [7] ) . We further assumed that all motifs would contact an RNA base , but not necessarily contiguously . These assumptions are based on the similarity between the number of repeats and the number of nucleotides in well-characterized PPR/RNA pairs [12] , [14] , and by a length polymorphism in the middle of PPR10's two binding sites ( Figure 2A ) . Given these constraints , there are 420 possible arrangements of PPR10's PPR motifs in contact with its RNA footprint ( see Materials and Methods ) . One of these arrangements stood out because it showed strong correlations between the RNA base and the amino acids found at positions 1 and 6 ( Table S1 and Figure 2A ) , which were suggested to be specificity-determining positions based on their patterns of evolutionary selection [10] . The alignment to amino acid 6 is offset by one nucleotide from the alignment to amino acid 1 , such that the base that correlates with position 6 of PPR motif n also correlates with position 1 of the n+1 motif; hereafter we shall refer to this position as 1′ , to distinguish it from position 1 in motif n . This offset is physically plausible ( Figure 2B ) , and it is supported by an in vitro analysis of a pair of PPR motifs [15] . The optimal alignment contains a gap that breaks the protein-RNA duplex into two segments . The gap corresponds with the position of a single nucleotide insertion in PPR10's psaJ binding site ( Figure 2A ) , providing evidence for relaxed selection in this region of the binding site . This alignment highlights the following correlations: every N6 aligns with a pyrimidine , each purine corresponds to S6 or T6 , and every D1′ aligns with a U . These correlations are maintained by covariation when one considers the orthologous protein and binding site in Arabidopsis ( Figure 2A ) . These correlations were extended by analysis of the PPR protein HCF152 [13] , which binds to sequences within its 17-nt footprint in the chloroplast psbH-petB intergenic region [16] , [17] . When HCF152's 13 PPR motifs were compared with this sequence , the optimal alignment spanned 12 nucleotides and preserved the correlations observed for PPR10 ( Figure 2C ) . Furthermore , this alignment is maintained through covariation in rice ( Figure 2C ) . The maize protein CRP1 further strengthens these correlations . CRP1 leaves a ∼30-nt footprint in the chloroplast petB-petD intergenic region [16] , [18] . CRP1's 14 PPR motifs can be aligned within this footprint in a manner that retains the correlations noted above ( Figure 2C ) . Similar to the PPR10 alignments , the CRP1 alignment involves ∼7 contiguous matches at each end , with “unpaired” nucleotides in the central region . Notably , the PPR10 , HCF152 , and CRP1 alignments are all placed very similarly within their RNAse-resistant footprints , as is to be expected given that each protein blocks access by the same exonucleases in vivo . Finally , an alignment that follows the same rules can be made between CRP1 and a sequence in the psaC 5′-UTR that maps within the 70-nt segment that is most strongly enriched in CRP1 coimmunoprecipitations [19] ( Figure 2C ) . PPR proteins can be separated into two classes , denoted P and PLS . PPR10 , HCF152 , and CRP1 are examples of P-class proteins , which contain tandem arrays of 35 amino acid PPR motifs . Members of this class have been implicated in RNA stabilization , processing , splicing , and translation . PLS-class proteins contain alternating canonical ‘P’ motifs and variant ‘long’ and ‘short’ PPR motifs [20] , and typically function in RNA editing . PPR editing factors can be aligned to sequences upstream of the edited nucleotide such that the amino acids at position 6 of the ‘P’ motifs and the amino acids at position 1′ of the following ‘L’ motif correlate with the matched nucleotide in a similar manner to that found for the P-class proteins ( Figure 2D ) . Importantly , the editing factors can all be aligned such that their C-terminal motif is at the same distance from the edited cytidine residue . This not only explains how the target C is defined , it allows the motif-nucleotide correlations in the editing factors to be evaluated without using them to make the alignment . Correlations between the aligned base and the amino acids at positions 6 and 1′ are highly significant across all alignments for both ‘P’ and ‘S’ motifs ( Table S2 ) . Apart from these two positions , only the amino acid at 4′ is also significantly correlated with the aligned nucleotide . Sequence logos constructed from PPR motif pairs aligned with either A , G , C , or U are shown in Figure 3 and Figure 4 . From these alignments , a set of rules can be derived that seem likely to represent a combinatorial amino acid code for nucleotide recognition by PPR motifs: T6D1′ = G; T/S6N1′ = A; N6D1′ = U; N6N/S1′ = C . The diversity of amino acid combinations at these positions implies that the code may be degenerate ( Table S3 ) . However , the above-mentioned amino acid combinations are the most commonly observed , and together represent 64% of all canonical PPR motif pairs in Arabidopsis and rice ( Figure S2 ) . To test whether the correlations between amino acid identities at PPR positions 6 and 1′ and the associated nucleotide reflect a recognition code , we generated a set of PPR10 variants in which residues ( 6 , 1′ ) in a pair of adjacent repeats ( motifs six and seven ) were modified to either T6D1′ , T6N1′ , N6D1′ , N6N1′ , or N6S1′ ( Figure 5A ) . Our model aligns PPR10 repeats 6 and 7 with U and C nucleotides , respectively . PPR10 does not bind significantly to RNA in which these nucleotides are substituted with either AA or GG ( Figure 5B ) . A PPR10 variant in which motifs 6 and 7 were modified to ( T , D ) did not bind to the wild-type RNA , but bound with high affinity to RNA with the GG substitution . Likewise , the variant in which these motifs were modified to ( T , N ) did not bind to wild-type RNA , but bound with high affinity to RNA with the AA substitution . Neither variant bound significantly to any of the other substituted RNAs . These results confirmed the proposed polarity and register of the PPR10/RNA complex , and show that ( T , D ) and ( T , N ) at positions ( 6 , 1′ ) are highly specific for binding G and A , respectively . The ( N , D ) , ( N , N ) , and ( N , S ) combinations at ( 6 , 1′ ) correlate with recognition of pyrimidines ( Figure 4 and Table S3 ) . As predicted , PPR10 variants with these amino acid combinations strongly favored binding to pyrimidine-substituted RNAs ( Figure 5B ) . The ( N , D ) variant bound the U and C substituted RNAs with Kds of ∼3 nM and 17 nM , respectively , indicating a clear preference for U over C ( Figure 5C ) . Conversely , the ( N , S ) variant favored C over U , albeit only slightly ( Kds of 9 nM and 20 nM for the C and U substituted RNAs , respectively ) . The ( N , N ) variant is less discriminating , binding the U and C substituted RNAs with similar affinities ( Figure 5C ) .
Results presented here provide strong evidence that PPR tracts bind RNA in a parallel orientation via a modular recognition mechanism , with nucleotide specificity relying primarily on the amino acid identities at positions 6 and 1′ in each repeat . Modification of amino acids at these positions in the context of two adjacent PPR motifs was sufficient to change the nucleotide preference , suggesting that other amino acid positions make no more than a small contribution to nucleotide specificity . Position 4′ correlates weakly with the aligned nucleotide , but threonine is preferred at 4′ for all four nucleotides ( Figure 3 ) and we have not investigated the effect of any other amino acid at this position . Although similar in concept to Puf/RNA recognition , PPR/RNA complexes have the opposite polarity and involve distinct amino acid combinations . The polarity and code we demonstrate for PPR/RNA interactions differ from those proposed by Kobayashi et al [15] , who concluded that the PPR protein HCF152 binds anti-parallel to an A-rich RNA sequence . This model was based on a shallow HCF152 SELEX dataset , from which similarities were sought to a presumed HCF152 binding site that was recently shown not to bind HCF152 with high affinity [16] . Our results define a combinatorial two-amino acid code that can specify binding of a PPR motif to either A , G , U>C , C>U , or U = C . With this knowledge , the engineering of PPR tracts to bind a wide variety of RNA sequences is within reach . However , prediction of the natural binding sites of PPR proteins , and prediction of off-target binding by engineered PPR proteins remains challenging for two reasons . First , the natural diversity of amino acid identities at positions 6 and 1′ implies a degenerate code , and less than two-thirds of naturally occurring combinations can currently be interpreted . Second , an understanding of the energetic parameters required to establish a physiologically meaningful PPR/RNA interaction and the energetic costs of mismatches at various positions along a PPR/RNA duplex will be required to accurately predict potential binding sites . The prediction of microRNA targets is similar in concept and provides a glimpse into the challenge to come: despite the simplicity of RNA base pairing rules , the parameters that dictate microRNA targets are still being worked out [21] . Prediction of binding sites is further complicated by the fact that gaps in a PPR/RNA duplex can be tolerated in some contexts , as exemplified by PPR10's natural targets ( Figure 2A ) . Indeed , the optimal alignments of the P-class PPR proteins HCF152 and CRP1 also contain a gap , with the predicted protein/RNA duplex containing non-contiguous segments of either RNA ( PPR10 and CRP1 ) or protein ( HCF152 ) . These gaps break the protein-RNA duplex into two segments in a manner that resembles Puf-RNA duplexes , which require contiguous protein-RNA matches at each end but can accommodate various flipped base conformations in the central region [22] . Our findings imply considerable flexibility in the length of the “looped out” RNA between contiguous PPR-RNA segments . These RNA loops may be analogous to internal loops in RNA duplexes , which adopt diverse architectures due to the great flexibility of the RNA backbone and to the wealth of opportunities for non-canonical base-base interactions ( reviewed in [23] , [24] ) . Our alignments of P-class PPR proteins to their cognate RNAs include contiguous duplexes consisting of no more than nine motifs and eight nucleotides . This is reminiscent of the binding of 8–9 nucleotides by the eight repeats in Puf proteins ( reviewed in [25] ) . The number of contiguous interactions between helical repeats and RNA bases may be constrained by the minimum distance between parallel alpha helices . The minimum theoretical helix-helix distance is c . 9 . 5 Å [26] , which is approached by the helix-helix distance in Puf motifs [27] . In contrast , adjacent nucleotides in Puf:RNA complexes are 7 Å apart , close to the maximally extended conformation , and resulting in a distance mismatch that is only partially accommodated by curvature of the RNA-binding surface . A similar constraint may limit the maximum number of contiguous RNA bases bound by tandem PPR motifs . There is no evidence for gaps in the alignments between PLS-class editing factors and their RNA targets . However , the representation of amino acids at position 6 differs between P and S versus L-type PPR motifs . Thus , we suspect that L motifs do not bind nucleotide bases , allowing a ‘mini-gap’ every third nucleotide that may relax the structural constraints . The well-defined code for RNA recognition by Puf domains provides a means to engineer proteins to bind specified RNA sequences . Results presented here imply that PPR tracts could be exploited for similar purposes . In fact , PPR tracts may well offer functionalities beyond those achievable with engineered Puf domains due to their more flexible architecture . Unlike Puf domains , whose 8-repeat organization is conserved throughout the eucaryotes , natural PPR proteins have between 2 and ∼30 repeats and rapidly evolve to bind new RNA sequences and fulfill new functions ( reviewed in [9] ) . The unusually long surface for RNA interaction that is presented by long PPR tracts has the potential to sequester an extended RNA segment , which can impact RNA function in novel ways [12] . PPR proteins play essential roles in all eucaryotes by enabling the expression of specific mitochondrial and chloroplast genes . Even for well-studied PPR proteins such as human LRPPRC ( e . g . [8] ) , the exact binding sites still await discovery . The results and approaches described here offer the potential to eliminate this bottleneck by permitting candidate sites to be postulated from simple sequence analysis , providing information that will have broad application in the medical and agricultural sciences .
rPPR10 and its variants were expressed in E . coli and purified as in [11] . In brief , mature PPR10 ( lacking the plastid targeting peptide ) was expressed as a fusion to maltose binding protein ( MBP ) , purified by amylose affinity chromatography , separated from MBP by cleavage with TEV protease , and further purified by gel filtration chromatography in 250 mM NaCl , 50 mM Tris-HCl pH 7 . 5 , 5 mM ß-mercaptoethanol . The elution peak was diluted in the same buffer for AUC , or dialyzed against 400 mM NaCl , 50 mM Tris-HCl pH 7 . 5 , 5 mM ß-mercaptoethanol , 50% glycerol prior to use in RNA binding assays . PPR10 variants were obtained by PCR-mutagenesis using the following primers ( lower case indicates mutations ) : TD Variant: 5′ GGTCTGTTGCCAgACGCATTCACG; 5′ CGTGAATGCGTcTGGCAACAGACC; 5′ GCTGTGACGTACAcCGAGCTCGCCGGAACG ; 5′ CGTTCCGGCGAGCTCGgTGTACGTCACAGC ; 5′ CACCTGGAGCAACGCGgTGTACGTGACGACGCAC . TN Variant: 5′ CGTGAATGCGTtTGGCAACAGACCC; 5′ GGGTCTGTTGCCAaACGCATTCACG ; 5′ GAACGGCTGCCAGCCAaAcGCTGTGACGTAC ; 5′ CGgTGTACGTCACAGCgTtTGGCTGGCAGCCG . NN Variant: 5′ GGAGCAGAACGGCTGCCAGCCAaacGCTGTGACG; 5′ CGTCACAGCgttTGGCTGGCAGCCGTTCTGCTCC . ND Variant: 5′ GGTCTGTTGCCAgACGCATTCACG; 5′ CGTGAATGCGTcTGGCAACAGACC . NS Variant: 5′ GCTGCCAGCCAagcGCTGTGACG; 5′ CGTCACAGCgctTGGCTGGCAGC;5′ GTCTGTTGCCAagcGCATTCACGTACAACACC; 5′ GGTGTTGTACGTGAATGCgctTGGCAACAGAC SV-AUC was performed in a Beckman Optima XL-I ultracentrifuge with a Beckman An60Ti rotor . 400 µl of sample and 410 µl of reference buffer were analyzed in a 1 . 2 cm double-sector standard AUC cell . Experiments were run at 20°C at 50 , 000 rpm and monitored with an interference optical system . Data were collected at 3 min intervals for 8 hrs , and analyzed with SedFit [28] , using a partial specific volume for rPPR10 of 0 . 73543 calculated from its amino acid composition . The residuals in all experiments were randomly distributed , and 95% of the residuals had a value <10% of the signal . The alignment of PPR10 to its atpH binding site was generated de novo as follows . Thirty-five 17-mers were constructed , each corresponding to the amino acids at a specific position within the 17 sequential PPR motifs in PPR10's interior . Terminal PPR motifs were excluded , as they have distinct properties that may adapt them to their terminal position . These 17 motifs can be arranged in 420 different ways on the 24-nucleotides that are protected by PPR10 , assuming that all the motifs contact the RNA sequentially but not necessarily contiguously , and permitting gaps of any length at any position . The number of arrangements is doubled if both polarities of the protein on the RNA are considered . For each of the 840 arrangements , contingency tables were constructed for each of the 35 17-mers , scoring the number of co-occurrences of each possible amino acid/nucleotide pair ( i . e . a total of 29400 20×4 tables ) . Fisher's Exact Test was used to test for independence of amino acid and nucleotides classes , as implemented in R version 2 . 14 . 2 by fisher . test . The tables were ranked by p-value . The top ranked alignment ( 1/29400 ) was for position 1 . The best alignment for position 6 was also retained ( ranked 71/29400 ) . No other highly ranked alignments were physically compatible with the motif arrangement required for the alignment shown in Figure 2A ( i . e . contained a gap of the same length in the same place ) . The Figure 2A alignments are empirically supported by the boundaries of the PPR10 footprint and minimal binding site , by covariations among PPR10 orthologs and their binding sites , by natural variation in the central region of PPR10's two native binding sites , and by binding affinities of PPR10 for variant atpH sites with various insertions and point mutations [12] . | RNA binding proteins dictate RNA fate and function by modulating RNA processing , localization , translation , and stability . The consequences of RNA/protein interactions are determined , in part , by the position at which the protein binds the RNA . However , it is impossible to predict the target sites of most RNA binding proteins or to design them to bind chosen RNA sequences . In contrast , we show that the pentatricopeptide repeat ( PPR ) protein family holds exceptional promise for the rational design of specified RNA–binding properties . PPR proteins harbor tandem arrays of a repeating structural unit that form a surface for binding single-stranded RNA . We show that PPR tracts bind specific RNA nucleotides via the combinatorial action of two amino acids in each repeat . This mechanism mimics the simplicity and predictability of the Watson-Crick pairing between nucleic acid strands , but at a protein/RNA interface . Our findings will facilitate the prediction of binding sites for the large number of PPR proteins found in nature . Additionally , our demonstration that a PPR tract can be engineered to bind specified RNA sequences implies that PPR proteins can be designed to bind desired RNA targets for applications in biotechnology , medicine , and basic research . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"biotechnology",
"biochemistry",
"plant",
"science",
"molecular",
"cell",
"biology",
"plant",
"biology",
"genetics",
"biology",
"evolutionary",
"biology",
"agricultural",
"biotechnology",
"genetics",
"and",
"genomics",
"agriculture"
] | 2012 | A Combinatorial Amino Acid Code for RNA Recognition by Pentatricopeptide Repeat Proteins |
BLyS/BAFF is recognized for its role in B-cell ontogenesis , as well as cell fate decision towards the first-line/innate marginal zone ( MZ ) B-cell pool . Excess BLyS/BAFF is associated with hyperglobulinemia and increased frequencies of activated precursor-like MZ B-cells . Herein , we show that HIV highly-exposed seronegative ( HESN ) commercial sex workers ( CSWs ) had lower soluble BLyS/BAFF levels and relative frequencies of BLyS/BAFF expressing cells in their genital mucosa when compared to those from HIV-infected CSWs and HIV-uninfected non-CSWs . Furthermore , we identified genital innate and/or marginal zone ( MZ ) -like CD1c+ B-cells that naturally bind to fully glycosylated gp120 , which frequencies were lower in HESNs when compared to HIV-infected CSWs and HIV-uninfected non-CSWs . Although genital levels of total IgA were similar between groups , HESNs had lower levels of total IgG1 and IgG3 . Interestingly , HIV-gp41 reactive IgG1 were found in some HESNs . Low genital levels of BLyS/BAFF observed in HESNs may allow for controlled first-line responses , contributing to natural immunity to HIV .
Worldwide , most HIV infections are acquired through heterosexual intercourse , and in sub-Saharan Africa , 60% of new HIV infections affect women [1] . Observations made in the context of natural immunity to HIV may help identify important clues for the development of protective devices . As such , we established a cohort of female commercial sex workers ( CSWs ) , in Cotonou ( Benin ) , in which we have identified HIV highly-exposed seronegative ( HESN ) individuals , who remain uninfected after more than 4 years of active prostitution . Beninese HESN CSWs have significantly lower genital levels of pro-inflammatory cytokines and chemokines when compared to both HIV-infected CSWs and HIV-uninfected non-CSWs [2 , 3] . Previous studies from Kenyan female CSWs demonstrated that HESNs have a low activation T-cell profile in both the blood and vaginal mucosa , which corresponds with a greater ability to proliferate in response to HIV-p24 peptides when compared to HIV-infected CSWs [4–7] . Furthermore , we and others have demonstrated elevated frequencies of T-regulatory lymphocytes in the blood [8] and genital tract [9] of HESN CSWs , the latter which were concomitant with increased frequencies of dendritic cells ( DC ) bearing a tolerogenic profile . Altogether , these findings suggest that the capacity to regulate the activation/inflammatory profile is associated with protection against HIV infection . Consistent with their low-inflammatory profile , we recently reported that Beninese HESNs have lower levels of B Lymphocyte Stimulator ( BLyS/BAFF ) in their blood when compared to HIV-uninfected non-CSWs [10] . BLyS/BAFF is highly recognized for its role in B-cell ontogenesis , as well as cell fate decision towards the first-line/innate marginal zone ( MZ ) B-cell pool [11 , 12] . As such , HESNs have reduced frequencies of mature MZ B-cells in their blood when compared to HIV-uninfected non-CSWs [10] . In contrast , HIV-infected CSWs have higher levels of BLyS/BAFF , hyperglobulinemia and increased frequencies of activated precursor-like MZ B-cells in their blood when compared to those in HESNs [10] . These findings suggest that control of BLyS/BAFF and innate B-cell status could play a role in natural immunity against HIV infection . Based on these observations , we have now assessed BLyS/BAFF expression levels and innate B-cell status in the genital tract of these women , which is a main portal of entry for HIV . We have been using the “lipid presenting” MHC class I-like molecule CD1c [13] , which is a marker shared by “innate-like” populations , to help track MZ-like B-cells [11 , 12] . In the present study , we show that as for blood , HESNs have lower levels of BLyS/BAFF and MZ-like CD1c+ B-cells in their genital tract when compared to HIV-infected CSWs and HIV-uninfected non-CSWs .
Female CSWs were recruited through a dedicated sex worker clinic in Cotonou , Benin . HIV-uninfected non-CSW control women at low risk for exposure were enrolled from a general health clinic in Cotonou . Women were invited to participate in the study as they attended clinics . Women were excluded from the study if , they were less than 18 years old , menstruating or pregnant . At enrolment , participants were asked to answer a questionnaire about demographic information , sexual behavior , duration of sex work , number of sex partners , condom use , vaginal douching practices , and reproductive history . Each participant underwent a genital examination by a physician . Vaginal specimens were obtained for diagnosis of candidiasis , trichomoniasis and bacterial vaginosis by microscopic examination and herpes simplex virus ( HSV ) infection by PCR . Endocervical swabs were obtained to test for Neisseria gonorrhoeae and Chlamydia trachomatis infection using BD ProbeTec ET system ( Strand Displacement Assay , Becton Dickinson , Heidelberg , Germany ) . Peripheral blood was taken for HIV , syphilis , HSV and progesterone testing by immunoassays . HIV-1 positivity was defined by the presence of HIV specific IgG tested with Vironostika HIV Uni-Form II Ag/Ab ( Organon Teknika , Boxtel , The Netherlands ) . Non-reactive samples were considered HIV seronegative , whereas reactive samples were tested with Genie II HIV-1/HIV-2 ( Bio-Rad , Hercules , CA ) . Genie II dually reactive samples ( to HIV-1 and HIV-2 ) and discordant samples ( Vironostika reactive/Genie II non-reactive ) were further tested by INNO-LIA HIV I/II Score ( Innogenetics NV , Technologiepark 6 , Gent , Belgium ) . HSV infection and shedding was determined by testing for the presence of HSV in the CVLs of the women by PCR assay . For the present study , we selected samples from 10 HIV-uninfected and 11 treatment-naïve HIV-infected CSWs , and 10 HIV-uninfected non-CSW control subjects from the general population . None of these women were injecting drug users . The three study groups were all in the follicular phase of their menstrual cycle , as determined by blood progesterone levels , not taking oral contraception or injectable contraception such as DMPA or implanted ring , had no HSV , N gonorrhoeae , C trachomatis infection , bacterial vaginosis , trichomoniasis or candidiasis . Written informed consent was obtained from all subjects who participated in the study . The methods reported in this paper were performed in accordance with the relevant guidelines and regulations and all experimental protocols were approved by the Comité National Provisoire d’Éthique de la Recherche en Santé in Cotonou and the Centre Hospitalier de l’Université de Montréal ( CHUM ) Research Ethics Committees . Mucus was removed initially prior to performing the Cervico-vaginal lavage ( CVL ) . CVL samples were obtained from all study participants by a physician , using a 10-ml syringe filled with sterile 1x phosphate-buffered solution ( PBS ) and aimed directly into the cervical os . CVL fluids were then collected , transferred immediately into 20 ml of RPMI-1640 , kept on ice , and processed within 1 hour . CVL samples were centrifuged at 1500 rpm for 10 min and supernatants were concentrated on a 3 KDa Amicon membrane and stored at -80°C . The CVL cellular fractions were cryopreserved in liquid nitrogen . BLyS/BAFF levels were determined by using a commercial ELISA kit , R&D systems ( Minneapolis , USA ) . CVL cells were thawed and washed with RPMI 1640 followed by 1X PBS . Briefly , a maximum of 2×105 cells per well were used for staining . Live/dead exclusion was performed using Aqua-LIVE/DEAD Fixable Stain ( Invitrogen Life technologies , Eugene , OR , USA ) . Non-specific binding sites were blocked using fluorescence-activated cell sorting ( FACS ) buffer ( 1x PBS , 2% heat inactivated ( hi ) -FBS , and 0 . 1% sodium azide ) supplemented with 20% hi-FBS and 10 ug mouse IgG ( Sigma-Aldrich , St-Louis , MO , USA ) . CVL cells were stained using the following conjugated mouse anti-human monoclonal antibodies: BUV395 anti-CD45 and BV786 anti-CD14 ( BD-Biosciences , San Jose , CA , USA ) , PeCy5 . 5 anti-CD11c , PeCy7 anti-CD66b , AlexaFluor 700 anti-CD3 and PE anti-BLyS ( ebiosciences , San Jose , CA , USA ) , APC CDK PAN cytokeratin for epithelial cells ( Cedarlane , Burlington , ON , CA ) . CVL cells were fixed with 1 . 25% paraformaldehyde and kept at 4°C for a minimum of 12 hours before flow-cytometry analysis . Briefly , live epithelial cells and leucocytes were analyzed after FSC/SSC gating to remove debris , and removal of doublets , for HIV-infected CSWs: epithelial cells constituted 18% and leukocytes 48% of recovered live cells , which summed to a mean of 17 336 ±6743 total events . For HESNs: epithelial cells constituted 9 , 3% and leukocytes 37% of recovered live cells , which summed to a mean of 14 867±8938 total events . For HIV-uninfected non-CSWs: epithelial cells constituted 8% and leucocytes 26% of recovered live cells , which summed to a mean of 5650±992 total events . Acquisition was with an LSRFortessa ( BD-Biosciences , San Jose , CA , USA ) and analyzed with FlowJo7 . 6 . 3 software ( TreeStar , Ashland , OR , USA ) . Flow-cytometry data analysis quadrants were set based on the expression values obtained with fluorescence minus one ( FMO ) and isotype controls . Cell processing , staining and analysis were performed as mentioned above . The following conjugated mouse anti-human monoclonal antibodies were used: BUV395 anti-CD45 , BUV737 anti-CD138 , BV605 anti-CD19 , APC/H7 anti-IgG and APC anti-CD1a ( BD-Biosciences , San Jose , CA , USA ) , PerCP efluor710 anti-CD1c and AlexaFluor 700 anti-CD3 ( ebiosciences , San Jose , CA , USA ) . Cells were pre-incubated or not with mannose ( 5 ug ) for 40 minutes on ice , followed by incubation with or without fully glycosylated biotinylated gp120 IIIB ( ImmunoDX Inc ) at 5 ug/ml for 40 minutes on ice prior to adding the staining cocktail and streptavidin-PE ( BD-Biosciences , San Jose , CA , USA ) . Levels of total immunoglobulin ( Ig ) isotypes IgG1 , IgG2 , IgG3 , IgG4 , IgM and IgA were measured in CVL supernatants using the multiplex bead assay Milliplex Map Kit with human immunoglobulin isotyping Magnetic Bead panel by EMD Millipore ( Billerica , USA ) according to manufacturer’s protocol . Analysis was performed on a Luminex 200 System ( Luminex Corporation , Austin , TX , USA ) . HIV-gp120 and -gp41 Ig reactivity was detected based on the method previously described [14 , 15] . Briefly , non-concentrated CVL supernatants were incubated for 18 hours at 4°C rotating with protein G-agarose ( ThermoFisher ) , and eluted with elution buffer ( ThermoFisher ) to recover IgG . Subsequently , remaining supernatants were incubated for 18 hours at 4°C rotating with peptide M-agarose ( Invivogen ) and eluted to recover IgA . IgG and IgA recovery in eluates and presence within remaining supernatants were verified by performing human total IgG and IgA ELISAs ( ThermoFisher ) . The remaining supernatants following IgA recovery were used to detect IgM reactivity . Eluates were neutralized with TRIS 1M pH 7 . 5 ( ThermoFisher ) and incubated with gp120 M . CONS-D11 and MN gp41 ( NIH AIDS-Reagent program ) coated magnetic microspheres ( Radix ) for 18 hours at 4°C rotating , followed by incubation with either of PE conjugated mouse anti-human IgG1 , IgG2 , IgG3 , IgG4 ( Southern Biotech ) , IgA1 , IgA2 and IgM ( ebioscience ) , and detection by Luminex 200 system . The cut–off for positivity was set at mean fluorescence intensity value obtained for 10 HIV-uninfected non-CSWs + 3 SD . Data from HESNs were compared separately to those of HIV-infected CSWs and HIV-uninfected non-CSWs . The statistical significance of difference between groups was determined by Fisher’s exact test for categorical variables and Unpaired T-test or Mann-Whitney U test analysis for continuous variables . The D’Agostino-Pearson normality test was used to determine whether the values were sampled from a Gaussian distribution . Analyses were performed using GraphPad Prism 5 . 00 for Windows ( GraphPad Software , San Diego , California , USA ) .
The socio-demographic characteristics of female CSWs and non-CSWs are shown in Table 1 . There were no statistical differences for age between HESNs and the two other groups . All women were practicing vaginal douching . Duration of sex work , average number of clients and condom use were not significantly different between the HESN and HIV-1-infected CSW groups . Thus CSWs and non-CSWs were comparable in terms of socio-demographic characteristics .
We [2 , 3 , 9] and others [8 , 18 , 19] have shown that HESN female CSWs have lower genital inflammation when compared to both HIV-infected CSWs and HIV-uninfected non-CSWs . We hypothesized that maintenance of low-inflammatory conditions in the female genital tract of HESN individuals may help to prevent excessive immune activation and lower HIV target availability , likely maintaining the integrity of the mucosal barrier to protect from HIV infection [4 , 20] . In agreement with this , HESNs had lower blood [10] and genital levels of soluble BLyS/BAFF and lower frequencies of BLyS/BAFF expressing cells in their genital mucosa . In contrast , the genital cells of HESNs expressed relatively higher levels of BLyS/BAFF than the cells of the other groups of women . These observations are suggesting that upregulation of BLyS/BAFF expression seems required , but regulated as to prevent deleterious effects . Recent studies have shown that plasmacytoid DCs exposed to HIV in vitro upregulate BLyS/BAFF cell surface expression without releasing the molecule [21] . This raises the possibility that the low levels of BLyS/BAFF measured in blood [10] and CVL supernatants of HESNs may be linked to the signals leading to BLyS/BAFF release . As to whether these are related to advantageous genetic polymorphisms remain to be established . We have recently analyzed BAFF promoter -871 , -2841 and -2701 mutations associated with elevated BLyS/BAFF plasma levels and susceptibility to auto-immune diseases such as Systemic Lupus Erythematosus and hepatitis C associated cryoglobulinemia [22–24] in our Benin cohort and found no association between BAFF promoter mutations and either blood and CVLs BLyS/BAFF levels nor HIV infection ( S4 Fig ) . The relatively high levels of BLyS/BAFF observed in the blood [10] and CVL supernatants of HIV-infected CSWs are consistent with our previous reports for HIV-infected rapid and classic progressors [16] , and likely due to direct and indirect factors associated with HIV infection [20] . Plasma [10] and CVL levels , but not cell surface expression , of BLyS/BAFF measured in HIV-uninfected non-CSWs were similar to those observed in HIV-infected CSWs . This may be due to inflammatory/infectious conditions other than HIV in HIV-uninfected non-CSWs that can stimulate soluble release of BLyS/BAFF [10] . Consistent with lower levels of soluble BLyS/BAFF , HESNs had reduced frequencies of innate MZ-like CD1c+ B-cells in their genital tract when compared to HIV-infected CSWs and HIV-uninfected non-CSWs . Growing importance is given to innate MZ B-cells in health and disease [11] , as they constitute early first-line defense against invading pathogens and participate in the development of adaptive antibody ( Ab ) responses by trafficking to follicular B-cell areas of lymphoid structures and promoting germinal center reactions [25] . MZ B-cells are capable of isotype switching and can present a somatically mutated pre-diversified low affinity polyreactive BCR repertoire [11] , which comprises usage of the IGHV1-2 gene [26] , shown to take part in HIV-ENV reactive broadly neutralizing Abs ( bNAbs ) such as VRC01 [27] . Interestingly , repeated treatment of mice with BLyS/BAFF increased their MZ compartment , and generated an increased response to ENV immunization and bNAbs [28] . In agreement with the observations made by Cerutti and colleagues [17] , the innate/MZ-like CD1c+ B-cell populations we identified in the genital tract of Beninese women also bind to fully glycosylated gp120 , and to a greater frequency than CD1c- B-cells . Although this suggests that these cells have the capacity to transfer HIV to target cells , it is unlikely that they get infected by the virus since it has not yet been convincingly shown to infect or replicate in B-cells in vivo [29] . Interestingly , although they have lower frequencies of CD1c+ MZ-like B-cells , HESNs have higher relative frequencies of total B-cells binding gp120 when compared to the other groups . The fact that there were no significant differences in frequencies of CD1c+ B-cell sub-populations binding gp120 between the different groups suggests that in HESNs , genital B-cells other than CD1c+CD1a- and CD1c+CD1a+ subsets have a greater capacity to bind gp120 that those in HIV-infected CSWs and HIV-uninfected non-CSWs . It is possible that the relative binding capacity was lower in the HIV-infected CSWs because gp120 receptors were saturated in these individuals . In contrast to that observed by He et al [17] , pre-incubation of total B-cells with mannose did not significantly diminish gp120 binding ( S2B Fig ) , suggesting receptors of various types might be involved . Identifying these receptors and B-cell sub-population ( s ) binding gp120 that are increased in HESNs will require further experimentation . Also , the exact nature of the innate MZ-like CD1c+ B-cells we identify in the genital tract has yet to be confirmed , and as to whether they have a direct link with those we previously observed in blood [10] remains to be established . In humans , MZ B-cells recirculate and have been found in front-line areas such as the sub epithelial lamina propria of mucosal associated lymphoid tissues ( MALT ) [11] . To our knowledge , we show for the first time that MZ-like B-cells can be found in the female genital tract , which is part of the MALT and is populated by a commensal microflora [4 , 30] . It is thus conceivable that the innate MZ-like CD1c+ B-cells observed in the genital tract of Beninese women participate in local immune responses to control microflora and pathogens . Moreover , it has been shown that the gut lamina propria can be a T-independent inductive site in humans [31] and likely similar mechanisms operate at the genital lamina propria . The recent characterization of elevated BLyS/BAFF levels and transient Gp41-specific IgA in mucosal genital fluids from patients within the first weeks after HIV transmission , suggest that these Abs might have originated from first-line B-cell populations [32] . The fact that HESNs who undergo sex-break eventually seroconvert [33] , suggests natural immunity involves populations of which pool maintenance in the genital mucosal niche requires frequent antigen exposure , and this is consistent with first-line responses . As to whether first-line responses are actually polyreactive with shared HIV-specificity and/or involve shared antigenicity with the local microbiota remains to be determined . Depending on the level of inflammation , CD1c+ B-cells may contribute to natural immunity against HIV or conversely promote disease progression . Indeed , as shown previously [10 , 16] , elevated BLyS/BAFF levels increase expansion , activation and dysregulation of innate B-cell populations such as precursor-like MZ B-cells , likely contributing to the over-representation of low affinity , polyreactive and auto-reactive Abs [34] at the expense of high affinity polyfunctional eradicating anti-HIV Ab responses . As such , we found hyperglobulinemia in the blood [10] and CVL supernatants of HIV-infected CSWs . Although we found no significant difference in the frequencies of total or IgG+ plasmablasts/plasma cells in the genital tract of Beninese women , relative percentages of total CD138+CD1c+CD1a+ cells bearing IgG were significantly higher in HIV-infected CSWs when compared to HESNs . It has yet to be determined whether these cells are substantially involved in the relatively high IgG1 and IgG3 levels measured in the CVLs of HIV-infected CSWs . Most genital immunoglobulins ( Ig ) are found in the mucus [35] , unfortunately the latter was removed prior to CVL sample collection in our study . Nevertheless , IgG1 and IgA1 reactivity to both gp120 and gp41 , as well as IgG2 , IgG3 , IgA2 and IgM reactivity to gp41 were observed in CVL supernatants of the majority of HIV-infected CSWs . However , despite the elevated frequencies of B-cells binding to gp120 , we found no Ig reactivity to gp120 in the CVL supernatants of HESNs . It is possible that lower levels of Ig are present in the samples of HESNs but mucus removal has precluded their detection . Accordingly , we have previously detected anti-HIV-1-Env-specific IgG , neutralizing or ADCC activities in blood and CVL samples from HIV-infected CSWs but not in those from HESNs [36] . Interestingly , we could detect IgG1 reactivity to gp41 in some HESNs , which could be derived from a microbiota reactive , possibly first-line B-cell pool [37] , as most gp41 reactive Abs cross-react with microbiota [38] . Whether the gp41 binding IgG1 Abs detected in the CVL of HESNS can confer some level of protection remains to be established . There is increasing evidence for non-neutralizing functions of antibodies in decreasing the viral load , and in conferring some level of protection [39] . In this view , anti-gp41 IgG antibodies are found in the plasma of HIV-infected individuals shortly after transmission , and form antibody-virion complexes , which although ineffective at controlling disease progression [40] , have been associated with infectivity decay [41] . We could not detect substantial IgA1 and IgA2 reactivity to gp120 or gp41 in the CVL supernatants of HESNs . To date , studies have reported contradictory results regarding the presence of anti-HIV specific IgA responses in the genital tract of HESNs [42–47] . The discrepancies between studies may be due to relatively small sample size of these studies and/or the different techniques used to detect ENV-reactive Abs . Because of the cross-sectional design , the present study cannot address whether the lower levels of BLyS/BAFF and CD1c+ B-cells , as well as gp41 reactive IgG1 found in the genital tract of HESNs have a protective role against HIV infection . Comparison between HESN and women involved in sex work but not yet HESN should also be done to control the effects of sex work itself on genital immunology . Longitudinal studies and further phenotypic and functional characterizations are required to confirm a protective role , and the exact nature of genital CD1c+ B-cells and their responses . To wrap-up , understanding the dynamics of BLyS/BAFF and its role in homeostasis of immune responsiveness appears pivotal to the design of vaccine strategies soliciting first-line B-cell responses to help protect from HIV infection . Based on our observations , the capacity to contain BLyS/BAFF expression levels seems concomitant with natural immunity against HIV , whereas excessive BLyS/BAFF may promote immune dysregulation , risk of infection and disease progression . The fact that human genital innate MZ-like B-cells naturally bind to fully glycosylated gp120 renders these cells of particular interest because MZ B-cells can acquire Ig somatic mutations and could be harnessed to increase HIV-ENV affinity . | Worldwide , most human immunodeficiency virus ( HIV ) infections affect women through heterosexual intercourse . We and others have identified African female commercial sex workers ( CSWs ) , who remain seronegative despite high exposition to HIV ( HESNs ) . Innate marginal zone ( MZ ) B-cells recirculate in humans and have been found in front-line areas such as the sub-epithelial lamina propria of mucosal associated lymphoid tissues . MZ B-cells can bind to fully glycosylated gp120 and produce specific IgG and IgA , and have a propensity for B regulatory potential , which could help both the fight against HIV and maintenance of low inflammatory conditions reported for HESNs . Here we identify genital MZ-like B-cells , which frequencies are lower in the genital tract of HESNs when compared to HIV-infected CSWs and HIV-uninfected non-CSW women . Furthermore , this coincides with significantly lower genital levels of B lymphocyte stimulator ( BLyS/BAFF ) , known to shape the MZ pool and which overexpression leads to MZ deregulation in HIV-infected progressors . HESN individuals provide an exceptional opportunity to determine important clues for the development of protective devices . Here we show that contained BLyS/BAFF levels are concomitant with natural immunity against HIV , and may prevent dysregulated first-line responses . MZ-like B-cells could be harnessed in preventive strategies viewed at soliciting quick first-line to be adjunct to matured long term protection . | [
"Abstract",
"Introduction",
"Materials",
"and",
"methods",
"Results",
"Discussion"
] | [
"blood",
"cells",
"hiv",
"infections",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"body",
"fluids",
"pathology",
"and",
"laboratory",
"medicine",
"immune",
"physiology",
"pathogens",
"immunology",
"microbiology",
"retroviruses",
"viruses",
"immunodefici... | 2019 | Natural Immunity to HIV is associated with Low BLyS/BAFF levels and low frequencies of innate marginal zone like CD1c+ B-cells in the genital tract |
Adhesion and detachment are coordinated critical steps during cell migration . Conceptually , efficient migration requires both effective stabilization of membrane protrusions at the leading edge via nascent adhesions and their successful persistence during retraction of the trailing side via disruption of focal adhesions . As nascent adhesions are much smaller in size than focal adhesions , they are expected to exhibit a stronger adhesivity in order to achieve the coordination between cell front and back . Here , we show that Nudel knockdown by interference RNA ( RNAi ) resulted in cell edge shrinkage due to poor adhesions of membrane protrusions . Nudel bound to paxillin , a scaffold protein of focal contacts , and colocalized with it in areas of active membrane protrusions , presumably at nascent adhesions . The Nudel-paxillin interaction was disrupted by focal adhesion kinase ( FAK ) in a paxillin-binding–dependent manner . Forced localization of Nudel in all focal contacts by fusing it to paxillin markedly strengthened their adhesivity , whereas overexpression of structurally activated FAK or any paxillin-binding FAK mutant lacking the N-terminal autoinhibitory domain caused cell edge shrinkage . These results suggest a novel mechanism for selective reinforcement of nascent adhesions via interplays of Nudel and FAK with paxillin to facilitate cell migration .
In order to achieve efficient migration , cell adhesion and detachment must be properly coordinated . Cells attach to the substratum via punctate focal contacts ( FCs ) . FCs contain integrin family members of transmembrane receptors and a variety of intracellular “adhesion” proteins and function to connect the extracellular matrix ( ECM ) to the actin cytoskeleton [1] , [2] . During cell migration , membrane protrusions at the leading edge are triggered after activation of the Rho-family small GTPases Cdc42 and Rac1 [3] . Activated integrin dimers situated at the tip of protrusions then search for and bind to their ligands in the ECM to form nascent adhesions [4] . Nascent adhesions can mature into focal complexes ( FXs ) , which are located mainly at the base of lamellipodium [5] , [6] . FXs associate with branched F-actin and are thought to facilitate the propulsive effect of the lamellipodium . Some FXs then further evolve into the largest form of FC , namely focal adhesion ( FA ) . FAs associate with the termini of F-actin bundles , or stress fibers , which provide cells with contractile forces [1] , [6]–[8] . It is known that only moderate concentrations of the ECM are associated with maximal cell motility [9] , [10] . Conceptually , fast migration would require efficient adhesion of leading-edge protrusions and rapid retraction of the trailing side [10] , [11] . These two requirements could be satisfied if nascent adhesion sites exhibit stronger ECM-binding strengths than adhesion sites in FXs as well as FAs . Stronger adhesions at nascent sites would both promote the stabilization of membrane protrusions and facilitate persistency of the leading edge during cell retraction to allow efficient forward movement . In fact , tension on each contact site within FAs , which are caps of stress fibers [1] , [7] , is relatively constant in a cell [12] . Therefore , smaller FAs can only transmit weaker traction forces [12] . In contrast , compared to FAs , nascent adhesions , despite their submicroscopic sizes , have been shown to transmit stronger traction forces [13] . This would physically require a stronger integrin-ECM interaction at nascent adhesion sites than at adhesion sites of FXs and FAs . Whether mechanical strengths of different adhesion sites are indeed modulated and the underlying molecular mechanism ( s ) , however , remain unclear . FCs are dynamic structures . They are assembled through a hierarchical process . Paxillin and talin are believed to bind directly to integrin at adhesion sites [2] . Other proteins such as vinculin and focal adhesion kinase ( FAK ) are then recruited to form dot-like FXs , whereas FA formation is accompanied by the appearance of zyxin [6] , [14] . FAK can be activated by multiple signaling pathways and is crucial for FC dynamics and membrane protrusion [2] . Its FA-targeting ( FAT ) domain , located at the C-terminus , interacts with talin and the LD domains of paxillin [2] , [15] . In addition to assembly , FCs are subjected to dynamic disassembly as well [7] . Both nascent adhesion sites and FXs can be rapidly disassembled if they failed to evolve [6] , [14] . FAs are relatively long-lived . Their disassembly often occurs at the trailing side of migrating cells . Moreover , FA formation can be promoted by internal and external tensions [12] , [16]–[18] . Tensions on stress fibers can also lead to a net disassembly of distal adhesion sites and assembly of proximal sites , resulting in centripetal movement of FAs [19] . Mammalian Nudel ( also named Ndel1 ) and Lis1 are essential for cell viability [20] , [21] and for functions of the microtubule ( MT ) -based , minus end–directed motor cytoplasmic dynein in diverse processes including mitosis , neuronal migration , and intracellular transport [20] , [22]–[27] . In addition , Nudel can also stabilize active Cdc42 by sequestering a negative regulator , Cdc42GAP , at the leading edge during migration of NIH3T3 cells [28] . Nudel confers homodimerization and Lis1 binding through its N-terminal coiled-coil region , whereas its C-terminus is able to interact with dynein heavy chain , Cdc42GAP , and other proteins [23] , [26] , [28]–[30] . In this report , we describe a novel mechanism we identified that regulates adhesivity of integrin-mediated adhesions . Our results indicate that Nudel selectively strengthens FC sites in nascent adhesions through a direct interaction with paxillin to facilitate stabilization of membrane protrusions at the leading edge , whereas structurally activated FAK can displace Nudel from paxillin in a kinase-independent manner , thus reducing the strength of FC sites in FXs and FAs to promote retraction of the trailing side .
We have previously shown that Nudel knockdown markedly inhibited pseudopodial formation in mouse fibroblast NIH3T3 cells [28] . To clarify whether this is solely related to defects in membrane protrusion , human epithelial ECV304 cells were chosen for analysis because they migrated with typical fan-shaped lamellipodia ( Figure 1A; Videos S1 and S2 ) . For convenient identification of live transfectants , the interference RNA ( RNAi ) constructs , pTER-Nudi for Nudel and pTER-Luci as a control [31] , were modified to coexpress green fluorescent protein ( GFP ) or red fluorescent protein ( RFP ) . As in NIH3T3 cells [28] , Nudel RNAi in sparse ECV304 cells significantly repressed membrane protrusions and thus migration ( Figures 1A , S1A , and S1B; Videos S1 and S2 ) . Overexpression of Nudel with an RNAi-resistant construct ( Nudel-R ) rescued both lamellipodial formation and cell motilities ( Figure S1C–S1E ) , thereby excluding a possible off-target effect of the RNAi construct . Nudel RNAi has been shown to cause inactivation of Cdc42 [28] , which could in turn repress Rac1 activity [32] , [33] . If the lack of lamellipodia in Nudel-depleted cells ( Figure 1A ) was simply due to inhibition of Rac1 , introduction of a constitutive active form of Rac1 ( Rac1CA ) should be able to fully restore lamellipodium formation [3] , [34] . Consistent with a previous report [34] , 76% of GFP-Rac1CA–positive cells cotransfected with pTER-Luci-RFP ( n = 233 ) became flat and circular in shape , due to extensive formation and spreading of lamellipodia ( Figure 1B , panels 1 and 2 ) . In contrast , although 67% of pTER-Nudi-RFP transfectants overexpressing GFP-Rac1CA ( n = 316 ) formed lamellipodia , as judged by the existence of F-actin–rich membrane ruffles , they failed to spread extensively ( Figure 1B , panels 3 and 4 , arrows ) . Quantitation also indicated that they generally exhibited obviously reduced circularity and area as compared to control cells ( Luci+Rac1CA ) ( Figure 1C ) . To corroborate these results , we applied a dominant-negative Cdc42 ( Cdc42DN ) to repress Cdc42 activity ( unpublished data ) [35] and found that as expected , its overexpression failed to repress cell spreading stimulated by Rac1CA ( Figure 1B and 1C ) . Therefore , the spreading defect associated with Nudel depletion is not solely due to inhibition of Cdc42 and Rac1 . We then performed time-lapse microscopy to examine why Nudel-depleted cells failed to fully spread even in the presence of Rac1CA . The control transfectants , which were much larger in size than surrounding untransfected cells , showed vigorous membrane ruffling at cell edges ( Figure 1D; Video S3 ) [34] . In contrast , although GFP-Rac1CA induced active membrane protrusions in Nudel-depleted cells ( Figure 1E vs . 1A ) , the protrusions were not persistent and usually retracted back within a few minutes ( Figure 1E; Video S4 ) , indicating lack of stable attachment to the substratum . As a result , the cells failed to spread even when monitored for more than 500 min ( Figure 1E; Video S4 ) . We further excluded the possibility that Nudel RNAi repressed lamellipodial formation through inhibition of dynein because NudelC36 , a deletion mutant whose overexpression inhibits dynein [22] , [23] , failed to affect ECV304 cell migration ( Figure S2A and S2B ) . Normal lamellipodial formation was seen as well in cells overexpressing either GFP-tagged NudelC36 or another dynein inhibitor , p50dynamitin ( Figure S2C ) [36] , [37] . Taken together , these results strongly suggest a critical role of Nudel in stable attachment of nascent membrane protrusions to the substratum . Importantly , such a role is distinct from the previous ones in regulation of Cdc42 and dynein [28] , therefore defining a novel function of Nudel in cell migration . We then examined detailed distributions of FCs and F-actin in ECV304 cells with Nudel knockdown . Indeed , compared to the typical arc-like lamellipodial formation in most sparse transfectants of pTER-Luci-GFP ( 71 . 0%; n = 356 ) ( Figure 2A , panels 1 and 2 ) , transfection with pTER-Nudi-GFP resulted in severe cell edge shrinkage in both subconfluent cells ( 63 . 1%; n = 388 ) and confluent cells scratched to induce migration [38] , [39] ( Figures 2A , panels 3 and 4 , and S3A ) . Moreover , robust FAs and stress fibers at the cell periphery were seen ( Figures 2A , panels 3 and 4 , and S3A ) . Similar phenotypes were also observed in HeLa cells , independent of cell densities ( Figure S3B ) . The FAs/stress fibers can develop in response to forces provided either intrinsically through contraction of myosin on stress fibers or externally by mechanical strains [12] , [17] , [18] , [40] . To better understand the phenotypes of Nudel RNAi , we disrupted the intrinsic contractile forces using blebbistatin , a small-molecule inhibitor of myosin II ATPase activity [41] . After blebbistatin treatment for 45 min , FAs and stress fibers were mostly disassembled in control cells , as expected ( Figure 2B ) [5] , [41] . Nevertheless , they were still largely preserved in Nudel RNAi cells ( Figure 2B ) , suggesting that the robust FAs/stress fibers in Nudel-depleted cells ( Figures 2A and S3 ) were formed in response to tensions from the collapsing cell edges in order to resist further shrinkage , instead of from the contractile forces of myosin II . To understand why cell edges tended to shrink upon Nudel RNAi , we examined FCs in Nudel-depleted cells overexpressing Rac1CA . In control cells , Rac1CA induced typical FX around the entire cell periphery ( Figure 2C and 2D ) [34] . In contrast , although FXs were readily observed in pTER-Nudi-RFP transfectants overexpressing Rac1CA , they only appeared in less than half of the cell periphery in approximately 82% of cells ( Figure 2C and 2D ) , indicating a markedly reduced efficiency of FX formation . We then treated such cells with blebbistatin for 25 min to block maturation of their nascent adhesions into FXs [5] . In contrast to the appearance of a rim of tiny , dense nascent adhesions within the lamellipodium in control cells ( Figure 2E ) [5] , Nudel RNAi cells overexpressing Rac1CA showed little accumulation of nascent adhesions around the cell periphery ( Figure 2E ) , though vigorous membrane protrusions still occurred ( Figure 2E ) as in untreated cells ( Figure 1B , panels 3 and 4 , and 1E ) . Therefore , the negative effect of Nudel RNAi on stabilization of membrane protrusions ( Figures 1 and 2A ) is attributed to poor formation of nascent adhesions . To understand how Nudel could affect nascent adhesions , we performed a screen for its partner ( s ) in FCs . FLAG-Nudel coexpressed with a GFP-tagged FC protein such as vinculin , paxillin , or FAK was subjected to coimmunoprecipitation ( co-IP ) . Ponceau S staining revealed GFP-paxillin as the major protein associated with FLAG-Nudel ( Figure 3A and 3B , lane 6 ) , strongly suggesting a direct interaction . GFP-paxillin was also associated with FLAG-NudelN20 , a mutant lacking Lis1-binding activity [22] , but not with FLAG-NudelC36 ( Figure 3A and 3B , lanes 7 and 8 ) . The failure of NudelC36 to interact with paxillin was also consistent with the results that , unlike the wild-type Nudel ( Figure S1D and S1E ) , NudelC36 overexpressed from an RNAi-resistant construct failed to restore the motility of Nudel-depleted cells ( Figure S2A and S2B ) . FLAG-Nudel was able to associate with Tyr/Ser/Thr-phosphorylated isoforms important for physiological functions of paxillin ( Figure 3C ) [42]–[44] , further suggesting a functional interplay between the two proteins . To confirm their direct interaction , GST-paxillin and FLAG-Nudel were expressed in Escherichia coli . Glutathione S-transferase ( GST ) -pulldown assays indeed indicated their interaction ( Figure 3D , lane 6 ) . Moreover , when paxillin mutants containing either the LD domains or the LIM domains ( Figure S4A ) [15] were assayed , only PaxLIM interacted with Nudel ( Figure 3D , lanes 8 ) . Reciprocal experiments also support a direct Nudel-paxillin interaction ( Figure S4B ) . In contrast , vinculin , a paxillin-associated FC protein [15] , failed to bind directly to Nudel ( Figure S4B ) . As paxillin exists in all types of FCs and is a scaffold/adaptor protein critical for cell migration [6] , [15] , it may serve as the target of Nudel in cell adhesion . Consistently , FLAG-Nudel formed a complex with endogenous paxillin and vinculin ( see below ) . We then examined localization of Nudel and paxillin in ECV304 cells migrating into an artificial “wound” [38] , [39] . As in NIH3T3 cells [28] , Nudel was enriched at the leading edge , and colocalized with paxillin there ( Figure 3E , arrowheads ) . Moreover , both proteins were enriched in areas of cell protrusions , indicated by the presence of active actin polymerization ( Figure 3E ) [4] , [45] . In contrast , Nudel did not show colocalization with the paxillin puncta , which typically represent FXs and FAs ( Figure 3E ) [46] . Quantitation analyses also indicated a significant correlation between Nudel and paxillin at the leading edge ( Figure 3F–3H ) . These results imply interaction of both proteins in early stages of FC formation and are consistent with the role of Nudel in nascent membrane adhesion ( Figures 1 and 2 ) . We then tried to assess whether the Nudel-paxillin interaction indeed contributed positively to nascent cell adhesion . As integrin-mediated nascent adhesion sites are submicroscopic structures and only represented a portion of total adhesion sites ( Figures 2 and 3 ) [7] , direct assays on them would not be feasible . We thus reasoned that a fusion protein , paxillin-GFP-Nudel ( PGN ) , would make all adhesion sites Nudel-containing , thus allowing convenient examination of Nudel's effect on adhesion . Although such a construct is somewhat artificial , a similar strategy has been successfully used in other studies [47] . Similar to Pax-GFP ( Figure 4A and 4B ) [48] , PGN was also located in FCs ( Figures 4A , 4B , S5A , and S5B ) . Moreover , PGN still bound to Lis1 ( Figure S5C ) , a protein associated with the N-terminal portion of Nudel [30] . Therefore , both paxillin and Nudel in the fusion protein are still functional . To verify whether PGN stabilized the cell–substratum adhesion , we first examined FA motilities [19] , which may reflect the stability of individual adhesion sites of FAs against tension . For easy comparison , image sequences at 0 , 10 , and 20 min were pseudocolored red , green , and blue , respectively , and merged . Motile FAs would thus display rainbow colors , whereas nonmotile ones would be white [19] . Upon overexpression of Pax-GFP , FAs in both nonmotile ( Figure 4A ) and motile cells ( Figure 4B ) exhibited similar active centripetal movement , as judged by the appearance and orientation of rainbow colors . The average velocity was 0 . 0434 µm/min ( Figure 4C ) , about 3-fold lower than that of 3T3 fibroblasts [19] . It should be noted that Smilenov and colleagues [19] considered cells just after division as “migrating” cells and defined the remaining population as “stationary” cells . Therefore , the population analyzed herein is equivalent to the “stationary” population in the previous study [19] . Just as in ECV304 cells ( Figure 1 ) , this population of fibroblasts is in fact not truly stationary [28] . In cells overexpressing PGN , FA motilities were largely reduced , as judged by the obvious appearance of white color ( Figure 4A and 4B ) . The average velocity of FAs was reduced by approximately 3-fold ( 0 . 0146 µm/min ) as compared to that in cells overexpressing Pax-GFP ( Figure 4C ) . Moreover , as FCs close to the cell edges where membranes are dynamic showed obvious turnover in PGN-positive cells as well ( Figure 4B ) , the reduced FA motility is unlikely due to defects in FA disassembly . Rather , it suggests an increased strength of the FC sites . To further corroborate the above results , we investigated whether PGN-positive cells exhibited enhanced adhesion on a laminin- and fibronectin-coated surface against different shear forces [49] . HEK 293T cells were used instead of ECV304 because the latter cells tended to aggregate , thus precluding sorting with fluorescence-activated cell sorter ( FACS ) to eliminate untransfected cells . At the wall shear stress of 1 and 2 dyne/cm2 , PGN-expressing cells accumulated more rapidly than Pax-GFP–positive cells ( Figure 4D ) . They maintained a higher resistance to the increasing shear stress from 2 to 16 dyne/cm2 ( Figure 4D ) . Taken together , these results strongly suggest that the Nudel-paxillin association can enhance adhesion strength of FC sites . This further explains why Nudel is critical for stabilization of nascent membrane protrusions ( Figures 1 and 2 ) . As Nudel was not seen in either FXs or FAs ( Figure 3E ) , we speculated that it might be displaced by certain paxillin-binding protein ( s ) that are recruited during the maturation of nascent adhesion sites [6] . Indeed , co-IP results indicated that overexpression of GFP-tagged FAK , but not vinculin , both of which are paxillin-binding FC proteins [15] , abolished the interaction between GFP-paxillin and FLAG-Nudel ( Figure 5A and 5B , lanes 1 and 2 vs . 9 and 10 ) . FAK mutants lacking the paxillin-binding FAT domain [2] , e . g . , FAKΔFAT and FAKKin , were completely ineffective ( Figure 5A and 5B , lane 7 vs . 15; Figure S6A and S6B , lane 1 vs . 4 ) . In contrast , mutants containing FAT or even FAT alone , e . g . , FAKΔFERM , FAKFAT , and FRNK , a naturally occurring FAK isoform [50] , were all potent in disrupting the Nudel-paxillin interaction ( Figure 5A and 5B , lanes 4 and 8 vs . 12 and 16; Figure S6A and S6B , lanes 2 vs . 5 ) . When the paxillin-interacting ability of FAKΔFERM was abolished by point mutations V954A/L961A [51] or I936E/I998E [52] , the resultant mutants , FAKΔFMPX1 and FAKΔFMPX2 , failed to disrupt the Nudel-paxillin interaction ( Figure 5A and 5B , lanes 5 and 6 vs . 13 and 14 ) . In contrast , the kinase-dead mutant FAKKD carrying a K454R mutation [53] still showed strong activity in competing for paxillin binding ( Figure 5A and 5B , lane 3 vs . 11 ) . Therefore , FAK can , in a kinase-independent manner , disrupt the Nudel-paxillin interaction through its physical association with paxillin . To understand how FAK abolishes the Nudel-paxillin interaction , GFP-tagged PaxLD and PaxLIM ( Figure S4A ) [15] were tested for their ability to bind Nudel in the presence of GFP-FAKΔFAT or FAKΔFERM ( Figure 5C ) . As expected ( Figure 5A and 5B ) , the Nudel-paxillin interaction was not affected by FAKΔFAT but was disrupted by FAKΔFERM ( Figure 5C , lanes 3 and 4 vs . 9 and 10 ) . PaxLIM associated with Nudel in both cases ( Figure 5C , lanes 1 and 2 vs . 7 and 8 ) , whereas PaxLD failed to do so in either case ( Figure 5C , lanes 5 and 6 vs . 11 and 12 ) . These data further confirm that Nudel interacts with paxillin via the LIM domains ( Figure 3D ) . Moreover , competition by FAK is mediated through its direct interaction with the LD domains of paxillin [2] , [15] . To further investigate whether FAK indeed regulates the Nudel-paxillin interaction at physiological conditions , we performed co-IP experiments to check whether a decrease in endogenous FAK levels could affect the Nudel-paxillin interaction . As we were not able to detect endogenous paxillin in co-IP experiments using anti-Nudel IgY , possibly due to a steric effect of the antibody , we overexpressed in HEK293T cells low levels of FLAG-GFP-Nudel ( at 3–6-fold of endogenous Nudel level ) through the internal ribosome entry site ( IRES ) and performed co-IP assays with anti-FLAG resin ( Figure 5D ) . The levels of endogenous FAK were reduced sequentially through transfection of increasing amounts of the RNAi plasmids ( Figure 5D , lanes 1–6 ) . Indeed , association of endogenous paxillin with FLAG-GFP-Nudel was markedly enhanced following the reduction of endogenous FAK levels ( Figure 5D , lanes 7–12 ) . Association of vinculin was also detected , with its levels paralleling those of paxillin ( Figure 5D , lanes 7–12 ) . Therefore , endogenous FAK can regulate the Nudel-paxillin interaction as well . If the Nudel-paxillin interaction was indeed critical for nascent membrane adhesion , according to the above results ( Figure 5A and 5B ) , overexpressing FAK or any FAK mutant that binds paxillin should displace Nudel prematurely from nascent adhesion sites and consequently impair cell spreading . Indeed , overexpression of any FAT-containing deletion mutant , i . e . , FAKΔFERM , FRNK , or even FAKFAT , resulted in high incidences ( ≥72% ) of cell shrinkage: affected cells usually lacked FXs and were typically polygonal in shape , with cell edges supported by F-actin bundles and FAs ( Figures 6A , 6B , and S6C ) . In contrast , mutants lacking the FAT domain , e . g . , FAKΔFAT and FAKKin , or containing FAT but lacking paxillin-binding activity , e . g . , FAKΔFMPX1 and FAKΔFMPX2 , only generated background levels of shrunken cells ( Figures 6A , 6B , and S6C ) . Moreover , although FAKΔFAT and FAKKin failed to show FA localization , FAKΔFMPX1 was still efficiently targeted to FAs ( Figures 6A and S6C ) [51] . FAKΔFMPX2 exhibited weak , but clear , FA localization as well ( Figure S6C ) . Because FAKΔFMPX1 and FAKΔFMPX2 do not bind to paxillin , their localization to FA is probably mediated by talin [2] , [51] , [52] , [54] . Thus , the paxillin-binding activity of FAK is essential for the induction of cell edge shrinkage , whereas its kinase domain is dispensable . In contrast to the intact FAT-containing mutants , full-length FAK only showed a mild effect . Although GFP-FAK–positive cells with the shrinkage phenotypes were approximately 2-fold higher in percentages than surrounding untransfected cells , the majority of cells overexpressing GFP-FAK ( 66 . 4% in average ) showed normal lamellipodia ( Figure 6A and 6B ) . The kinase-dead mutant GFP-FAKKD had a similar effect ( Figure 6A and 6B ) , whereas FAKOpn , a mutant containing two point mutations ( Y180A/M183A ) that abrogate the autoinhibitory effect of the FERM domain [55] potently induced cell edge shrinkage upon its overexpression ( Figure 6A and 6B ) . Similar phenotypes were seen in ECV304 cells grown on fibronectin- and/or laminin-coated substratum as well as in CV1 and NIH3T3 cells ( Figure S7; unpublished data ) . Therefore , the “open” structure of full-length FAK is important for both full activation of FAK [55] and induction of cell edge shrinkage . FAK is believed to promote cell migration through its kinase activity [2] . To understand why both FAKOpn , which exhibits robust kinase activity in cells ( unpublished data ) [55] , and FRNK , which is a dominant-negative mutant on kinase activity of endogenous FAK [50] , caused similar cell shrinkage phenotypes ( Figure 6A and 6B ) , we monitored behaviors of live cells . Consistent with results in fixed cells ( Figure 6A and 6B ) , ECV304 cells overexpressing GFP-FAKOpn were narrow or polygonal in shape ( Figure 6C; Video S5 ) . Whereas surrounding untransfected cells migrated through typical arc-like lamellipodia , these transfectants extended long processes rich in transient filopodium-like projections ( n = 19/20 ) and migrated like fibroblasts ( Figure 6C and 6D; Video S5 ) [28] . In contrast , cells overexpressing GFP-FRNK showed markedly reduced motilities ( Figures 6C and S6D; Video S6 ) . Such cells also failed to show active membrane protrusions ( n = 21/21 ) ( Figure 6D ) , consistent with the lack of FAK kinase activity . Therefore , although cells overexpressing FAKOpn or FRNK showed different motilities , they share similar shrinkage phenotypes . These results identify a novel kinase-independent role of FAK in cell spreading . As this role of FAK depends on its interaction with paxillin , the “shrunken” phenotype caused by FAK overexpression is attributed to poor adhesions of nascent membrane protrusions due to premature disruption of the Nudel-paxillin interaction at the leading edge .
We have previously shown that Nudel is required for membrane protrusions in NIH3T3 cells [28] . Here , we further showed that Nudel depletion markedly repressed lamellipodial formation in ECV304 cells ( Figure 1A; Video S2 ) , indicating a general requirement of Nudel in membrane protrusion . Lamellipodial formation requires Rac activity , whereas Cdc42 can activate Rac [3] , [32] , [34] . Therefore , the protrusion defect upon Nudel depletion is consistent with inhibition of Cdc42 activity [28] . By contrast , although Nudel is also essential for dynein functions [22] , [23] , dynein activity is not important for lamellipodial formation and free migration of ECV304 cells ( Figure S2 ) . Nudel is also critical for stabilization of membrane protrusions by facilitating nascent adhesion formation . First , Nudel depletion by RNAi primarily resulted in cell edge collapse ( Figures 2A and S3 ) . As the robust stress fibers in Nudel RNAi cells were not sensitive to blebbistatin treatment ( Figure 2B ) , their formation is unlikely due to increased contractile forces on stress fibers , e . g . , through activation of Rho GTPase [34] , [40] . Rather , it is attributed to mechanical strains caused by cell edge shrinkage because stress fibers induced by mechanical forces do not depend on myosin II activity [18] , [40] , Second , although overexpression of Rac1CA rescued the membrane protrusion defect of Nudel depletion , cells still failed to fully spread due to poor adhesions of their protruded membranes ( Figure 1B and 1E; Video S4 ) . Importantly , this phenotype is not caused by the inactivation of Cdc42 per se , as coexpression of a dominant-negative form of Cdc42 with Rac1CA failed to repress cell spreading ( Figure 1B and 1C ) . Third , the markedly reduced formation of FXs as well as nascent adhesions upon Nudel depletion even in the presence of Rac1CA ( Figure 2C and 2E ) further indicates a positive role of Nudel in nascent adhesions . The Nudel-paxillin interaction further substantiated the role of Nudel in nascent adhesion because paxillin can bind directly to integrin and is thus one of the earliest intracellular proteins at nascent adhesions [2] , [5] , [6] . As the interaction between exogenous Nudel and paxillin was readily detected even by Ponceau S staining after co-IP ( Figure 3A ) , these two proteins appear to interact with high affinity . Moreover , they interacted directly through their C-terminal domains ( Figures 3 and S4 ) . The interaction between endogenous paxillin and FLAG-Nudel expressed at a relative low level can also be detected in vivo , especially upon knockdown of FAK expression ( Figure 5D ) . Our results suggest that Nudel interacts with paxillin in nascent adhesions . Complex formation of Nudel with both paxillin and vinculin ( Figures 5B , lane 9 , and 5D ) suggests its localization in certain FCs . Nevertheless , it was not detected in FXs or FAs , but enriched and colocalized with paxillin at the leading edge in areas of active membrane protrusions ( Figure 3E–3H ) , where nascent adhesions occur [4] , [5] . Moreover , a localization of Nudel in nascent adhesions is also consistent with its functions there ( Figures 1 and 2 ) . We provided evidence showing that the presence of Nudel in FCs can indeed stabilize integrin-ECM interactions using PGN ( Pax-GFP-Nudel ) ( Figures 4 and S5 ) . PGN was specifically located in FCs ( Figure S5B ) and reduced FA motilities by approximately 3-fold , compared to Pax-GFP ( Figure 4A–4C ) , suggesting elevated stability , or strength , of individual adhesion site . Furthermore , the elevated adhesiveness of PGN-positive cells over Pax-GFP-positive ones , measured through their resistance to shear forces ( Figure 4D ) [49] , further supports the increase in adhesion strengths . Although PGN is an artificial protein and may not precisely reflect situations in vivo , its effects on FA motility and cell adhesion are well in agreement with other results that suggest a role of Nudel in stabilization of nascent adhesions ( Figures 1–3 ) . Consistently , paxillin-deficient cells exhibit a delayed rate of spreading [56] . Nevertheless , it is currently not known whether the Nudel-paxillin interaction stabilizes integrin-ECM ligations by modulating the integrin conformation or by regulating other intracellular adhesion molecules . We demonstrated that FAK is a key regulator of the Nudel-paxillin interaction . FAK was able to disrupt the interaction via direct binding to paxillin ( Figure 5 ) . Such a competition effect may be mediated through steric hindrance . Alternatively , given that the FAT domain alone , which covers only one-eighth of FAK , was already sufficient to disrupt the Nudel-paxillin interaction ( Figure 5A and 5B ) , FAK binding may induce in paxillin a conformational change that abrogates Nudel binding . That FAK and Nudel bound to distinct regions of paxillin ( Figures 3D and 5C ) [15] also supports the latter speculation . In addition to its known kinase-dependent functions in cell migration [2] , we found that FAK can negatively regulate nascent adhesions . Overexpression of FAK resulted in an approximately 2-fold increase in incidence of cells with shrunken edges comparing to surrounding untransfected populations ( Figure 6A and 6B ) . Deleting the FERM domain ( e . g . , FAKΔFERM ) or abolishing its autoinhibitory role through point mutations ( e . g . , FAKOpn ) [55] considerably augmented incidences of the shrunken phenotype ( Figure 6A and 6B ) . Such a phenotype , however , is not correlated with the kinase activity of FAK because it is similar in cells overexpressing either the hyperactive ( e . g . , FAKOpn and FAKΔFERM ) [55] or the dominant-negative ( e . g . , FRNK and possibly FAKFAT ) [50] mutants ( Figures 6A , 6B , and S6C ) . In addition , in the absence of the FERM domain , the potency of FAK to induce cell edge collapse is only correlated with its interaction with paxillin ( Figures 6A , 6B , and S6C ) . Localization of FAK in FCs , however , is not sufficient: the point mutants FAKΔFMPX1 and FAKΔFMPX2 localized to FAs but failed to cause the shrunken phenotype ( Figures 6A and S6C ) . We therefore propose a model to explain how paxillin , Nudel , and FAK cooperate to modulate integrin-mediated adhesivity in cell migration ( Figure 7 ) : during membrane protrusion , activated integrin molecules located on polymerizing F-actin [4] bind to ECM to form nascent adhesion sites containing paxillin [6]; association of Nudel with paxillin strengthens such sites; upon formation of the open conformation in response to external signals , possibly through interaction of integrin and/or growth factor receptors [55] , FAK displaces Nudel from paxillin; adhesion sites now exhibit a lower strength than those containing Nudel . The antagonizing roles of Nudel and FAK in adhesivity provide a mechanism for cells to properly coordinate adhesion and migration . The positive effect of Nudel on adhesion strength can stabilize nascent adhesion sites and thus facilitate stabilization of membrane protrusions at the leading edge . Stronger adhesiveness would also allow nascent sites to transmit stronger traction forces [13] and to resist retraction . On the other hand , because FAs are large in size , a decreased strength of their individual FC sites would facilitate FA movement ( Figure 4 ) [19] and retraction of the trailing side . Our findings also help to understand how cells orchestrate different events in migration . As formation of the open structure of FAK depends on upstream signals and serves as a prerequisite for activation of the kinase [55] , disruption of the Nudel-paxillin interaction , thus down-regulation of adhesivity at nascent adhesions , is likely to precede other events associated with the kinase activity of FAK [2] . Such an ordered sequence of action appears important for cell migration because premature disruption of the Nudel-paxillin interaction and/or interference with the kinase activity of FAK affects cell motility . For instances , although excess wild-type FAK failed to interfere with lamellipodial formation in the majority of cells ( Figure 6A and 6B ) , overexpression of FAKOpn to prematurely disrupt the Nudel-paxillin interaction ( Figures 5 , 6 , and S6 ) while provoking a hyperactive kinase activity [55] , impaired the arc-like lamellipodium formation in ECV304 cells and resulted in cell migration through transient filopodium-like membrane projections ( Figure 6; Video S5 ) . In contrast , overexpression of FRNK to similarly abrogate the Nudel-paxillin interaction ( Figure S6 ) while also inhibiting endogenous FAK activity [50] , [57] caused cell shrinkage but poor migration ( Figures 6 and S6; Video S6 ) [57] . Furthermore , FAK-null cells have been shown to exhibit robust FC formation at the cell periphery [58] , [59] , reminiscent of enhanced cell edge adhesions to the substratum . These cells also show poor migration [58] , [59] . We have previously shown that Nudel can stabilize active Cdc42 at the leading edge by sequestering Cdc42GAP in NIH3T3 cells [28] . Nudel also contributes to dynein functions at the leading edge [28] , [60] . Moreover , similar to paxillin ( Figure 3 ) , both Cdc42GAP and dynein heavy chain bind to the C-terminus of Nudel [22] , [28] . How these functions of Nudel are coordinated is not yet clear . One possibility is that Nudel interacts with different partners for different functions or in different cell types . Another possibility is that these partners use Nudel as a common platform to achieve orchestration of different functions . Interestingly , in co-IP experiments , we found that associations of Cdc42GAP , paxillin , and dynein with Nudel were significantly enhanced upon overexpression of both paxillin and Cdc42GAP ( Figure S8 ) . If such a synergetic effect on Nudel binding occurred at the leading edge due to enrichment of these proteins there ( Figure 3 ) [28] , [60] , the Nudel-paxillin interaction and the regional activation of Cdc42 and dynein would become spatiotemporally coupled events to eventually facilitate establishment of a polarized lamellipodium . These issues will be worthy of future investigations .
Expression plasmids for human Nudel , its mutants , and p50dynamitin were described previously [22] , [28] , [61] . pLV-IRES-FLAG-GFP and pLV-IRES-FLAG-GFP-Nudel were constructed from a lentiviral vector ( a gift from Qiwei Zhai , Institute of Nutritional Science , Shanghai Institutes for Biological Sciences [SIBS] ) for low-level expression of FLAG fusion proteins via the internal ribosome entry site ( IRES ) . pTER-Nudi , a Nudel RNAi construct , and a control construct pTER-Luci [31] were further modified to coexpress GFP or RFP . The RNAi-resistant Nudel constructs contained three silent mutations in the short hairpin RNA ( shRNA ) -target region . The expressed proteins , despite unchanged amino acid sequences , were named Nudel-R and NudelC36-R sheer for presentation purposes . To silence FAK expression , pTER-FAKi1 and pTER-FAKi2 were constructed and cotransfected at a 1∶1 ratio . Their targeting sequences are 5′GGTACTGGTATGGAACGTTCT3′ and 5′GCCTTAACAATGCGTCAGT3′ , respectively . Expression plasmid for GFP-vinculin was kindly provided by Benjamin Geiger ( Weizmann Institute , Israel ) . pGFP-hPaxillin and pVSV-mFAK/FRNK were gifts from Kenneth M . Yamada ( National Institute of Dental and Craniofacial Research [NIDCR] , National Institutes of Health [NIH] ) . To express fusion proteins paxillin-GFP or Paxillin-GFP-Nudel , the coding sequence of paxillin was amplified by PCR and inserted in-frame between the NheI and AgeI sites of pEGFP-C1 or pEGFP-C1-Nudel . FAK and paxillin mutants were created by PCR as well . Plasmids for expression of GFP-tagged Rac1 , FLAG-tagged Cdc42 , and mutants were from Xiaobing Yuan ( Institute of Neuroscience , SIBS ) and Michiyuki Matsuda ( Osaka University , Japan ) . Plasmids containing PCR fragments were subjected to sequencing confirmation . Mouse monoclonal antibodies ( mAbs ) to α-tubulin , vinculin , FLAG , and phospho-Tyr , and rabbit antibodies to FAK and FLAG were purchased from Sigma-Aldrich . mAbs to paxillin and phospho-Ser/Thr were from BD Biosciences Transduction Laboratories . Rabbit antibody to GFP was from Santa Cruz Biotechnology . Anti-GST mAb was from Wolwo Biotech . Anti-Nudel IgY was generated from chicken and affinity-purified [23] . Secondary antibodies conjugated with peroxidase or Alexa Fluor-405 , -488 , -546 , or -647 were purchased from Invitrogen . Phalloidin-Alexa-647 was from Invitrogen . Phalloidin-TRITC and blebbistatin were from Sigma-Aldrich . All cells were cultured at 37°C and 5% CO2 in Dulbecco's modified Eagle's medium ( Invitrogen ) supplemented with 10% ( v/v ) bovine serum ( Sijiqing ) . Human embryonic kidney ( HEK ) 293T cells were transfected by using the conventional calcium phosphate method . This cell line was used for assays involving immunoprecipitation due to its high transfection efficiency . Human bladder epithelial ECV304 [62] and cervical carcinoma HeLa cells were transfected with Lipofectamine2000 ( Invitrogen ) . In overexpression experiments , cells were harvested approximately 48 h posttransfection for biochemical assays or fixed approximately 20 h posttransfection for microscopy . In RNAi experiments , they were used 48–72 h posttransfection . To determine RNAi efficiency in ECV304 , GFP-positive transfectants were enriched to approximately 90% by using a BD FACSAria cell sorter 48 h after transfection . To prevent cell aggregation , Latrunculin A ( 0 . 1 µg/ml; Invitrogen ) was added prior to sorting . Transfectants were cultured for an additional 24 h and then collected for immunoblotting ( IB ) . Approximately 1×107 HEK293T cells were lysed in co-IP buffer ( 20 mM Tris-HCl [pH 7 . 5] , 100 mM KCl , 0 . 1% NP-40 , 1 mM EDTA , 10% glycerol , 1 mM DTT , 50 mM NaF , 10 mM Na-pyrophosphate , 1 mM Na-Vanadate , and protease inhibitors cocktail [Calbiochem] ) by repetitive pipetting through a 1-ml tip . After centrifugation at 10 , 000 g for 10 min to remove debris , lysates were incubated with anti-FLAG M2 agarose beads ( Sigma ) for 2 h on a rotator at 4 °C . The beads were then washed with the buffer for three times , followed by elution with synthetic FLAG peptide [63] . For pull-down assays , bacterial lysates containing GST fusion proteins or FLAG-Nudel were premixed for 1 h and then incubated with glutathione or anti-FLAG agarose beads ( Sigma-Aldrich ) for another 1 h at 4°C with agitation . Proteins binding to the beads were then boiled in SDS-sample buffer and subjected to IB . When necessary , membranes were stripped and blotted with different antibodies . Experiments were repeated at least three times . Unless indicated , cells were grown sparsely on sterile glass coverslips without pre-coating of ECM . They were fixed with 4% paraformaldehyde ( Sigma-Aldrich ) for 15 min , followed by permeabilization with 0 . 5% Triton X-100 ( v/v ) for 10 min . For scratch wound assays , confluent cell monolayers cultured in serum-free medium for 12 h were scratched with yellow tips [28] and then cultured in serum-containing medium for an additional 3 h prior to fixation . Immunofluorescence staining was performed with appropriate combinations of antibodies . F-actin was decorated with fluorochrome-labeled phalloidin . Images were captured with a Leica TCS SP2 laser-scanning confocal microscope . Grayscale images were converted to pseudocolor using Adobe Photoshop . Statistical data were presented as mean±standard deviation ( SD ) from at least three experiments . Cell area and circularity ( 4π×area/perimeter2 ) were measured using ImageJ ( NIH ) . To quantify fluorescent colocalizations along the leading edge , intensity profiles were obtained using ImageJ . Cross-correlations and Pearson correlation coefficients of the intensity profiles were calculated with Matlab ( MathWorks ) [4] . ECV304 cells were cultured in L-15 medium ( Invitrogen ) supplemented with 10% ( v/v ) bovine serum . Image sequences for cell migration were collected by using an Olympus IX81 microscope with 37 °C-incubation chamber , motorized stage , and Evolution QEi CCD camera ( Media Cybernetics ) , or a Leica AS MDW workstation with a heating hood and a CoolSNAP HQ CCD camera ( Roper Scientific ) [28] , [64] . For FA motility assays , cells were imaged by using an Olympus FluoView 1000 inverted confocal microscope with a heating stage at 5-min intervals . ImageJ ( NIH ) was used for measurement . Migration tracks were determined as tracks of nuclei [28] . Average velocity of a sparse cell was calculated using its track length of free migration . Flow chamber assays were performed basically as described [49] . A polystyrene Petri dish coated with purified human laminin and fibronectin ( 12 . 5 µg/ml each; Sigma-Aldrich ) was used as the lower wall of the chamber . HEK293T transfectants were trypsinized and sorted . GFP-positive cells were diluted to 1×106/ml in complete culture medium and infused into the flow chamber immediately . Cells were allowed to accumulate for 30 s at 0 . 3 dyne/cm2 and for 10 s at 0 . 4 dyne/cm2 . Shear stress was then increased every 10 s from 1 dyne/cm2 up to 32 dyne/cm2 in 2-fold increments . The number of cells remaining bound at the end of each 10-s interval was counted . | Cell migration is an essential process in both single-cell and multicellular organisms . In higher animals , cell migration is important for many biological processes , including embryonic development , the immune response , and wound healing . Cancer cell invasion into healthy tissues occurs as a result of inappropriate cell migration . As can be easily visualized when cultured in the lab , mammalian cells attach to surfaces through focal adhesions , cellular structures characterized by complexes of the transmembrane protein integrin and intracellular proteins including paxillin and focal adhesion kinase ( FAK ) . In order for cells to move , they must coordinate two processes: extension of the front edge of the cell and retraction of the back edge . To accomplish this , a cell first protrudes membranous structures from the front edge and then establishes adhesion structures known as nascent adhesions to hold the extensions in place . At the same time , the focal adhesions that hold a cell in place must be disrupted in order for the back edge of the cell to retract . Here , we show that a protein called Nudel is enriched at the front edge of moving cells , where it interacts with paxillin but is not detected in focal adhesions . We further show that the focal adhesion protein FAK is able to abolish the Nudel-paxillin interaction , leading to repression of the formation of nascent adhesions and to the loss of cell extensions . We therefore propose a model in which modulation of paxillin interactions in nascent adhesions and in focal adhesions is critical for coordinated cell movement: the Nudel-paxillin interaction enhances the strength of nascent adhesions to promote the attachment of membrane protrusions at the front edge of the cell , whereas FAK prevents the Nudel-paxillin interaction in focal adhesions in order to facilitate retraction of the back edge of the cell . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"cell",
"biology",
"cell",
"biology/cell",
"adhesion",
"cell",
"biology/cytoskeleton"
] | 2009 | Nudel and FAK as Antagonizing Strength Modulators of Nascent Adhesions through Paxillin |
One of the most influential observations in molecular evolution has been a strong association between local recombination rate and nucleotide polymorphisms across the genome . This is interpreted as evidence for ubiquitous natural selection . The alternative explanation , that recombination is mutagenic , has been rejected by the absence of a similar association between local recombination rate and nucleotide divergence between species . However , many recent studies show that recombination rates are often very different even in closely related species , questioning whether an association between recombination rate and divergence between species has been tested satisfactorily . To circumvent this problem , we directly surveyed recombination across approximately 43% of the D . pseudoobscura physical genome in two separate recombination maps and 31% of the D . miranda physical genome , and we identified both global and local differences in recombination rate between these two closely related species . Using only regions with conserved recombination rates between and within species and accounting for multiple covariates , our data support the conclusion that recombination is positively related to diversity because recombination modulates Hill–Robertson effects in the genome and not because recombination is predominately mutagenic . Finally , we find evidence for dips in diversity around nonsynonymous substitutions . We infer that at least some of this reduction in diversity resulted from selective sweeps and examine these dips in the context of recombination rate .
Homologous meiotic recombination has an important role in molecular evolution . Sufficient recombination uncouples the evolution of different sites on the same chromosome allowing positive or negative selection at one site to act independently from selection at another site . If there is less than effectively free recombination between two selected sites , then linkage results in selection at one site interfering with selection at another site . This has been termed “Hill–Roberson interference” [1]–[6] . Hill–Robertson interference increases the probability of fixation of deleterious mutations , decreases the probability of fixation of advantageous mutations , and reduces overall DNA sequence diversity . Thus , the breakdown of linkage disequilibrium between loci experiencing Hill–Robertson interference allows selection to act more efficiently , purging deleterious mutations and accelerating adaptation [1]–[6] . Such indirect effects of recombination on the genome [7] result in a positive association between the rate of recombination and adaptive evolution [8]–[10] . For example , recombination rate is positively associated with codon usage bias , whereby those codons coded by the most abundant tRNAs are “preferred” and used more often [11] , [12] . Recombination has direct effects on a genome sequence as well , because recombination influences base composition through biased gene conversion and the distribution of repetitive elements , hotspot sequences , and indels [7] , [13]–[17] . Understanding the magnitude of indirect effects in light of these direct effects has proved challenging [7] . One striking association is a positive relationship of local recombination rate and nucleotide diversity [13] , [18] , [19] . Originally described in Drosophila melanogaster [13] , the positive relationship between recombination rate and nucleotide diversity has been demonstrated in a wide range of taxa , including humans , mice , yeast , maize , and tomatoes ( reviewed in [20] ) . It is not fully understood how much of this relationship results from recombination's indirect versus direct effects on the genome . For instance , mutations created during crossing over or double-strand break repair may generate new polymorphisms and hence increase diversity [21]–[27] . Alternatively , recombination may indirectly influence genetic diversity by mitigating the genomic footprint of selective sweeps and background selection [28]–[30] . One way to distinguish between these general explanations is to evaluate the relationship of between-species nucleotide divergence at neutral sites and local recombination rate , because truly neutral mutations are substituted at the same average rate between species as they appear between generations , even if linked to sites under selection [31] , [32] . This allows us to predict that both within-species nucleotide diversity and between-species nucleotide divergence would have a positive relationship with local recombination rate [13] , if the recombination–diversity association was purely caused by mutation . In contrast , selective sweeps and background selection will cause an association between recombination and within-species nucleotide diversity , but not a relationship between recombination and between-species nucleotide divergence [30] , [32] . The absence of an association of between-species nucleotide divergence and local recombination rate suggests that variation in recombination rate translates to variation in the efficiency of selection [13] . Past work relating nucleotide divergence to recombination rate found no relationship between these two variables in several species of Drosophila , mouse , beet , yeast , and other species [13] , [20] , [33]–[37] . Furthermore , in several species , evidence indicates that segregating ancestral polymorphisms may be responsible for correlations between divergence and recombination rate ( [38]–[40] , also suggested by [25] , [41] ) . The test above , however , implicitly assumes that local recombination rates are conserved between the two species used to generate the nucleotide divergence measure . If recombination rate has diverged between the two species , no relationship between local recombination rate and nucleotide divergence may be detected even when recombination is mutagenic ( see Figure S1 ) . Recombination rates , especially at fine scales , are often not conserved among closely related species , as is the case between humans and chimpanzees [42]–[44]; thus , the assumption of conservation of recombination rates may be violated in previous studies , and a more definitive understanding of the diversity–recombination association awaits estimates that are free from this assumption . Though there are theoretical expectations concerning how recombination rate should affect selection efficiency [45] , [46] , it is unclear empirically whether variation in local recombination rates translates into significant variation in the efficiency of selection [7] . Several empirical studies have tackled this problem [12] , [38] , [47]–[52] , and many findings suggest that recombination rate influences the efficiency of positive or negative selection in regions of moderate or high recombination . Still , various confounding factors ( e . g . , biased gene conversion , gene density ) may produce spurious correlations between both recombination and substitution rate , and some authors suggest that there is no strong empirical evidence for recombination affecting the efficiency of selection ( apart from reduced selection in regions with essentially no recombination [7] ) . The Drosophila pseudoobscura system is ideal for pursuing questions about recombination rate variation and its molecular evolutionary consequences . The average crossover rate of D . pseudoobscura ( about 7 cM/Mb in females ) is over twice that of D . melanogaster [53] . There is also considerable fine-scale ( <200 kb windows ) variation in the local recombination rate within the genome of D . pseudoobscura and within the genome of its sister species , D . persimilis [25] , [33] , [54] . While some recombination data are available for D . pseudoobscura and D . persimilis , these sister taxa interbreed in the wild [55]–[57] and are , therefore , not ideal for examining the divergence–recombination association . For example , shared polymorphism due to hybridization and recent speciation may be responsible for the positive divergence–recombination association found in a previous study [25] ( see also [38] , [39] ) . Fortunately , a third species exists ( D . miranda ) that is phylogenetically close to D . pseudoobscura but does not interbreed with D . pseudoobscura . Since there is still some residual shared ancestral polymorphism [58] , we also obtained the genome sequence for a slightly more distantly related outgroup species , D . lowei ( Figure S2 ) . Sequence from D . lowei is useful for generating a proxy for neutral mutation rate across the genome . In this work , we generate and compare two fine-scale recombination maps for D . pseudoobscura , which each cover approximately 43% of the D . pseudoobscura physical genome and one fine-scale recombination map that covers approximately 31% of the D . miranda physical genome . In order to circumvent the assumption of classic studies , we analyze the relationship of local recombination rate to nucleotide diversity and divergence in regions with very similar recombination rates between the two species . By employing a linear model framework to account for multiple covariates , we conclude that the contribution of recombination to diversity is significant and positive , but recombination contributes little to divergence . This indicates that recombination is likely to modulate the footprint of selection in the genome . Next , we tested the impact of recombination rate on the efficiency of selection . We examined whether recombination rate ( 1 ) affects the distribution of nonsynonymous substitutions across the genome and ( 2 ) affects the pattern of diversity around nonsynonymous and synonymous substitutions . In particular , we use a generalized linear model to test how recombination modulates the magnitude and physical extent of the loss of diversity surrounding substitutions . Our analysis of these putative selective sweeps should be less sensitive to common confounding factors such as gene expression and GC content than previous measures . In total , this work allowed us to determine that recombination rate has an important impact on how selection shapes diversity across the genome of Drosophila pseudoobscura and its close relatives .
We generated linkage maps for chromosome 2 and parts of the X chromosome for D . pseudoobscura and D . miranda . Using a backcross design and inbred lines , we developed two replicate recombination maps ( referred to here as “Flagstaff” and “Pikes Peak” ) for D . pseudoobscura and one recombination map for D . miranda using the Illumina BeadArray platform to distinguish heterozygotes from homozygotes of the inbred lines used in the backcross design . These maps ( Table S1 ) measure recombination rate across <200 kb windows , and we refer to these as “fine-scale” maps . Recombination was surveyed across approximately 43% of the D . pseudoobscura physical genome and about 31% of the D . miranda physical genome ( Tables S1 and S2 ) . For each of the three maps , nearly the entire assembled region of chromosome 2 ( 97 . 8%–99 . 4% ) , the majority of the XR chromosome arm ( 70 . 8%–89 . 4% ) , and part of the XL chromosome arm ( ∼22%–23% ) were surveyed ( Table S2 ) . After removal of likely erroneous putative double recombinants , ambiguous genotypes , and markers that did not work or gave inconsistent genotypes , recombination was measured for three different crosses for 1 , 158–1 , 404 individuals per map ( Table S1 ) . Excluding larger intervals at the telomeres and centromeres , intervals between markers had a median size across the three maps of 141–148 kb for chromosome 2 and 146–160 kb for the XR chromosome arm ( Table S1 ) . For chromosome 2 , recombination rates ranged from 0–30 . 8 cM/Mb in D . pseudoobscura and 0–24 . 0 cM/Mb in D . miranda ( Table S2 ) . The number of individuals surveyed is often slightly different per interval; therefore , for all intervals where no recombination was detected , we report 0 cM/Mb . The recombination rate for those intervals with “0 cM” should be interpreted as <1 recombination event per total number of individuals surveyed for each interval ( Dataset S1 ) . Recombination near the telomere and centromere was measured at a broader scale than the remainder of chromosome 2 because we expected these regions to have lower crossover rates than the center of the chromosome ( chromosome 2 is telocentric ) . Because of this limitation , comparisons of recombination rates between the ends of the chromosome and the center are more tentative . Nonetheless , examining recombination across roughly 3 Mb of sequence at the telomeric end and 3 Mb at the centromeric end , we found up to an 8 . 9-fold difference between the recombination rates at the middle of chromosome 2 relative to the centromeric end . The Pikes Peak D . pseudoobscura map exhibited the largest reduction of recombination at the telomeric or centromeric ends relative to the center of the chromosome for all three maps , though in the Flagstaff D . pseudoobscura map and the D . miranda map , recombination rates were reduced by at least 2 . 6-fold in the centromere and telomere relative to the center of the chromosome ( Table S3 ) . For the XR chromosome arm , recombination rates ranged from 0–25 . 2 cM/Mb in D . pseudoobscura and 0–32 . 3 cM/Mb in D . miranda ( Figure S3 presented with 95% confidence intervals; see also Dataset S1 , Table S2 ) . The number of crossovers per individual for both chromosome 2 and the XR arm was close to 1 ( 1 . 01–1 . 06 ) for D . pseudoobscura and was 1 . 40–1 . 54 for D . miranda , illustrating that a greater overall recombination rate in D . miranda relative to D . pseudoobscura is observed in both an autosome and a sex chromosome . The XL chromosome arm was not surveyed as intensively ( ∼22%–23% of the XL arm in Pikes Peak and D . miranda and ∼60% of the XL arm in Flagstaff; Figure S4 presented with 95% confidence intervals; Dataset S1 ) . The number of crossovers per individual appears consistent with ∼1 crossover per chromosome arm , as in D . pseudoobscura XR and chromosome 2 , but the average number of crossovers per individual on the XL reflects how much of the arm was surveyed . For example , when ∼22%–23% of the arm was surveyed , crossovers per individual ranged from 0 . 23–0 . 26 ( Table S2 ) . A binomial Generalized Linear Model ( GLM ) with size of the interval as a covariate and interval identity as a factor in the model indicated significant heterogeneity in recombination rate among intervals for chromosome 2 , XR , and XL ( each tested separately ) for each of the three maps ( each tested separately , interval identity p<0 . 00001 , χ2≥64 . 67 , df≥3 , in all cases ) . Furthermore , 95% confidence intervals ( generated via the same method in [54] ) do not overlap in many cases between different intervals ( shown in Figures 1 , S3 , S4; Dataset S1 ) . Overall , we observe heterogeneity in fine-scale recombination rates within each of the three maps ( see Figures 1 , S3 , and S4 with 95% confidence intervals plotted; Dataset S1; statistical quantification between maps given in section below ) , and we note a reduction in recombination rate around the telomeric and centromeric ends consistent with other studies in Drosophila [33] . Our three fine-scale crossover maps utilized markers on average 141–160 kb apart ( median interval size for each of the three maps , with the exception of XL where the median distance between markers was 200–1 , 775 kb for the three crosses ) . We additionally examined three regions on chromosome 2 in more detail . Each of these regions spanned a total of 99–125 kb , and we placed markers every ∼20 kb within the region ( 16 total intervals; Tables S4 and S5 ) . These regions were originally picked because previous data [25] , [33] indicated that recombination rates for each of these regions differed ( regions are referred to as 6 Mb , 17 Mb , and 21 Mb , which indicate approximate location on chromosome 2 ) . We refer to these as “ultrafine-scale” maps . For these ultrafine maps , we followed the same backcross scheme as above , and we scored approximately 10 , 000 individuals for each marker ( Table S5 ) . For the 16 ultrafine intervals ( Tables S4 and S5 ) , each interval was on average 20 . 61 kb long ( range 12 . 6–27 . 4 kb ) . Recombination rates range from 1 . 6–21 . 2 cM/Mb for these ∼20 kb intervals ( Figure 2; see Table S5 for 95% CI ) . The ultrafine-scale map uncovered variation in recombination rates that was not apparent with the fine-scale maps . For example , for the 17 Mb ultrafine-scale region on chromosome 2 , the recombination rates for the two fine-scale intervals spanning this region ( 17 . 5–17 . 7 Mb ) are 5 . 6 and 4 . 4 cM/Mb . The ultrafine-scale recombination rates , in contrast , ranged from 3 . 5–21 . 2 cM/Mb ( markers spanning 17 . 5–17 . 7 Mb ) . This heterogeneity in recombination rates within the ultrafine regions was statistically significant ( binomial GLM similar to that described in fine-scale section above: p = 0 . 0011 , df = 14 , χ2 = 35 . 91; 95% confidence intervals given in Table S5 ) and highlights the fact that “broader” scale measures of recombination rates ( such as the fine-scale measures here ) are averages of true variation in recombination rate . For comparisons of recombination rates between fine-scale maps , we restricted our analysis to intervals that were condensed to have nearly identical physical marker placement between the three fine-scale maps ( Figures S5 and S6; Table S6 ) . Recombination was estimated as detailed above , using the number of crossovers spanning the newly defined physical intervals . After condensing across all three maps , 97 intervals remained for chromosome 2 and 44 intervals for XR ( see Tables 1 and S6 fornumber of individuals , size , range of these condensed intervals , and base pairs between markers on each map ) . The XL chromosome arm was not included in the analysis that used condensed intervals across maps because too few intervals overlapped between all three maps . When comparing two maps , intervals were condensed between those two maps only ( see Datasets S2 and S3 for rare events logistic regressions for all two-map and three-map comparisons ) . Recombination rates did not differ significantly between the two D . pseudoobscura maps for either the XR or chromosome 2 for the two-map comparisons ( each chromosome analyzed separately , rare events logistic regression , absolute value of z>0 . 3901 , p>0 . 236 , in both cases; Dataset S2 ) . For chromosome 2 , one interval was significantly different in recombination rate after correcting for multiple tests [59] . For the XR , no intervals between the two D . pseudoobscura maps were significantly different in recombination rate after correcting for multiple tests . The 95% confidence intervals for the odds ratio of the difference between maps were narrow and located around zero , indicating that the maps are likely very similar ( chromosome 2 , 0 . 87–1 . 10; XR , 0 . 94 , 1 . 28; within-species two map comparison ) . It is unlikely that the single significant difference observed within the same species is because of slight differences in marker placement between the two maps . The marker placement for this interval was nearly identical between the two maps ( left marker , 102 nucleotides different between maps; right marker , 17 nucleotides ) . For both chromosome 2 and the XR chromosome arm , Drosophila miranda had significantly higher recombination rates than both D . pseudoobscura maps ( Figure S5 , Table 1 , Datasets S2 and S3 ) . A rare events logistic regression of two-map comparisons indicated that the recombination rate of the D . pseudoobscura crosses we surveyed is about 76%–78% of the D . miranda recombination rate we observed on chromosome 2 ( absolute z value>4 . 5374 , p<0 . 001 for D . miranda relative to either D . pseudoobscura map , Table 1 ) . The recombination rate of D . pseudoobscura is about 68%–71% of the D . miranda recombination rate on the XR chromosome arm ( rare events logistic regression absolute z value>5 . 101 , p<0 . 001 for D . miranda relative to either D . pseudoobscura map , Table 1 ) . After the global difference between D . miranda and D . pseudoobscura is accounted for by the rare events logistic regression , recombination rates within and between species appear very similar for chromosome 2 ( Figure S5; Datasets S2 and S3 ) . None of the intervals for the two-map comparison between D . miranda and D . pseudoobscura–Flagstaff were significantly different after correction for multiple tests , though power to detect significant differences on a per interval basis was likely weak ( see confidence intervals in Datasets S2 and S3 ) . For example , 15 of the 115 intervals on chromosome 2 showed at least a 3-fold difference in recombination rate between maps ( Datasets S2 and S3 ) , though this magnitude of difference was not significant in our rare events logistic regression after correcting for multiple tests . Likewise , only one of the intervals for the two-map comparison between D . miranda and D . pseudoobscura–Pikes Peak was significantly different after correction for multiple tests , but 19 of the 123 intervals exhibited at least a 3-fold difference in recombination rate between maps for chromosome 2 . The XR chromosome exhibited a qualitatively larger difference in recombination rate between species than chromosome 2 . After the global difference between D . miranda and D . pseudoobscura is accounted for by a rare events logistic regression , two of the intervals between D . miranda and D . pseudoobscura–Flagstaff for the two-map comparison and seven of the intervals between the D . miranda and D . pseudoobscura–Pikes Peak two-map comparison were significantly different after correction for multiple tests . Six of the 72 intervals between D . miranda and D . pseudoobscura–Flagstaff two-map comparison exhibited at least a 3-fold difference , and 12 of 102 intervals between D . miranda and D . pseudoobscura–Pikes Peak exhibited at least a 3-fold difference ( Dataset S2 ) . Twenty-seven of 97 condensed intervals ( three-map comparison , condensed between all three maps ) for chromosome 2 were considered to be “conserved” within and between species . This means that they displayed a nonsignificant difference across all three maps when analyzed with a rare events logistic regression and had an odds ratio between 0 . 62 and 1 . 615 after the effect of map identity was taken into account . These “conserved” intervals were used for further downstream analyses ( see “Diversity , Divergence , and Recombination”; Table S7 ) . For the XR , seven of 44 intervals condensed between all three maps were conserved within and between species according to the criteria outlined above . In sum , we observe strong conservation in recombination rates within a single species , while between species , we see globally elevated recombination rates in D . miranda . Once the global difference is accounted for , there are few intervals with significant differences in recombination rate within and between species . Thus , it is possible and parsimonious that recombination rate is generally conserved at the scale examined here ( ∼180 kb ) over moderate evolutionary timescales ( 2–2 . 5 my ) . We used various Illumina platforms to resequence genomic DNA from 10 D . pseudoobscura lines using virgin females from lines that were inbred for five or more generations with full-sibling single-pair mating ( Table S8 ) . Drosophila pseudoobscura populations across North America display very little differentiation , as indicated by low FST values ( always<0 . 10 , often<0 . 05 for loci located outside of the inversion polymorphisms of the third chromosome ) [60] , [61] . Therefore , the choice of strains sequenced for estimating diversity covered much of the species range but was fairly random . We also sequenced two lines of D . persimilis ( one of these was provided by S . Nuzhdin ) , two lines of D . pseudoobscura bogotana ( one of these was provided by S . Nuzhdin ) , one line of D . lowei , and three lines of D . miranda ( two provided by D . Bachtrog , Table S8; Short Read Archive accession numbers SRA044960 . 1 , SRA044955 . 2 , and SRA044956 . 1; see also http://pseudobase . biology . duke . edu/ ) . The divergence between D . persimilis and D . lowei was used to generate measures of a proxy for neutral mutation rate across the genome . In all diversity and divergence calculations , the reference sequences for the D . pseudoobscura and D . persimilis genomes were both included [62] , [63] . Details of diversity and divergence calculations are discussed in Text S1 ( see section titled “Fine-Scale Recombination Maps: Computational Methods for Diversity and Divergence Measures” ) . Briefly , average pairwise diversity and divergence was calculated for 4-fold degenerate sites , focusing exclusively on unpreferred codons [64] , though we obtained very similar results when using all 4-fold degenerate sites . Overall , recombination is significantly and positively associated with average pairwise diversity but not average pairwise divergence at 4-fold degenerate sites of unpreferred codons . We examined this relationship in several ways . We analyzed each chromosome for each uncondensed recombination map independently using a generalized linear model for diversity and a separate model for divergence ( Tables S9 , S10 , and S11 ) . After accounting for multiple covariates , diversity at 4-fold degenerate sites of unpreferred codons shows a significant , positive relationship with recombination , while divergence at 4-fold degenerate sites of unpreferred codons does not ( Tables S9 and S10 ) . This result is consistent for each of the three recombination maps ( D . pseudoobscura–Flagstaff , D . pseudoobscura–Pikes Peak , and D . miranda ) for both chromosome 2 and the XR chromosome arm ( Tables S9 and S10 ) . The XL chromosome arm contained too few intervals for analysis for D . pseudoobscura–Flagstaff . For D . pseudoobscura–Pikes Peak and D . miranda , diversity showed a significant , or nearly significant , positive relationship with recombination , while divergence did not ( Table S11 ) . The analysis above suggests that the recombination–diversity relationship is probably the result of the effect of recombination on selection at linked sites ( sensu [13] , [18] ) ; however , inadvertently including regions with discordant recombination rates between species in the analysis above could result in a pattern that supports this hypothesis—even when recombination is predominantly mutagenic ( Figure S1 ) . To resolve this potential bias , we restricted analysis to only regions that exhibited conserved recombination rates between all three chromosome 2 maps ( N = 27 intervals; described above ) and examined recombination in association with average pairwise D . pseudoobscura diversity at 4-fold degenerate sites of unpreferred codons ( Table 2; Figures S7 and S8 ) and average pairwise D . pseudoobscura–D . miranda divergence at 4-fold degenerate sites of unpreferred codons ( Table 3; Figures S7 and S8 ) . The effect of recombination on diversity was significant when the analysis was restricted to only those regions with the most conserved recombination rates ( quasibinomial GLM , F = 6 . 123 , p value = 0 . 024 ) , and the effect of recombination on divergence remained nonsignificant ( quasibinomial GLM , F = 0 . 138 , p value = 0 . 714 ) . These regions contained only one interval within 4 Mb of the telomeric end and no intervals within 4 Mb of the centromeric end of the chromosome; thus , these results are not a function of broad-scale regional recombination rate differences across the chromosome . These results support the hypothesis that recombination affects diversity through the effect of selection on linked sites . We did not perform an analysis on conserved windows for the X chromosome , as only seven intervals were conserved within and between species . To determine the impact of recombination rate on selection at linked sites in the genome , we used two generalized linear models to analyze the relationship of recombination rate and several measures that may be indicative of the efficiency of selection: ( 1 ) abundance of nonsynonymous substitutions and ( 2 ) average pairwise nucleotide diversity at 4-fold degenerate sites around nonsynonymous substitutions . We analyzed the association of recombination rate with these two measures in a generalized linear model framework to account for covariates such as gene density , GC content , and a proxy for neutral mutation rate . Biased gene conversion may influence substitution rates; thus , we controlled for GC content in all of the analyses below [7] , [16] , [65] , [66] . We did not consider gene expression as a covariate , though some studies point to a negative relationship with recombination rate [67] . The relationship of recombination rate to nonsynonymous substitution abundance was examined with the D . pseudoobscura Flagstaff fine-scale recombination maps . Nonsynonymous substitution abundance was measured as the nonsynonymous substitutions on the branch leading to D . pseudoobscura+D . persimilis as identified with PAML . The response variable was the number of nonsynonymous substitutions in each gene , and the covariates of the linear model included ( 1 ) the number of synonymous substitutions in the gene in question allowing for inclusion of genes where Ks = 0 ( sensu [50] ) , ( 2 ) , GC content of the gene , ( 3 ) gene density of 50 kb on either side of the midpoint of the gene , and ( 4 ) average pairwise divergence at 4-fold degenerate sites of unpreferred codons between D . persimilis and D . lowei as a proxy for neutral mutation rate within the gene . We found no relationship ( Table 4 ) between recombination and nonsynonymous substitution abundance with the fine-scale data ( generalized linear model with Poisson distribution , z = −0 . 614 , p = 0 . 539 ) . In response to selective sweeps , a trough in diversity should be visible around selected variants [30] , [68]–[72] . We analyzed diversity surrounding the nonsynonymous substitutions along the lineage leading to D . pseudoobscura+D . persimilis identified by PAML . We compared the average pairwise diversity patterns at 4-fold degenerate sites surrounding these substitutions in relation to the Flagstaff recombination rate and distance in basepairs from the substitution ( Text S1 ) . In regions with high recombination rates , the footprints of selection are thought to be narrower than in regions with low recombination rates , where strong linkage between sites will create a stronger signature of sweeps [39] , [69] , [71] , [73] . As a control , similar analyses were performed using synonymous substitutions along the D . pseudoobscura+D . persimilis lineage following [68] . Synonymous substitutions , in many cases , evolve in a more neutral fashion than nonsynonymous substitutions ( [68] , but see [74] , [75] ) . In a recent genome-scale analysis conducted with data similar to what are presented here , little reduction in diversity was seen around synonymous substitutions [68]; this study instead saw an increase in diversity , which disappeared after correction for local mutation rates . We considered 60 kb on either side of the substitution along the D . pseudoobscura lineage divided into 1 , 000 bp nonoverlapping windows ( sensu [68] ) . For each 1 , 000 bp window , the response variable was the number of polymorphic 4-fold degenerate sites . The generalized linear model included the following covariates: ( 1 ) total 4-fold degenerate sites , ( 2 ) GC content , ( 3 ) proportion of coding bases , ( 4 ) divergence of D . lowei–D . persimilis at 4-fold degenerate sites as a proxy for neutral mutation rate , and ( 5 ) proportion of bases that were nonsynonymous substitutions . The identities of each nonsynonymous substitution were included as random effects . This generalized linear mixed model with Poisson distribution included the following factors: absolute physical distance from the substitution , fine-scale-derived estimates of recombination rate , and the interaction between these two factors . A negative interaction term means that short distances from a substitution and high recombination rates have similar effects on diversity as large distances and low recombination rates . We expect the interaction term for distance and recombination rate to be much reduced in magnitude for synonymous substitutions in comparison to the nonsynonymous analysis . We found a small but significant negative interaction term of physical distance from the nonsynonymous site and recombination rate on nucleotide diversity around nonsynonymous substitutions ( Poisson GLMM , z = −7 . 52 , p<0 . 001; Table 5 , Figures 3 and S9 ) . In other words , higher rates of recombination allow for recovery of diversity at shorter physical distances from the nonsynonymous site than lower recombination rates ( Figure S9 ) . In contrast , a weaker interaction was detected for the interaction of distance and recombination rate on diversity around synonymous substitutions along the D . pseudoobscura lineage ( Poisson GLMM , z = −2 . 43 , p = 0 . 015; Table 6 , Figures 3 and S9 ) . GLM plots for the very low recombination rates of <0 . 5 cM/Mb show wider dips in diversity ( and more associated noise; Figure S9 ) than plots for recombination rates of >0 . 5 cM/Mb ( Figure S9 ) . Distance from a substitution had a positive , significant effect on diversity as expected if linked selection of substitutions generates a dip in diversity ( Tables 5 , 6 , and S12 ) . Recombination rate also had a positive , significant effect on diversity as expected , if either recombination was mutagenic or if positive/negative selection was operating on the chromosome ( Tables 5 , 6 , and S12 ) . The proportion of nonsynonymous substitutions around a substitution had a negative significant effect on diversity surrounding a nonsynonymous site as expected if many of these substitutions combine forces to generate stronger selective sweeps ( Tables 5 , 6 , and S12 ) . The interaction term pointing to deeper dips in diversity for lower recombination rates is no longer significant when examining only 5 kb or 15 kb on either side of the focal substitution ( it is negative for nonsynonymous substitutions and positive for synonymous substitutions ) , but it is conceivable that this lack of significance represents an issue with window size or sampling .
Here and in other recent work [54] , we demonstrate that ultrafine-scale patterns of crossover rate ( intervals spanning 20 kb ) are also significantly heterogeneous in D . pseudoobscura . In each ultrafine region on chromosome 2 , recombination rates varied by up to 6-fold ( 17 Mb region ) over only approximately 120 kb ( 6 Mb region variation is 3 . 6-fold , and 21 Mb region variation is 5 . 1-fold ) , and ultrafine-scale maps reveal variation not detected in the fine-scale maps . This was especially apparent for the 17 Mb region , where ultrafine-scale recombination rates ranged from 3 . 5 to 21 . 2 cM/Mb , and fine-scale recombination rates in the same area ranged only from 4 . 4 to 5 . 6 cM/Mb . This heterogeneity suggests that our fine-scale measures ( intervals spanning <200 kb ) are averages of actual variation in recombination rate . In humans , broad-scale variation averages over the density and intensity of ∼2 kb hotspots that occur in clusters every 60–90 kb [78] , [79] . The majority of recombination occurs at these hotspots , and the majority of recombination is governed by the DNA binding protein PRDM9 and its recognition motifs in humans [17] , [80]–[84] . Interestingly , several studies in different regions of the D . melanogaster genome indicate that linkage disequilibrium decays rapidly [37] , [85]–[87] , suggesting that the heterogeneity we observed in ultrafine-scale maps may not be governed by clustered hotspots similar to those in humans , or at least that a nontrivial amount of recombination may occur outside such “hotspots . ” To assess whether “hotspots” of some sort exist in D . pseudoobscura , genome-wide patterns of linkage decay need to be investigated or incredibly fine-scale maps ( interval size <5 kb ) need to be made . Such a line of inquiry would help address basic questions about the requirements for functional recombination across various taxa . For example , there are several notable differences regarding the formation and function of the synaptonemal complex and the role of double-strand breaks across taxa [88]–[93] . Furthermore , the Drosophila lineage completely lacks several proteins essential for generating crossovers and double-strand break repair in other organisms [89] , [94] . It is likely that understanding particular sequence features associated with recombination on a kilobase scale in Drosophila will uncover more details about the mechanistic underpinnings of meiosis that differentiate these species and the distribution of crossovers across the genome . Recombination rates at broad scales are conserved between populations and species [33] , [95]–[100] ( see also review in [20] ) . Our fine-scale data are generally consistent with these findings except that D . pseudoobscura has about three-fourths the rate of recombination , on average , as D . miranda for chromosome 2 and about three-fifths the rate of recombination of D . miranda on the XR chromosome arm . Notably , D . melanogaster has one of the lowest recombination rates in the genus , as evidence indicates that D . mauritiana , D . simulans , D . virilis , D . pseudoobscura , D . miranda , and D . persimilis all exhibit higher rates of recombination [33] , [53] , [99]; this should be considered when interpreting hitchhiking and linkage data from D . melanogaster to patterns of recombination in Drosophila in general . Our results indicate that recombination affects diversity through mediating selection in the genome . While accounting for multiple covariates , we found no association between recombination and average pairwise divergence at 4-fold degenerate sites of unpreferred codons , and a significant , positive association of recombination with average pairwise diversity at 4-fold degenerate sites of unpreferred codons . Using data from our fine-scale maps , we ensured that recombination rates are nearly identical between the species used to generate divergence estimates; thus , we absolved a key assumption made in previous studies ( see Figure S1 ) . Data from Drosophila suggest both positive and negative selection are markedly less efficient in nearly nonrecombining regions of the genome [12] , [47] , [76] , [101] , [102] , and a relationship of diversity but not divergence to recombination is apparent for other species of Drosophila [13] , [33] , [40] , [49] , mouse [36] , beet [35] , tomato [103] , [104] , Caenorhabditis [38] , and yeast [105] . This last example is especially interesting because recombination is known to be mutagenic in yeast [24] , [27] , but there is a negative or absent divergence–recombination correlation [34] , [105]; thus , it may be that recombination is somewhat mutagenic in many organisms , but the power of recombination to modulate the diversity eroding effects of selection likely has a much greater impact on the genome . In other systems , the divergence–recombination association is positive , which may be interpreted as evidence that recombination is predominately mutagenic . A positive divergence–recombination association is apparent for humans [106] , [107] , maize [108] , and in an inverted region between D . pseudoobscura and D . persimilis [25] . This association may be attributable to mutation [21] , but unmeasured variables or segregating ancestral polymorphism could predispose a system to exhibiting a positive divergence–recombination relationship [34] , [38]–[41] . For instance , in C . briggsae , segregating ancestral polymorphism leads to the signature of recombination-associated mutation ( i . e . , a positive divergence–recombination association ) , but further examination shows the majority of polymorphism heterogeneity is caused by recombination affecting the impact of selection at linked sites [38] . Since recombination probably mediates the effects of hitchhiking in our system , we sought to understand whether this hitchhiking is primarily positive or negative ( background , purifying ) selection and if recombination rate variation has a significant impact on the potential efficacy of selection . Evidence is emerging that in many organisms , especially those with large population sizes , selection may play a substantial role in shaping the genome [109] . For partial selfers , it seems that background selection substantially affects the genome [110]–[113] , while in outcrossing species Drosophila , mice , and Capsella grandiflora a large fraction of the genome may be influenced by positive selection [40] , [114]–[116] . The majority of studies find strong support that recombination can shape adaptive evolution when comparing regions of no recombination to regions with some or abundant recombination ( reviewed in [7] ) . However , after accounting for multiple covariates in regions with detectable recombination rates , there is often very little relationship between recombination rate and the efficacy of selection [7] , [12] , [65] . Across chromosome 2 , we found no relationship between the number of nonsynonymous substitutions and the recombination rate as measured with our fine-scale Flagstaff map . Reanalysis of the fine-scale data after removal of the first and last 3 Mb of the chromosome did not change the relationship of fine-scale recombination rate to nonsynonymous substitutions . Our observation of a reduction of average pairwise diversity at 4-fold degenerate sites around nonsynonymous substitutions ( Figure S9 ) is consistent with the idea that positive selection may have fixed many nonsynonymous substitutions along the ancestral lineage leading to D . pseudoobscura+D . persimilis , as has been argued elsewhere for other Drosophila species [68] , [117] . While potentially less common , dips in diversity could also be caused by deleterious mutations that can get fixed by chance if deleterious selection coefficients are small enough—a situation we call “loser's luck” ( Figure S10; but see [117] , [118] ) , and theoretical investigations of entirely neutral substitutions showed that their quick fixation can also lead to dips in diversity [119] . Thus , while many of the dips in diversity we see may be caused by positive selection , both loser's luck and fixation of neutral substitutions may also contribute . Diversity may be recovered slightly farther from a nonsynonymous substitution in areas of low recombination than in areas of high recombination , and such a relationship is not as pronounced for synonymous substitutions fixed along the lineage leading from the common ancestor of D . pseudoobscura and D . persimilis ( Tables 5 and 6; Figure S9 ) . Similarly , in Arabidopsis , haplotype blocks around nonsynonymous SNPs are larger than around synonymous SNPs [120] . Our data agree with theoretical expectations [69] , [71] and past studies that show negative correlations of polymorphisms and nonsynonymous substitutions in Drosophila ( [40] , [68] , [121] , [122]; indeed , our data also show a significant negative relationship for nonsynonymous substitutions and within-species polymorphisms , generally ( Tables 5 and 6 ) . Yet the negative interaction term between recombination rate and distance from focal substitutions we observed is dependent on window size and distance from the substitution examined . Our study documented global and local differences in recombination rate between two closely related species , and these data indicate that recombination probably modulates Hill–Robertson effects in the genome , causing a positive association of diversity with recombination . While we found no overall association of recombination rate with the number of nonsynonymous substitutions at the fine scale , we found evidence for dips in diversity around nonsynonymous substitutions that are dependent on the distance from the substitution , local recombination rate , and a number of other factors . In total , our study adds to the growing literature that indicates that selection must be a ubiquitously important factor for shaping diversity across much of the genome [30] , [69] , [71] .
Using a backcross design , we developed two recombination maps for D . pseudoobscura ( Flagstaff and Pikes Peak ) and one recombination map for D . miranda ( Text S1 ) . For each cross , Duke's Genomic Analysis Facility genotyped 1 , 440 individual backcrossed flies for 384 line-specific SNP markers ( see “SNP Development” section in Text S1 ) using the Illumina BeadArray platform ( Illumina , San Diego , CA ) [123] . Recombination events were scored when an individual fly's genotype changed from heterozygous to homozygous ( for the parent in the backcross ) or vice versa for autosomes and when the fly's genotype changed between the possible allele combinations for the sex chromosome arms XL and XR . Double crossovers were defined as adjacent intervals with different genotypes on both sides ( for instance , a single homozygote genotype call nested in a tract of heterozygote genotype calls ) . We deemed these as genotyping errors as crossover interference is high within 2 Mb [124] and removed the single inconsistent genotype , scoring it as missing data . CentiMorgans were defined as the number of recombination events over the total number of individuals examined for each recombination interval , and we scaled this raw measure with a correction for recombination interference [125] . Throughout the article , recombination rates are given in Kosambi centiMorgans [125] per Megabase ( cM/Mb ) . Approximately 1 , 400 backcross progeny were scored for the Pikes Peak D . pseudoobscura map , approximately 1 , 250 backcross progeny were scored for the Flagstaff D . pseudoobscura map , and approximately 1 , 170 backcross progeny were scored for the D . miranda map ( see Table S1 for the final number of individuals , number of intervals , and size of intervals over which recombination was measured ) . Physical genomic distances used to calculate centiMorgans per Megabase ( cM/Mb ) per interval were based on the D . pseudoobscura reference genome v2 . 6 ( Flagstaff ) and v2 . 9 ( Pikes Peak , D . miranda ) . Marker order was confirmed by the R ( The R Foundation for Statistical Computing 2010 ) package OneMap [126] using the algorithms Recombination Counting and Ordering [127] and Unidirectional Growth [128] . Onemap does not accommodate backcrossed designs for sex chromosomes; therefore , we specified an F2 intercross design in these cases . We found one small inversion in D . miranda relative to D . pseudoobscura on chromosome 2 . We estimated the left breakpoint was between the markers at 10 , 491 , 527 and 10 , 660 , 216 bp , and the right breakpoint was between the markers at 13 , 318 , 705 bp and 14 , 068 , 383 bp from the telomeric end of chromosome 2 . This inversion corresponds to one previously documented between D . miranda and D . pseudoobscura between markers rosy and nop56 [129] . Figure S6 illustrates that recombination rate differences are probably not due to differences in gene order; thus , we used the D . pseudoobscura orientation for this inversion when comparing recombination between maps and excluded intervals that included the breakpoints . Confidence intervals ( 95% ) for cM/Mb for each recombination interval were calculated by permutation [33] , [54] . Confidence intervals for those intervals where we did not find a single recombinant individual were estimated from a binomial distribution—simply , we solved the equation ( 1−x ) N = 0 . 05 , where x is the 95% upper bound of recombination frequency , and N is the number of individuals surveyed . The rationale for regressing out the effect of species ( when identifying conserved intervals ) was to account for the globally higher recombination rate in D . miranda relative to D . pseudoobscura and to identify regions where the recombination profile overlapped ( e . g . , where peaks and troughs can be overlaid ) . To delimit conserved regions using data that have not been corrected for elevated recombination rate of D . miranda , one might identify a region with very similar recombination rates between D . miranda and D . pseudoobscura , but this region may be a trough in recombination rate for D . miranda and a peak in recombination rate for D . pseudoobscura . Not correcting for the global elevation of D . miranda may lead to falsely concluding that a region has a conserved recombination profile between two maps . Thus , we used a rare events logistic regression ( Zelig package in R ) between each set of condensed fine-scale recombination maps to identify regions of conserved recombination after accounting for map identity ( Flagstaff–Pikes Peak , Flagstaff–D . miranda , Pikes Peak–D . miranda ) . The package Zelig uses the same model as a logistic regression , but it corrects for a bias that is introduced when the sample contains many more of one of the dichotomous outcomes than the other . Recombination events conditioned on the total number of observations was the response variable , and species , interval , and species-by-interval were included as factors in the model . We defined “divergent” intervals as those where tests in each interval between the species from the rare events logistic regression had a q-value of <0 . 05 after correction for multiple tests [59] . “Conserved intervals” were those intervals that displayed a nonsignificant difference across all three maps when analyzed with a rare events logistic regression and had an odds ratio between 0 . 62 and 1 . 615 , after accounting for a species effect . We did not correct for multiple tests in defining conserved intervals . The effect size , the confidence intervals for the effect size , p values , and multiple-test corrected q-values are available in Datasets S1 , S2 , and S3 . In this way , only intervals that were conserved within and between species were delineated as conserved intervals . The final dataset used to differentiate between the mutagenic and selection hypotheses contained 27 conserved intervals on chromosome 2 . We did not use the XR to differentiate between the mutagenic and selection hypotheses—of the 44 intervals condensed across three XR maps , only seven were conserved within and between species . We chose not to combine data from chromosome 2 and XR , as there is some evidence for different evolutionary patterns between autosomal and sex chromosomes in Drosophila [130] . Details of how diversity and divergence were measured from the next generation sequencing data are given in Text S1 . We analyzed the effect of recombination on diversity and divergence by applying a quasibinomial GLM as the data were overdispersed , which has several statistical properties favorable to analyzing proportions such as pairwise diversity [131] , [132] . Diversity or divergence was used as a response variable by binding the number of SNP bases to the number of non-SNP , eligible bases with cbind in R . We included recombination rate , proportion of G or C bases within the recombination interval , gene density ( measured as a proportion of nucleotides within the recombination interval that are coding ) , a proxy for neutral mutation rate ( see Text S1 ) , and interaction terms as factors in the model . See Text S1 for filtering steps that were required for a nucleotide to be considered an eligible base . For these models , the analysis presented is restricted to those conserved , condensed intervals with highly similar recombination rates between all three maps , unless otherwise noted . This restriction removes a classic bias by requiring that the intervals have similar recombination rates between the two species compared for the divergence measures ( Figure S1 ) . Similar linear models were also analyzed using the uncondensed intervals for each of the three maps individually ( Tables S9 , S10 , and S11 ) . All statistics were performed in R version 2 . 12 . 1 ( The R Foundation for Statistical Computing 2010 ) unless otherwise noted . Using Flagstaff 16 and Flagstaff 14 , we followed the same backcross scheme described in the section “Fine-Crossover Maps: Crosses and Technical Details . ” Over 10 , 000 progeny from this backcross were stored in 96-well plates , frozen at −20°C and amplified for markers over these three regions . PCR products were visualized on a polyacrylamide gel using LICOR 4300 ( see the section “Ultrafine Crossover Maps” in Text S1 ) . The number of nonsynonymous substitutions , specific to the D . pseudoobscura+D . persimilis lineage , were calculated for each gene using PAML using the resequenced genomic and reference genomic data described in Table S8 ( one D . lowei , three D . miranda , three D . persimilis , two D . pseudoobscura bogotana , and 11 D . pseudoobscura genomes , filtered for quality as described above ) . We used a tree rooted with D . lowei and considered the branches leading to [D . persimilis ( D . pseudoobscura , D . pseudoobscura bogotana ) ] to be the foreground branches ( additional details in Text S1 ) . We included D . persimilis a part of the foreground branch because relatively extensive interbreeding occurs between D . pseudoobscura and D . persimilis across much of the genome , aside from a few inverted regions [133]–[135] . Following [50] , we used a GLMM with Poisson distribution to examine the potential for recombination rate to shape the distribution of nonsynonymous substitutions along the D . pseudoobscura+D . persimilis lineage . The model contained the following main effects: the number of silent segregating sites in each gene , GC content in each gene within Flagstaff 16 , the proportion of coding bases 50 kb on either side of the gene's midpoint , weakly selected average pairwise divergence within the gene between D . persimilis and D . lowei at 4-fold degenerate sites of unpreferred codons ( a proxy for neutral mutation rate ) , recombination rate observed for the interval containing the gene , and a random variable included to account for pseudoreplication of multiple genes per interval . The response variable was the number of nonsynonymous substitutions observed in each gene . This model construction allowed the inclusion of genes whose synonymous substitution count was zero ( sensu [50] ) . The GC content from Flagstaff16 was used as this was the line used for backcrossing in the crossing scheme , and the Flagstaff map ( D . pseudoobscura ) was used in this analysis . We used 4-fold degenerate sites of unpreferred codons to measure the average levels of diversity as a function of distance from amino acid substitutions along the D . pseudoobscura+D . persimilis lineage ( as identified by PAML , see above ) . Generalized linear mixed models with a Poisson distribution were used to compare the diversity around nonsynonymous substitutions along the D . pseudoobscura+D . persimilis lineage in relation to distance from the site and recombination rates measured in the Flagstaff cross . Measures of diversity at 4-fold degenerate sites were taken 60 kb ( sensu [68] ) from the site in either direction ( 120 kb total ) with nonoverlapping bins of 1 , 000 bp . The random effects of identities of each substitution were estimated . We included as covariates ( 1 ) divergence between D . persimilis and D . lowei at 4-fold degenerate sites ( a proxy for neutral mutation rate ) , ( 2 ) proportion of bases that were either G or C in Flagstaff 16 within the 1 , 000 bp window , ( 3 ) proportion of codons that were nonsynonymous substitutions within the 1 , 000 bp window , and ( 4 ) proportion of bases that were coding over each 1 , 000 bp window . The absolute value of the distance from the site and local recombination rate ( at the particular nonsynonymous substitution ) were included in the model as well as the interaction between distance and recombination rate . All effects in the model were standardized to mean zero and unit standard deviation . As a control , similar analyses were performed using synonymous substitutions along the D . pseudoobscura+D . persimilis lineage . Synonymous substitutions should evolve in a more neutral fashion; thus , less of an interaction between distance and recombination rate is expected . Any 1 , 000 bp window with less than 75 eligible , 4-fold degenerate sites was excluded from the analysis . Any nonsynonymous or nonsynonymous changes with less than 10 windows were excluded from the analysis . For the 60 kb analysis , after all filtering steps , our data consisted of 4 , 338 nonsynonymous and 8 , 670 synonymous substitutions along the D . pseudoobscura+D . persimilis lineage on chromosome 2 . Four-fold degenerate sites were used here , rather than 4-fold degenerate sites at unpreferred codons , because too little data were available in each 1 , 000 bp nonoverlapping window . | Individuals within a species differ in the DNA sequences of their genes . This sequence variation affects how well individuals survive or reproduce and is transmitted to their offspring . Genes near each other on individual chromosomes tend to be passed to offspring together—neighboring genes are unlikely to be separated by exchanges of genetic material derived from different parents during meiotic recombination . When genes are inherited together , however , the evolutionary forces acting on one gene can interfere with variation at its neighbors . Thus , variation at multiple genes can be lost if natural selection acts on one gene in close proximity . Recombination can prevent or reduce this loss of variation , but previous tests of this phenomenon failed to account for recombination rate differences between species . In this study , we show that some parts of the genome differ in recombination rate between two species of fruit fly , Drosophila pseudoobscura and D . miranda . Avoiding an assumption made in previous studies , we then examine sequence variation within and between fly species in those parts of the genome that have conserved recombination rates . Based on the results , we conclude that recombination indeed preserves variation within species that would otherwise have been eliminated by natural selection . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"genome",
"evolution",
"neutral",
"theory",
"population",
"genetics",
"genome",
"sequencing",
"mutation",
"genome",
"databases",
"genome",
"complexity",
"genetic",
"polymorphism",
"biology",
"genetic",
"drift",
"sequence",
"databases",
"natural",
"selection",
"genetics",
... | 2012 | Recombination Modulates How Selection Affects Linked Sites in Drosophila |
Chelt , a cholera-like toxin from Vibrio cholerae , and Certhrax , an anthrax-like toxin from Bacillus cereus , are among six new bacterial protein toxins we identified and characterized using in silico and cell-based techniques . We also uncovered medically relevant toxins from Mycobacterium avium and Enterococcus faecalis . We found agriculturally relevant toxins in Photorhabdus luminescens and Vibrio splendidus . These toxins belong to the ADP-ribosyltransferase family that has conserved structure despite low sequence identity . Therefore , our search for new toxins combined fold recognition with rules for filtering sequences – including a primary sequence pattern – to reduce reliance on sequence identity and identify toxins using structure . We used computers to build models and analyzed each new toxin to understand features including: structure , secretion , cell entry , activation , NAD+ substrate binding , intracellular target binding and the reaction mechanism . We confirmed activity using a yeast growth test . In this era where an expanding protein structure library complements abundant protein sequence data – and we need high-throughput validation – our approach provides insight into the newest toxin ADP-ribosyltransferases .
Sequence data from over 6 , 500 genome projects is available through the Genomes OnLine Database [1] and more than 60 , 000 protein structures are in the Protein Data Bank ( PDB ) . While these sequences represent large diversity , a limited number of possible folds – estimated at 1 , 700 [2] – helps researchers organize the sequences by structure . A single fold performs a limited number of functions , between 1 . 2 and 1 . 8 on average [3] . Therefore , structure knowledge helps pinpoint function . Researchers are combining sequence and structure data to expand protein families such as the mono-ADP-ribosyltransferase ( mART ) protein toxins that participate in human diseases including diphtheria , cholera and whooping cough [4] . ADP-ribosylation is a post-translational modification that plays a role in many settings [5] . ADP-ribosyltransferases ( ADPRTs ) bind NAD+ and covalently transfer a single or poly ADP-ribose to a macromolecule target , usually protein , changing its activity . Many prokaryotic ADPRT toxins damage host cells by mono-ADP-ribosylating intracellular targets . G-proteins are common targets including: eukaryotic elongation factor 2 ( ADP-ribosylation halts protein synthesis ) , elongation factor thermo unstable , Ras , Rho ( ADP-ribosylation locks Rho GTPase in the GDP-bound state and disaggregates the actin cytoskeleton ) and Gs-α ( ADP-ribosylation interrupts signal transduction ) . Other targets include actin ( ADP-ribosylation inhibits actin polymerization ) [6]; kinase regulators ( ADP-ribosylation inhibits phagocytosis ) [7] and RNA-recognition motifs ( ADP-ribosylation alters the transcriptome and weakens immunity ) [8] . Researchers use ADPRT toxins to develop vaccines [9] , as drug targets , to kill cancer cells [10] , as stent coatings to prevent restenosis after angioplasty [11] , as insecticides , to deliver foreign proteins into cells using toxin receptor-binding and membrane translocation domains , to study cell biology [12] , [13] , to understand the ADP-ribosylation reaction and to identify biosecurity risks . ADPRTs occur in viruses , prokaryotes , archaea and eukaryotes . Genomes acquire them through horizontal gene transfer [14]–[17] . Several authors have reviewed the prokaryotic ADPRT family [6] , [18] , [19] . Examples include Pseudomonas aeruginosa exoenzyme S ( ExoS ) , Vibrio cholerae cholera toxin ( CT ) , Bordetella pertussis pertussis toxin ( PT ) and Corynebacterium diphtheriae diphtheria toxin ( DT ) . Toxic ADPRTs are divided into the CT and DT groups to better organize the family . We focus on the CT group , which we divide into the ExoS-like , C2-like , C3-like and CT-PT-like toxins . CT group primary sequences are related through a specific structure-linked pattern ( Figures 1 and 2 ) [20] . The ADPRT pattern , updated from previous reports [4] , [21] and written as a regular expression is: The toxin catalytic domain consists of several regions . We describe them here going from the N- to C-terminus using previously introduced nomenclature [20] , [22] . Region A ( not shown ) is sometimes present and recognizes substrate , when ExoT recognizes Crk , for example . Its recognition of ExoT targets is an exception rather than a general rule for ADPRTs . Except for the CT-PT-like subgroup , region B – an active site loop flanked by two helices – appears early in the toxin sequence . It stabilizes the “catalytic” Glu , binds the nicotinamide ribose ( N-ribose ) and the adenine phosphate ( A-phosphate ) . It also stabilizes the target substrate and helps specific bonds rotate during the ADPRT reaction , in turn , helping to bring the nucleophile and electrophile together for reaction . ( The CT-PT-like subgroup lacks region B and instead has a knob region that precedes region 2; these might function interchangeably . ) Region 1 is at the end of a β-sheet , with sequence pattern [YFL]RX . It is important for binding A-phosphate , nicotinamide phosphate ( N-phosphate ) , nicotinamide , adenine ribose ( A-ribose ) and the target substrate . Region F ( not shown ) follows region 1 and sometimes recognizes substrate . The region 2 ( STS motif ) follows on a β-sheet with sequence pattern [YF]-X-S-T-[SQT] . It binds adenine , positions the “catalytic” Glu , orients the ADP-ribosyl-turn-turn ( ARTT ) loop and maintains active site integrity . The phosphate-nicotinamide ( PN ) loop ( also known as region E ) is immediately after the STS motif . It interacts with the target and binds N-phosphate . Menetrey et al . suggested the PN loop is flexible and implicated it in locking the nicotinamide in place during the reaction [23] . Region 3 ( also known as region C ) consists of the ARTT loop leading into the β-sheet with pattern [QE]-X-E . It recognizes and stabilizes the target and binds the N-ribose to create a strained NAD+ conformation . The ARTT loop is plastic , having both “in” and “out” forms that might aid substrate recognition [23] . The FAS region ( also known as region D , not shown ) mediates activator binding when present [6] , [22] , [24] , [25] . Researchers have long debated the ADPRT reaction details . Some suggest an SN2 mechanism [26] , [27] , but many now favor the SN1 mechanism [28]–[32] . Tsuge et al . recently devised a specific version of this mechanism for iota toxin , which we follow closely in this work [33] , [34] . The reaction follows three steps: the toxin cleaves nicotinamide to form an oxacarbenium ion , the oxacarbenium O5D-PN bond rotates to relieve strain and forms a second ionic intermediate . ( The electrophile and nucleophile might migrate by an unknown mechanism to further reduce the distance between them . ) Finally , the target makes a nucleophilic attack on the second ionic intermediate . The SN1mechansim – believed widely applicable to CT group toxins – is a template for new toxins given the historical structure similarity and consistent NAD+ conformation in the active site as shown in Figures 1 and 2 . Quaternary structure for the toxins is wide-ranging . Several combinations exist for toxin domains ( A ) and receptor binding or membrane translocation domains ( B ) . The B domains have diverse structures and functions and exist as fusions or separate polypeptides . Various formats include: A-only , two-domain AB ( single polypeptide ) , three-domain AB ( single polypeptide ) and AB5 ( multiple polypeptides ) . C3-like toxins are A-only . ExoS-like toxins have toxic A-domains and are often paired with Rho GTPase activating protein ( RhoGAP ) , which are not true B domains . C2-like toxins are AB toxins that contain B domains that are structural duplicates of the A domain . These B domains are not toxins; they bind proteins that are similar to anthrax protective antigen ( PA ) including Vip1 , C2-II and Iota Ib [35] , [36] . DT group toxins are three-domain , single polypeptide AB toxins where the B domain contains both a receptor-binding and a membrane-translocation domain . The CT-PT-like toxins are AB5 and have B domains that form a receptor-binding pentamer [37] . Low overall sequence identity hampers conventional sequence-based homology searches [17] , [20] , [38]–[40] . One challenge – key to filling gaps in the toxin family – is to link new sequences and known toxins . Depending only on amino acid sequence alignment techniques to discover new toxins is imprudent . Instead the trend is to use more structure information in the search because many primary sequences produce the same fold [41] . Researchers can then link these sequences through fold recognition [42] . Otto et al . used PSI-BLAST to identify new ADPRT toxins , including SpvB from Salmonella enterica [14] . More recently a similar strategy yielded 20 potential new toxins [15] . This led to interesting examples later characterized including: CARDS toxin from Mycoplasma pneumonia [43] , SpyA from Streptococcus pyogenes [44] and HopU1 from Pseudomonas syringae [8] . PSI-BLAST is a classic way to expand protein families , but it has limits . For example , unrelated sequences often “capture” the search . Also , nearly a decade has passed since Pallen et al . released the last detailed data mining results for the toxin family [15] . The sequence and structure databases – and remote homolog detection tools – have advanced during this time . Masignani et al . proposed that a match between the conserved ADPRT pattern with corresponding secondary structure is one way to reduce dependence on sequence identity . The pattern helps ensure function and reduces the total sequence set to a smaller subset for screening; secondary structure prediction ensures that key active site parts are present [17] . Our contribution is to expand ADPRT toxin family using a new approach . The difference is that we use fold-recognition searches extensively rather than relying on PSI-BLAST or secondary structure prediction . Our genomic data mining combines pattern- and structure-based searches . A bioinformatics toolset allows us to discover new toxins , classify and rank them and assess their structure and function . Often , data mining studies simply present a table of hits with aligned sequences , but do not interpret or analyze those hits in detail . Our aim – rather than to explicitly confirm the roles of the six proteins , 15 domains , 18 loops and 120+ residues discussed – is to develop a theoretical framework for understanding new toxins , based on 100s–1000s of jobs per sequence . We intend our in silico approach to guide and complement – rather than replace – follow-up in vitro and in vivo studies . Here , we extract features and patterns from known ADPRT toxins and explain how they fit new toxins . We use in silico methods to probe structure , secretion , cell entry , activation , NAD+ substrate binding , intracellular target binding and reaction mechanism . A computer approach is fitting for several reasons . Such an environment is a safe way to study new toxins . Challenges in cloning , expressing , purifying and crystallizing often prevent in vitro characterization . Also , ADPRTs are abundant within bacterial genomes and researchers make the sequences available faster than we can conduct biochemical studies . New toxins might play a role in current outbreaks and are also excellent drug targets against antibiotic resistance . Our new study design expands the family by ∼15% ( from 36 to 42 toxins ) . Cell-based validation complements our in silico approach . We use Saccharomyces cerevisiae as a model host to study toxin effects . Increasingly , researchers are turning to yeast to study bacterial toxins . Yeast are easy to grow , have well-characterized genetics and are conserved with mammals in cellular processes including: DNA and RNA metabolism , signalling , cytoskeletal dynamics , vesicle trafficking , cell cycle control and programmed cell death [45]–[47] . We place the toxin genes under the control of a copper-inducible promoter to test putative toxins for ADP-ribosyltransferase activity in live cells [48] . A growth-defective phenotype clearly shows toxicity . Substitutions to catalytic signature residues confirms ADP-ribosyltransferase activity causes the toxicity . Indeed , pairing in silico and cell-based studies helps identify and characterize new ADPRT toxins .
We searched fold-recognition databases – including Pfam 24 . 0 [49] , Gene3D 9 . 1 . 0 [50] and SUPERFAMILY 1 . 73 [51] – using SCOP and CATH codes of known toxins . These strategies relate sequences with profiles . We also used a sensitive profile-profile based search strategy , HHsenser 2 . 13 . 5 [52] . We combined the results from our various searches and filtered them by successively applying exclusions to discover new ADPRT toxins . First , we had 2106 hits . We kept only bacterial hits ( lost 1222 ) from pathogens ( lost 445 ) that tested positive for secretion ( lost 95 ) , had the conserved ADPRT pattern ( lost 218 ) and had less than 50% identity to a known toxin ( lost 87 ) . This left 39 hits . We reduced them to 29 by clustering at the 50% identity level . We removed one more sequence on the basis of genetic context ( a hydrolase gene was next to the toxin gene , suggesting possible de-ADP-ribosylation reactions ) . This left 28 sequences . Of these , we found 15 from Pfam , Gene3D and HHsenser; eight from both Gene3D and HHsenser; four from HHsenser only; and one from both Pfam and Gene3D . We chose five of the 28 sequences to analyze more thoroughly . We also present our analysis of TccC5 , a toxin we previously proposed [4] that Lang et al . biochemically characterized during this writing [53] . We count 36 known ADPRT toxins ( see [4] for a recent table and note that researchers recently characterized several [54]–[57] ) . The six described in this writing bring the total to 42 distinct ADPRT toxins that generally have identity <50% unless the species or domain organization is different . We may want to remove the pattern constraint in the future and further expand the toxin pattern . Here , we prefer higher accuracy at the risk of removing some true ADPRT toxins from our list . Five of the six toxins described appear in a simple protein-protein BLAST search . But identity is typically low enough that many false hits appear as well . This makes the simple BLAST search ineffective . Randomly created sequences , for example , regularly return BLAST hits at ∼25% identity . ( For example , we tried 10 BLAST searches using 200-residue random sequences with average Swiss-Prot amino acid composition . We received top hits of average length 99 and having 29% identity to a natural protein . ) We ranked the toxin candidates by relevance signalled by ISI Web of Knowledge hits to the species name ( Table 1 ) . As well , we list the fold prediction strength given by J3D-jury and catalytic domain novelty suggested by sequence identity to the nearest known toxin . 3D-jury accepts models from various servers and makes pair-wise comparisons . Each pair gets a similarity score that equals the total number of Cα atom pairs within 3 . 5Å after overlap . The final score is the sum of the similarity scores divided by the number of pairs considered plus one . A higher J3D-jury implies a stronger prediction . The closest toxin relative to a newly predicted toxin indicates the new toxin's novelty and aids function prediction . Identity to a known toxin ranges from 25% to 60% . We show predictions for the toxins in Table 2 . Aligned sequences of known and new CT group toxins are critical to further studies ( Figures S1 and S2 in Text S1 ) . We removed positions with gaps and represented the alignment in LOGO format for the ExoS-like , C2-like , C3-like subgroups ( Figure 1 ) and the CT-PT-like subgroup ( Figure 2 ) . Also , we correlated critical residues with previous X-ray structures and function information . We used an alignment that contained all CT group toxins to build a phylogenetic tree that groups known and new toxins into subgroups , shown in Figure 3 . We use this tree to show relationships between the toxins independent of any specific evolutionary pathway . Such a pathway is difficult or impossible to deduce because of horizontal , rather than vertical , gene transfer . We did not include eukaryotic ARTs in our tree because they are not within this paper's scope . But , they often group well with C3-like toxins , and many eukaryotic PARPs group with the DT group toxins . Also , we calculated a pair-wise identity matrix ( Table S1 in Text S1 ) , revealing identity between known and new CT group toxins . We invite readers to skip to the species or toxin of most interest; each one is described independently . V . cholerae produces cholera and cholix toxins [4] . Chelt ( UniProt A2PU44 ) is , to our knowledge , the third ADPRT toxin identified in V . cholerae , the bacterium responsible for cholera outbreaks and food poisoning . The genome sequence of V . cholerae strain MZO-3 serogroup O37 , isolated from a patient in Bangladesh ( Heidelberg , J . and Sebastian , Y . , 2007 , Annotation of Vibrio cholerae MZO-3 , TIGR ) encodes Chelt . It is specific to this strain . Chelt GC content is 14% lower than the overall genome ( 34% vs . 48% ) ; also , a transposase gene immediately follows the Chelt gene , indicating horizontal gene transfer typical of the ADPRT toxins . Chelt is a 601-residue , 69 kDa protein . It has a secretion signal ( ∼1–18 ) , followed by toxin domain Ia ( ∼19–179 ) and Ib ( ∼180–240 ) and a presumed cell-binding domain II ( ∼241–601 ) ( Figures 4A and 5A ) . Chelt is unusual in that it has a second domain attached to the catalytic domain ( Figure S3 in Text S1 ) . Because the genome does not obviously encode a B-domain pentamer , domain II could fulfill that role . After secretion , Chelt likely uses it to bind to the cell surface . Domain II has significant structure similarity to Psathyrella velutina lectin ( PDB 2BWR; 15% identity; J3d-jury = 152; an easy target for the Local Meta-Threading-Server LOMETS , which provides this high-confidence match ) . Weaker similarities also exist to human integrin αVβ3 ( PDB 2VDR; 11% identity; an easy target for LOMETS , which provides this high-confidence match ) . Prokaryotic lectins allow differential eukaryotic cell recognition . Indeed bacterial lectins can mimic eukaryotic adhesion motifs [58] . Structurally , the domain is a seven-bladed β-propeller ( SCOP b . 69 . 8 , CATH 2 . 130 . 10 ) , with each blade containing seven four-stranded β-sheet motifs that meander . The lectin suggests a role in sugar and Ca2+ , or possibly Mg2+ , binding and perhaps even integrin mimicry . Chelt is reminiscent of ricin toxin from the castor bean . Ricin is a two-domain toxin that contains both a lectin for binding the cell-surface galactosyl residues for cell-entry and a second domain that causes cell death [59] . Domain I , the catalytic domain , is 60% identical to LT-A from Escherichia coli . This toxin clearly fits into the Gαs–targeting CT-PT-like subgroup because sequence identity to LT-A is so high . Fold recognition returned a match to LT-A ( PDB 1LT4 , J3D-jury = 178 ) and our model against this template was also high quality . The Chelt catalytic domain adopts an α+β ADP-ribosylation fold consisting of anti-parallel β-sheets and having separate α and β regions . Chelt must likely be activated by reduction of a disulfide bond between Chelt C205 and C220; cleavage at or near I215 ( details are unclear due to a four amino acid deletion compared to LT-A between H214 and I215 ) ; and interaction with an ADP-ribosylating factor , perhaps ARF3 , in the Chelt regions ∼45–57 , ∼109–113 , ∼134–141 and ∼167–182 ( Figure S3 in Text S1 ) . We propose a likely mode of NAD+ binding , target binding and ADP-ribosylation based on alignment data and our modeling experiments . Once activated , Chelt binds NAD+ through hydrogen bonds , hydrophobic interactions and aromatic interactions ( Figure 6A , Figure S4 in Text S1 , Table 3 ) . We propose these H-bonds: Y41 binds to adenine , S28 binds to A-ribose , R43 binds to A-phosphate , R25 binds to A- or N-phosphate , E130 binds to N-ribose and A26 binds to nicotinamide . Chelt recognizes Gαs using the knob ( ∼66–71 ) , the α3 helical region ( ∼82–99 ) and the ARTT loop ( ∼104–129 ) ( Table 4 ) . The ARTT loop might plastically rearrange between the in and out conformation during this process . Anchor residues S123 and Q127 in the second part of the loop may act as hinges to reposition H125 to interact with Gαs . We propose an SN1 alleviated-strain mechanism ( Figure 7 ) . First , E130 H-bonds to the N-ribose while phosphate electrostatic interactions hold the NAD+ in a conformation that favors oxacarbenium ion formation . The reaction's progress is unclear . T71 might induce a rotation about the O5D-PN bond of the oxacarbenium ion to reduce the nucleophile-electrophile distance . A Gαs Glu or Asp stabilizes N-ribose , E128 stabilizes Gαs R201 and Gαs R201 attacks the oxacarbenium ion . Several residues hold the active site in place including: Chelt S79 , which H-bonds to E130; T80 , which stiffens the active site through H-bonding to a nearby β-sheet and T81 , which orients the ARTT loop and E128 . Hydrophobic interactions with NAD+ involve D27 , R29 , P42 , I90 , I94 and L95 . Also , H62 stabilizes E130 . Cell-based toxin expression in yeast , driven by the copper-inducible CUP1 promoter , shows cell death in the presence of the wild-type toxin . We observed mild growth restoration with the E128A mutant , dramatic growth restoration with the E130A mutant and near-complete growth restoration with the E128A/E130A double mutant ( Figure 8A ) . The wildtype growth-defective phenotype clearly shows Chelt toxicity . Substitutions to E128 and E130 confirm that this toxicity is because of Chelt ADP-ribosyltransferase activity . Researches may modify Chelt in the future with the E128A and E130A substitutions – or produce recombinant forms including domain II only – to make vaccines similar to the commercial Dukoral [60] . Certhrax ( UniProt Q4MV79 ) is encoded in B . cereus G9241 . ( A slightly larger relative exists in another B . cereus strain . ) Most B . cereus strains are harmless or cause foodborne illness , but researchers have implicated this strain in several severe pneumonia cases [61]–[63] . Certhrax , a 476-residue , 55 kDa protein , is the first anthrax-related ADPRT toxin to our knowledge . It is 31% identical to lethal factor from Bacillus anthracis . The closest fold recognition match is to anthrax toxin lethal factor ( LF , PDB 1J7N; J3D-jury = 239 , a high score reflecting a two-domain match ) . So we modeled Certhrax against LF . Certhrax has two domains: domain I ( ∼1–241 ) presumed to bind PA and domain II ( ∼242–476 ) is the toxin domain ( Figures 4B and 5B ) . B . cereus cells secrete this protein non-classically . Certhrax likely behaves similarly to LF in cell entry because of similarities in domain I , which is likely responsible for PA-binding . We describe a supposed model of Certhrax here using LF as a template [64] . Under harsh conditions , B . cereus forms spores that humans inhale into lung alveoli . Spores that escape from macrophages enter the lymph system where B . cereus germinates . Here , B . cereus produces protective antigen ( PA , UniProt Q4MV80 ) that may bind Certhrax and edema factor ( UniProt Q4MKW0 ) . Both Certhrax and LF have a PA binding domain; sequence identity over this domain is 36% , within the safe zone of homology . But , Certhrax lacks the catalytic zinc metalloprotease domain of LF that proteolyzes mitogen activated protein kinase kinase ( MAPKK or MEK ) . It contains a functional ADPRT domain instead of the vestigial ADPRT domain of LF ( Figure S5 in Text S1 ) . PA likely binds to ANTXR1/2 or LRP6 receptor . Furin proteolyzes PA so a PA heptamer can form . Certhrax and edema factor bind the PA heptamer and are translocated into the cell in a clathrin-coated pit . Low pH in the endosome causes a pore to form through which Certhrax and EF travel and enter the cytosol [64] . Domain II matches to iota toxin ( PDB 1GIQ , J3D-jury = 143 ) . Fold recognition and phylogenetic analysis suggest similarities to C3-like toxins . We propose a likely mode of NAD+ binding , target binding and ADP-ribosylation based on alignment data and our modeling experiments ( Figure 6B , Figure S4 in Text S1 , Table 3 ) . These H-bonds are likely: Q382 and N384 may bind to adenine , S344 binds to A-ribose , N288 and R292 bind to A-phosphate , R341 binds to A- or N-phosphate , T280 and E431 bind to N-ribose and R342 binds to nicotinamide . Active site residue Y398 in the flexible PN loop locks nicotinamide in the enzyme cleft during the reaction . Certhrax likely recognizes its target through the region B active site loop ( ∼295–314 ) , the PN loop ( ∼390–402 ) and the ARTT loop ( ∼420–430 ) ( Table 4 ) . The ARTT loop might plastically rearrange between the in and out conformation during target recognition . The second part may hinge on anchor residues S424 and Q429 to reposition Y426 to interact with the target substrate . We propose the reaction follows an SN1 alleviated-strain mechanism ( Figure 7 ) . First , E431 H-bonds to the N-ribose while phosphate electrostatic interactions hold the NAD+ in a conformation that favors oxacarbenium ion formation . Then Y284 induces a rotation about O5D-PN bond of the oxacarbenium ion that reduces the nucleophile-electrophile distance . Finally , a target Glu or Asp stabilizes the N-ribose , Q429 stabilizes the target Asn or Gln and the target Asn or Gln attacks the oxacarbenium ion . Several residues hold the active site in place including: S387 , which H-bonds to E431; T388 , which stiffens the active site through H-bonding to a nearby β-sheet and S389 , which orients the ARTT loop and Q429 . Another conserved residue is Y279 , which may participate in the reaction . Toxin gene expression in yeast , driven by the CUP1 promoter , shows cell death in the presence of the wild-type toxin . We observed mild growth restoration with the Q429A and E431A mutants and near-complete growth restoration with the Q429A/E431A double mutant ( Figure 8B ) . The wildtype growth-defective phenotype clearly suggests Certhrax toxicity . Substitutions to Q429 and E431 confirm that this toxicity is because of Certhrax ADP-ribosyltransferase activity . Researchers may eventually modify Certhrax with the Q429A and E431A substitutions – or produce recombinant forms of the toxin that include only the PA-binding domain I – to create vaccines similar to Biothrax that protects against B . antracis effects [65] . Mav toxin ( UniProt A0QLI5 ) from M . avium strain 104 is a predicted ADPRT with possible relevance to AIDS patients who face a high risk of M . avium infections [66] . ( Slightly larger relatives exist in M . avium subsp . paratuberculosis and M . avium subsp . avium ATCC 25291 . ) M . avium is both an environmental microbe and opportunistic pathogen causing chronic , pulmonary infections in immune-compromised individuals . Mav toxin is an 825-residue , 83 kDa protein with four putative domains: an ESAT6-like domain I ( ∼1–96 ) , a predicted helical pore-forming domain II ( ∼97–439 ) , a largely disordered domain III ( ∼440–674 ) and the toxin domain IV ( ∼675–825 ) ( Figures 4C and 5C ) . Domain I suggests secretion through the ESX ( type VII ) secretion system . This matches the non-classical secretion result . Fold recognition matches residues 1–95 to the 6 kDa early secreted antigenic target ( ESAT-6; PDB 1WA8; J3d-jury = 65; 16% identity ) . Virulent mycobacteria need the ESX secretion system for pathogenesis: ESX-1 deletion weakens virulence in M . tuberculosis , M . bovis and M . marinum [67] . ESAT-6 forms a heterodimer with the 10 kDa culture filtrate protein ( CFP-10 ) . Researchers believe the tight dimer binds an Rv3871-like ATPase for transfer to the Rv3877-like transmembrane pore through an Rv3870-like protein [68] . Domain II is α-helical , especially from 134–348 . It might be a multi-helical bundle of short and long helices poised to form pores for target cell entry . Fold recognition matches are to the soluble domain of bacterial chemoreceptors ( PDB 3G67 , J3d-jury = 93 ) , a tropomyosin leucine zipper ( PDB 2EFR , J3d-jury = 78 ) and spectrin-like repeats ( PDB 1QUU , J3d-jury = 76 ) . Domain III has slight propensity for forming β-sheets; but it is disordered . Its role is unknown , but it might recognize and bind cell-surface receptors . Combining domains II and III we found matches to the Cry insecticidal α-pore-forming toxins ( a hard target for LOMETS , which provides a high-confidence match to PDB 1CIY ) . Domain IV is the catalytic domain . Fold recognition suggests matches to Art2 . 2 ( PDB 1GXY , J3d-jury = 126 ) . Mav , compared with iota toxin , has an 18-residue deletion after region 1 between P735 and A736 . Also , and possibly affecting targeting , it has a two-residue PN-loop insertion ( S765–S766 ) . We propose a likely mode of NAD+ binding , target binding and ADP-ribosylation based on alignment data and our modeling experiments . NAD+ binding ( Figure 6C , Figure S4 in Text S1 , Table 3 ) likely involves these H-bonds: E750 binds to adenine , N733 and possibly T732 bind to A-ribose , N695 and R699 bind to A-phosphate , R730 binds to A- or N-phosphate , T687 and E795 bind to N-ribose and G731 binds to nicotinamide . Active site residue F768 on the flexible PN loop locks the nicotinamide in the enzyme cleft during the reaction . Mav toxin recognizes its target using the region B active site loop ( ∼701–705 ) , the PN loop ( ∼758–771 ) and the ARTT loop ( ∼784–794 ) ( Table 4 ) . The ARTT loop might plastically rearrange between the in and out conformation during this process . The first part of the ARTT loop , anchored between V784 and V787 , is likely less flexible than the second part . The second part hinges on S788 and E793 to reposition Y790 to interact with the target substrate . We propose the reaction follows an SN1 alleviated-strain mechanism ( Figure 7 ) . First , E795 H-bonds to the N-ribose while phosphate electrostatic interactions hold the NAD+ in a conformation that favors oxacarbenium ion formation . Then Y691 induces a rotation about O5D-PN bond of NAD that reduces the nucleophile-electrophile distance . Finally , a target Glu or Asp stabilized the N-ribose , E793 stabilizes the target Arg and the target Arg attacks the oxacarbenium ion . Several residues hold the active site in place including: S755 , which H-bonds to E795; T756 , which stiffens the active site through H-bonding to a nearby β-sheet and S757 , which orients the ARTT loop and E793 . Also , Y686 stabilizes E795 . Neighbourhood and co-occurrence evidence suggest Mav may interact with the exported repetitive protein ( UniProt A0Q9B3 ) – suggested as a virulence factor in Mycobacteria [69] – and several putative uncharacterized proteins . Cloning problems frustrated cell-based characterization in yeast . As well , we have several concerns about this prediction: a characteristic WXG motif is lacking in domain I and the whole protein is unusually long for ESX-1 secretion . Perhaps Mav toxin uses a variant of the ESX-1 system ( ESX-2 to ESX-5 ) . Also , the genetic context suggests a haloacid dehalogenase-like hydrolase is encoded nearby , making de-ribosylation reactions a concern . But , we believe this putative toxin is worth presenting despite these issues because of its potential health implications . EFV toxin ( UniProt Q838U8 ) is a medically relevant ADPRT candidate from a vancomycin-resistant E . faecalis strain , V583 [70] . This strain produces cytolysin toxin [71] and causes urinary infection , bacteremia and endocarditis [72] . A slightly smaller relative exists in Enterococcus faecalis CH188 . EFV toxin itself is a 487-residue , 56 kDa protein and has a needle-like helical domain I ( ∼1–309 ) and catalytic domain II ( ∼310–487 ) ( Figures 4D and 5D ) . The toxin is non-classically secreted ( i . e . , without a signal peptide ) . A type IV secretion system has been identified in E . faecalis [73] , but it is unclear if it mediates EFV toxin secretion . Genetic context suggests that EFV toxin may more likely travel through a phage infection conduit to target cells . Neighbourhood , gene fusion and co-occurrence evidence suggest it may interact with portal proteins ( UniProt Q838U9 and Q833E4 ) , a scaffold protein ( Q838U5 ) , a major tail protein ( Q835T7 ) , a Cro/CI family transcriptional regulator ( Q835K8 ) and several putative uncharacterized proteins . The phage origin makes it unclear whether EFV toxin acts mainly against bacterial or eukaryotic targets . Domain I bears large sequence similarity to phage minor head region from 147–268 that suggests a possible phage origin . The phage head match is reminiscent of the dual role of Alt in bacteriophage T4 as both a phage head structure component and a RNA-polymerase targeting ADPRT [74] . Fold recognition on domain I suggests matches to spectrin ( PDB 1U4Q , J3d-jury = 49; a hard target for LOMETS , which provides this high-confidence match ) and weaker matches to the pore-forming domain of colicin s4 ( PDB 3FEW , J3d-jury = 42 ) . Also genetic context suggests similarities to the bacteriophage P22 needle implicated in cell-envelope penetration [75] . Domain II is 25% identical to Bacillus thuringiensis VIP2 over 166 residues . EFV toxin has C2-like character based on its phylogenetic branching . It also has a region 3 EXE sequence pattern that suggests an Arg target . Fold recognition suggests that its closest structure match is to C2-I ( PDB 2J3Z , J3D-jury = 158 ) . The efforts of the Midwest Center for Structural Genomics have failed to produce a structure . We propose a likely mode of NAD+ binding , target binding and ADP-ribosylation based on alignment data and our modeling experiments ( Figure 6D , Figure S4 in Text S1 , Table 3 ) . These H-bonds are likely: S397 , N399 or E400 binds to A-ribose , N354 and R358 bind to A-phosphate , R394 binds to A- or N-phosphate , T346 and E463 bind to N-ribose and G395 binds to nicotinamide . Active site residue F426 in the PN loop locks the nicotinamide in the enzyme cleft during the reaction . EFV toxin recognizes its target using the region B active site loop ( ∼361–370 ) , the PN loop ( ∼418–436 ) and the ARTT loop ( ∼452–462 ) ( Table 4 ) . The ARTT loop might plastically rearrange between the in and out conformation during this process , hinging on S456 and E461 . Compared with iota toxin , and possibly influencing target recognition , EFV toxin has a 22-residue deletion in region F ( between regions 1 and 2 ) between A403 and I404 . Also possibly influencing targeting , EFV toxin has a six-residue PN loop insertion ( E424–F429 ) . We propose the reaction follows an SN1 alleviated-strain mechanism ( Figure 7 ) . First , E463 H-bonds to the N-ribose while phosphate electrostatic interactions hold the NAD+ in a conformation that favors oxacarbenium ion formation . Then F350 likely induces a rotation about the O5D-PN bond of the oxacarbenium ion bond to reduce the nucleophile-electrophile distance . Finally , a target Glu or Asp stabilizes the N-ribose , E461 stabilizes the target Arg which attacks the oxacarbenium ion . Several residues hold the active site in place including: S415 which H-bonds to E463; T416 , which stiffens the active site through H-bonds to a nearby β-sheet and S417 , which orients the ARTT loop and E461 . Also , Y345 stabilizes E463 . Other potential active site residues include T346 , E412 and F426 . EFV toxin expression in yeast , driven by the CUP1 promoter , shows cell death in the presence of the wild-type toxin . We observed dramatic restoration growth with the E461A and E463A mutants and near-complete growth restoration with the E461A/E463A double mutant ( Figure 8C ) . The wildtype growth-defective phenotype clearly shows EFV toxin toxicity . Substitutions to E461 and E463 confirm that this toxicity is because of EFV toxin ADP-ribosyltransferase activity . TccC5 ( UniProt Q7N7Y7 ) is an ADPRT from P . luminescens TT01 that we previously suggested as an ADPRT toxin [4] , which has gained significant attention recently [53] . Is distinct from the recently reported Photox [56] , but a close relative also exists in the W14 strain . TccC5 is 938-residue , 105 kDa protein and has four domains: domain I ( ∼1–341 ) , domain II ( ∼342–675 ) , domain III ( ∼676–738 ) and domain IV ( ∼739–938 ) ( Figures 4E and 5E ) . This toxin is non-classically secreted . Fold-recognition matches to TccC5 are to various tandem seven-bladed β-propellers , including the actin-interacting protein ( PDB 1NR0; J3D-jury = 71 ) and the Sro7 exocytosis regulator ( PDB 2OAJ , a high-confidence LOMETS match ) . These proteins are WD40 repeat-containing proteins ( SCOP b . 69 . 4 , CATH 2 . 130 . 10 . 10 ) . Also , we found matches to several tandem seven-bladed β-propeller xyloglucanase structures ( PDB IDs 3A0F , 2EBS , 2CN2; SCOP b . 69 . 13; CATH 2 . 130 . 10 . 140 ) that hydrolyze polysaccharides . Fold recognition on domain I , a hard target , produces matches to various β-propellers such as βγ-dimer of the heterotrimeric G-protein transducin ( PDB 1TBG , LOMETS high-confidence match ) , oxidoreductases ( PDB 1FWX , J3d-jury = 123 ) , outer surface protein OspA ( PDB 1FJ1 , J3d-jury = 83 , LOMETS high-confidence match to 2FJK ) , Tyr-Val-Thr-Asn ( YVTN ) domain from an archaeal surface layer protein ( PDB 1L0Q , a high-confidence LOMETS match ) , lyases ( e . g . , streptogramin B lyase , PDB 2QC5 , a LOMETS high-confidence match; and virginiamycin B lyase , PDB 2Z2P , J3d-jury = 51 ) , among others . Function prediction suggests domain I contains two YD repeats possibly involved in binding carbohydrates and heparin . Also , domain I contains a lipocalin pattern , hinting at a connection to small-molecule transporters . Fold recognition on domain II , also a hard target , shows there may be a second β-propeller after the first . Matches are to various β-propellers including OspA , YVTN from an archaeal surface layer protein and the extracellular domain of LDL receptor ( PDB 1N7D , a high-confidence LOMETS match ) , among others . The C-terminal end of domain II appears to have recombination hot spot ( Rhs ) repeats employed in other secreted bacterial insecticidal toxins and eukaryotic intercellular signalling proteins , and often involved in ligand binding . Rhs suggests horizontal transfer; it is related to YD repeats and also often contains VgrG , a type VI secretion protein . β-propellers are structurally conserved but functionally diverse , so it is difficult to pinpoint exact functions for domains I and II . While the exact role of these domains is unclear , a likely role is gaining cell entry . Domain III seems helical with unknown function . TccC5 domain IV best matches SpvB but identity is only 25% over the toxin core , making TccC5 among the most novel toxins discussed here . Fold recognition results suggest that TccC5 is similar to C3bot2 ( PDB 1R45 , J3d-jury = 92 ) throughout the catalytic domain . Recently , Lang et al . identified the cellular target as RhoA Q63 [53] . We propose a likely mode of NAD+ binding , target binding and ADP-ribosylation based on alignment data and our modeling experiments . TccC5 binds NAD+ through hydrogen bonds , hydrophobic interactions and aromatic interactions ( Figure 6E , Figure S4 in Text S1 , Table 3 ) . We propose these H-bonds: T777 binds to A-ribose , N742 and R746 bind to A-phosphate , R774 binds to A- or N-phosphate , R829 may bind N-phosphate , T735 and E886 bind to N-ribose and V775 binds to nicotinamide . Active site residue F819 in the flexible PN loop locks the nicotinamide in the enzyme cleft during the reaction . TccC5 recognizes RhoA using the region B active site loop ( ∼748–751 ) , the PN loop ( ∼812–828 ) and the ARTT loop ( ∼861–885 ) ( Table 4 ) . The ARTT loop might plastically rearrange between the in and out conformation during this process . Compared to SpvB , TccC5 has several key differences that may influence targeting including: a 30 amino acid deletion in region B between I750 and T751 , an eight-residue insertion in the PN loop ( F819–S826 ) and a 32-residue insertion in the ARTT loop between A854 and E885 . Other variations include a five-residue insertion between I779 and K783 and two deletions that follow the ARTT loop , namely , three residues between R901 and H902 and two residues between I914 and K915 . We propose the reaction follows an SN1 alleviated-strain mechanism ( Figure 7 ) . First , E886 H-bonds to the N-ribose while phosphate electrostatic interactions hold the NAD+ in a conformation that favors oxacarbenium ion formation . The reaction's progress is unclear . S738 might induce a rotation about the O5D-PN bond of the oxacarbenium ion to reduce the nucleophile-electrophile distance . A RhoA Glu or Asp likely stabilizes N-ribose , TccC5 Q884 likely stabilizes RhoA Asp , and finally RhoA Q63 attacks the oxacarbenium ion . Several residues hold the active site in place including: S809 , which H-bonds to E886; T810 , which stiffens the active site through H-bonding to a nearby β-sheet and S811 , which orients the ARTT loop and Q884 . Also , Y734 stabilizes E886 . Co-occurrence , neighbourhood , gene fusion and recent evidence [53] , suggest that TccC5 exists as part of a toxin complex with the TcdA1 toxin and TcdB2 potentiator . Full activity depends on these partners [76] . TccC5 expression in yeast , driven by the CUP1 promoter , shows cell death in the presence of the wild-type toxin . We observed mild growth restoration with the Q884A mutant , dramatic growth restoration with the E886A mutant and near-complete growth restoration with the Q884A/E886A double mutant ( Figure 8D ) . The wildtype growth-defective phenotype clearly shows TccC5 toxicity . Substitutions to Q884 and E886 confirm that this toxicity is because of TccC5 ADP-ribosyltransferase activity . Vis ( UniProt A3UNN4 ) is an ADPRT from a known pathogen , V . splendidus 12B01 , which causes vibriosis and afflicts oysters . Similar proteins exist in Vibrio harveyi strains HY01 and BB120 , Photobacterium sp SKA34 and Photobacterium angustum S14 . Vis toxin is 30% identical to VopT from Vibrio parahaemolyticus . This single-domain toxin has 249 residues and is 28 kDa . It harbors a secretion signal peptide with a cleavage site between position 18 and 19 ( Figures 4F and 5F ) . Fold recognition matches it to iota toxin ( PDB 1GIQ , J3D-jury = 135 ) . Vis entry into target cells is unclear . It may travel through a transporter , be aided by other pore-forming toxins or be directly released into the cytosol after V . splendidus invasion . We propose a likely mode of NAD+ binding , target binding and ADP-ribosylation based on alignment data and our modeling experiments . NAD+ binding ( Figure 6F , Figure S4 in Text S1 , Table 3 ) likely involves these H-bonds: E137 binds to adenine , W120 may bind to A-ribose , N76 and R80 bind to A-phosphate , R117 binds to A- or N-phosphate , S68 and E191 bind to N-ribose and G118 binds to nicotinamide . Active site residue F153 in the flexible PN loop locks the nicotinamide in the enzyme cleft during the reaction . Vis recognizes its target using the region B active site loop ( ∼82–91 ) , the PN loop ( ∼145–164 ) and the ARTT loop ( ∼180–190 ) ( Table 4 ) . Vis has a 24-residue deletion after the region 1 Arg between K122 and L123 . Also , and possibly affecting targeting , it has a four-residue region B insertion between V89-A92 and an eight-residue insertion in the PN loop between E148 and V155 . The ARTT loop might plastically rearrange between the in and out conformation during target recognition . The first part of the ARTT loop is anchored between hydrophobic residues I180 and L183 and is likely less flexible than the second part . This second part , which hinges on S184 and E189 , likely repositions Y186 to interact with the target substrate . We propose the reaction follows an SN1 alleviated-strain mechanism ( Figure 7 ) . First , E191 H-bonds to the N-ribose while phosphate electrostatic interactions hold the NAD+ in a conformation that favors oxacarbenium ion formation . Then Y72 induces a rotation about O5D-PN bond of the oxacarbenium ion that reduces the nucleophile-electrophile distance . Finally , a target Glu or Asp stabilizes the N-ribose , E189 stabilizes the target Arg or Cys which attacks the oxacarbenium ion . Several residues hold the active site in place including: S142 , which H-bonds to E191; T143 , which stiffens the active site through H-bonds to a nearby β-sheet and S144 , which orients the ARTT loop and E189 . Also , Y76 stabilizes E188 . F153 promotes NAD+ binding and glycohydrolase activity . F67 is another conserved residue possibly involved in the reaction . Vis toxin expression in yeast , driven by the CUP1 promoter , shows cell death in the presence of the wild-type toxin . We observed mild growth restoration with the E189A and E191A mutants and dramatic growth restoration with the E189A/E191A double mutant ( Figure 8E ) . The wildtype growth-defective phenotype clearly suggests Vis toxicity . Substitutions to E189 and E191 confirm that this toxicity is because of Vis toxin ADP-ribosyltransferase activity . We have combined computer results with cell-based data to improve toxin discovery and characterization . The six new toxins described here are a significant addition to the list of known ADPRTs . Interested readers may refer to Text S1 for further discussion of trends in structure and function . Future toxin discoveries will involve not only new entries to public sequence and structure databases , but also updates to the search pattern and perhaps even new folds . For example , Johnson et al . recently showed the region 2 STS motif is not strictly needed in an M . penetrans ADPRT [55] . Also , the PARP10 ADPRT does not need the hallmark “catalytic Glu” because it uses a substrate-assisted mechanism [77] . AexU from Aeromonas hydrophila [78] , [79] may reveal a new ADP-ribosylation fold: our preliminary fold-recognition tests suggest it does not adopt the typical ADPRT fold . We must do much work to characterize the new toxins in vitro . One challenge is developing a way to reliably overcome expression , purification and solubility problems , which seem typical in this family . If we can overcome these problems , we may pinpoint structure details through X-ray crystallography in cases where the toxin is amenable such techniques . Finding intracellular targets will also aid in elucidating functional details . Time-resolved crystallography , NMR spectroscopy and QM/MM simulations may one day further reveal reaction dynamics [80] . Our efforts in cell-based characterization may involve more complete in vivo characterization where we give purified toxin to suitable target cells or model organisms . Applying knowledge of these new toxins to improve human health and agricultural production is a large-scale but worthwhile challenge .
We used remote homolog detection strategies to find ADPRTs within the set of all known sequences . Authors have reviewed [81] , [82] and benchmarked [83] these strategies . Often the only way to find remote homologs to a query sequence is through structure links , so structure prediction and remote homolog detection often rely on the same strategies . One effective strategy is to pair structure prediction with matches to consensus patterns . Russell et al . described the leading structure classification databases [84] . We used the Structural Classification of Proteins ( SCOP ) [85] and Class Architecture Topology Homology ( CATH ) [86] databases . We extracted structure codes for the ADPRT family from these databases for further searches . We used these SCOP codes: d . 166 . 1 . 1 ( mART ) , d . 166 . 1 . 2 ( PARPs ) , d . 166 . 1 . 3 ( ARTs ) , d . 166 . 1 . 4 ( AvrPphF ORF2 , a type III effector ) , d . 166 . 1 . 5 ( Tpt1/KptA ) , d . 166 . 1 . 6 ( BC2332-like ) and d . 166 . 1 . 7 ( CC0527-like ) . We used these CATH codes: 3 . 90 . 175 . 10 ( DT Group mART ) , 3 . 90 . 176 . 10 ( C2- and C3-like mARTs , ARTs ) , 3 . 90 . 210 . 10 ( CT-PT-like mARTs ) and 3 . 90 . 182 . 10 ( Anthrax_PA-like ) . Teichmann et al . described several fold-recognition databases [87] . To get a putative ADPRT toxin list , we searched the structure classification codes for known ADPRTs against such databases , including Gene3D [50] and SUPERFAMILY [51] . We filtered the resulting sequences for ADPRT toxins by keeping only bacterial hits using NCBI taxon IDs , keeping only hits from pathogens using gene ontology data and the GOLD database [1] , keeping only hits that tested positive for secretion using SignalP 3 . 0 or Secretome P 2 . 0 and keeping only hits that had the conserved ADPRT pattern using ScanProsite [88] with this regular expression: [YFL]-R-X ( 27 , 60 ) -[YF]-X-S-T-[SQT]-X ( 32 , 78 ) -[QE]-X-E . We formed this pattern strictly using known 3D structures in 3dLOGO and changing the resulting regular expression to ensure that it captured known ADPRT toxins in ScanProsite searches . We kept only hits with less than 50% identity to a known toxin and further reduced the list by clustering at the 50% identity level . We checked genetic context for hydrolases using Entrez Gene [89] and removed sequences where one was encoded nearby . ( Ribosylhydrolases and ribosylglycohydrolases can de-ribosylate proteins . Hydrolases may suggest a regulatory cycle or toxin-antitoxin selfish genetic entities [90] . ) We selected several interesting examples to characterize and discuss . We ranked the final toxin list in order of decreasing ISI Web of Knowledge hits to the species name . For both the C2-C3-like toxins and the CT-PT-like toxins , we aligned known and new toxins using 3D-Coffee [91] , we visualized the alignment using ESPript [92] , we curated it to remove positions with gaps using Phylogeny . fr [93] and converted it to LOGO format using WebLOGO [94] . We produced a percent identity matrix using ClustalX [95] to reveal the relationships between the new and known ADPRT toxins . We curated an alignment containing all ADPRT toxins by removing positions with gaps to prepare it for phylogenetic analysis by Bayesian inference with the MrBayes algorithm [96] . The likelihood model included six substitution types with invariable and gamma rate variation across sites . Markov chain Monte Carlo parameters included 10 , 000 generations , sampling a tree every 10 generations . We discarded the first 250 trees sampled . Fisher reviewed fold recognition servers [97] . We sent the putative ADPRT toxins to fold-recognition meta-servers including: 3D-jury [98] , Pcons [99] , Genesilico [100] , LOMETS [101] and Atome2 [102] . Sequences with top hits to ADPRT toxins or ADPRT-related structures ( e . g . ART , PARP , LF , etc . ) remained on the list . We recorded the J3D-jury and structure match for each sequence . J3D-jury> = 40 is usually correct , but ideally we like it to be 100 or more for strong structure matches . We reassessed sequences showing no match to ADPRT-like proteins by using sliding-window fold–recognition ( see structure prediction: domain organization below ) . If no match to an ADPRT-related structure appeared , we removed them from the list . We checked ScanProsite matches against fold-recognition results , and adjusted them to ensure that we correctly identified the region 1 Arg , region 2 “STS” motif and region 3 ARTT motif . The CASP7 competition compared domain prediction tools [103] . We present domain assignments and boundaries that often differ from data in public domain databases or are unavailable . We used top performer DOMAC ( Accurate , Hybrid Protein Domain Prediction Server ) . It uses both template-based and ab initio methods and uses a PSI-BLAST generated profile to find templates . For significant matches it uses MODELLER for modeling and the protein domain parser ( PDP ) for domain parsing . If it does not find matches , it relies on neural networks or support vector machines ( SVMs ) [104] . We manually adjusted these results to match the sliding-window fold recognition data , testing sliding windows of 50 , 75 , 100 , 150 , 200 , 250 , 300 , 350 etc . amino acids on the fold-recognition meta servers to identify boundaries and fold type for the non-toxic domains . We mapped PDB hits to SCOP and CATH codes and interpreted the results to understand cell-entry strategies [105] . Nayeem et al . compared modeling software [106] . Prime works best for modeling in low sequence identity cases . But Modeller [107] is widely used , updated often and freely available , so we chose it for our work . For each candidate ADPRT , we used the alignments in Figures S1 and S2 in Text S1 and 3D-jury to select a suitable input alignment of the new toxin against a known template . We inspected the input alignments to ensure that we had properly aligned regions B , 1 , 2 and 3 . We modeled NAD+-bound structures using MODELLER and alignments to an NAD+-bound template: C3bot1 ( PDB 2A9K ) [108] , Iota toxin ( PDB 1GIQ ) [33] , SpvB ( PDB 2GWL ) [109] , EDIN-B ( PDB 1OJZ ) [110] , CdtA ( PDB 2WN7 ) [111] , Art2 . 2 ( PDB 1OG3 ) [112] , Vip2 ( PDB 1QS2 ) [39] and cholera toxin ( PDB 2A5F ) [113] . Except for Chelt , we used all templates to find invariant features between the resulting models and interpret the new toxins based on consistent NAD+-binding patterns . We modeled full-length ADPRT structures using I-TASSER , the top-ranked program for fully-automated structure prediction in CASP7 . It combines folds and supersecondary structures selected from the PDB with ab initio loop models . These elements are reassembled and refined to produce the final model . When I-TASSER failed to produce a result matching the sliding-window fold recognition data ( four cases ) , we selected suitable templates from this fold recognition data . We docked the templates using HADDOCK [114] and used them as MODELLER input . Where appropriate , we used VTFM and MD to optimize the models and repeated the modeling cycle at least two times to achieve an adequate objective function ( >1×106 ) . We refined loops automatically after model building and ranked them by Discrete Optimized Protein Energy ( DOPE ) statistical potentials to find the top model . We visualized the models using PyMol . Laskowski et al . reviewed model quality assessment programs ( MQAPs ) [115] . We assessed the ADPRT models using MetaMQAPII , a meta-server that considers results from VERIFY3D , PROSA , BALA , ANOLEA , PROVE , TUNE , REFINER and PROQRES [116] . We also gathered model data using MolProbity [117] . We assessed NAD+ binding using crystal structures solved with NAD+ in the active site: C3bot1 ( PDB 2A9K ) [108] , Iota toxin ( PDB 1GIQ ) [33] , SpvB ( PDB 2GWL ) [109] , EDIN-B ( PDB 1OJZ ) [110] , CdtA ( PDB 2WN7 ) [111] , Art2 . 2 ( PDB 1OG3 ) [112] , Vip2 ( PDB 1QS2 ) [39] and cholera toxin ( PDB 2A5F ) [113] . We used LigPlot [118] on the PDBsum server [119] ) to visualize the usual interactions in ADPRT NAD+ binding . We used the 3dLOGO [120] software to reveal equivalent positions in these structures . We used conserved residues from the alignment involved in typical NAD+ binding interactions in the known ADPRTs to identify the equivalent residues in the new ADPRTs . We also analyzed our NAD+-bound models and compared the ADPRTs modeled directly against the NAD+-bound templates using Modeller [107] . We developed the ADPRT toxin reaction mechanism for the new toxins using the SN1 alleviated-strain model , first proposed by Tsuge et al . , that many believe is widely relevant to the entire family [34] . As for NAD+ binding we used 3DLOGO [120] to reveal equivalent positions in these structures: C3bot1 ( PDB 2A9K ) , Iota toxin ( PDB 1GIQ ) , SpvB ( PDB 2GWL ) , EDIN-B ( PDB 1OJZ ) , Art2 . 2 ( PDB 1OG3 ) , Vip2 ( PDB 1QS2 ) and cholera toxin ( PDB 2A5F ) . We also matched residues involved in the iota toxin mechanism to residues in SpvB , EDIN-B and C3bot1 and to the new toxins using 3D-jury results . We exploited conservation of the hallmark catalytic Glu for step 1 , a conserved aromatic ( usually Tyr , but sometimes Phe ) for step 2 and the secondary Glu or Gln for step 3 . We also used the rule that region 3 [QE]XE pattern appears as EXE in ADPRTs that ribosylate Arg and as QXE in ADPRTs that ribosylate Asn , Gln or Cys . We cultured Saccharomyces cerevisiae strain W303 ( MATa , his3 , ade2 , leu2 , trp1 , ura3 , can1 ) on yeast-peptone-dextrose or synthetic dextrose ( SD ) drop-out medium . We performed the yeast growth-defective phenotypic test and quantified growth as previously described [48] . | Computer tools helped us uncover and understand potent protein toxins that empower bacterial pathogens against plants , animals and man . These toxins are potential drug targets and researchers can use them to make vaccines . New toxin knowledge aids the long-term goal of finding alternatives to antibiotics , to which pathogens are becoming more resistant . The toxins share similar structure despite low sequence identity , so our search links sequence and structure features . We present a ranked list and computational characterization of six new toxins combined with cell-based tests . | [
"Abstract",
"Introduction",
"Results/Discussion",
"Methods"
] | [
"computational",
"biology/sequence",
"motif",
"analysis",
"computational",
"biology/protein",
"homology",
"detection",
"computational",
"biology/protein",
"structure",
"prediction",
"computational",
"biology/macromolecular",
"sequence",
"analysis",
"biochemistry/biomacromolecule-liga... | 2010 | Cholera- and Anthrax-Like Toxins Are among Several New ADP-Ribosyltransferases |
A general paradigm to understand protein function is to look at properties of isolated well conserved domains , such as SH3 or PDZ domains . While common features of domain families are well understood , the role of subtle differences among members of these families is less clear . Here , molecular dynamics simulations indicate that the binding mechanism in PSD95-PDZ3 is critically regulated via interactions outside the canonical binding site , involving both the poorly conserved loop and an extra-domain helix . Using the CRIPT peptide as a prototypical ligand , our simulations suggest that a network of salt-bridges between the ligand and this loop is necessary for binding . These contacts interconvert between each other on a time scale of a few tens of nanoseconds , making them elusive to X-ray crystallography . The loop is stabilized by an extra-domain helix . The latter influences the global dynamics of the domain , considerably increasing binding affinity . We found that two key contacts between the helix and the domain , one involving the loop , provide an atomistic interpretation of the increased affinity . Our analysis indicates that both extra-domain segments and loosely conserved regions play critical roles in PDZ binding affinity and specificity .
PDZ domains are modular protein interaction domains specialized in binding short linear motifs at the C-terminus of their cognate protein partners [1] , [2] . In human , they are found in hundreds of different proteins and are mostly involved in cell-cell adhesion and epithelial junctions [3] . PDZ domains are often classified on the basis of their preferred C-terminal ligand sequences . Early studies organized binding specificity in three canonical classes: class-I involving C-terminal motifs of the type [x– ( s/t ) –x– ( v/i ) cooh] , class-II [cooh] and class-III [x– ( d/e ) –x–cooh] , where is a hydrophobic residue and x any amino acid [2] , [4] . This classification , though consistent with the highly conserved binding groove [2] , does not explain the large selectivity observed both in naturally occurring C-terminal peptides and synthetic peptide library screening [5]–[8] . Systematic investigations of PDZ domain specificity revealed that more distal C-terminal peptide residues are involved in the binding process [7] , [9] , suggesting a role for the loop following the binding site [10]–[15] . For example , the solution structure of the second domain of the hPTP1E protein showed that this loop interacts with the sixth amino acid from the peptide C-terminus [10] , while possible electrostatic contacts between the loop and peptide amino acids up to position eight were found in the Par3 PDZ3-VE-Cad domain [14] , [15] . It was recently suggested that specificity beyond the canonical classes can be obtained by long-range interactions involving non-conserved structural motifs specific to the domain [16] . For instance , the extra-domain helical extension characterizing the third PDZ domain of PSD95 ( also called DLG4 or SAP90 ) was shown to influence binding [17] . Although this helix is away from the binding groove , affinity is reduced by 21-fold upon truncation of this non-conserved structural motif . Titration calorimetry measurements indicated that the free-energy penalty is entropic in nature . It was proposed that enhanced side-chain flexibility upon helix truncation , which is subsequently quenched by peptide binding , might be the main reason for this effect . This exquisitely dynamical behavior , calling for a hidden dynamic allostery [17] , [18] , pinpointed the importance of conformational entropy upon binding mediated by structural elements not directly evident from structural inspection alone [19] , [20] . Here , we investigate the set of interactions beyond the binding site influencing peptide binding in the PSD95-PDZ3:CRIPT complex . Molecular dynamics ( MD ) simulations indicate that residues upstream of the 4th C-terminal amino acid are crucial for binding . Specifically , lysines residues at position −4 and −7 in the CRIPT peptide are observed to dynamically interact with the loop . Shorter peptides spontaneously unbind from the domain , indicating that canonical interactions within the binding site are not sufficient for binding . Further simulations of the DLG1-PDZ2:E6 complex suggest a wide spread presence of such peptide-loop interactions in the PDZ family . Finally , we find that the extra-domain helix of PSD95-PDZ3 helps stabilizing the loop via ionic interactions . Our results provide direct evidence of the role played by peptide amino acids away from the C-terminus and the interplay with previously unrecognized PDZ structural motifs .
Seminal X-ray crystallography experiments on the third PDZ domain of PSD95 in complex with the CRIPT C-terminal peptide indicated that peptide binding is realized through the last four residues ( peptide positions 0 to −3 ) , while the rest of the peptide is mostly disordered [1] ( the system was crystallized with a 9-mer peptide , see below ) . This observation suggested a minor role of residues upstream of the last four ones for binding . To test this hypothesis , four MD simulation runs were carried out using a 5-mer peptide from CRIPT ( -KQTSV-COOH , CRIPT5 ) , a natural class-I binder of PSD95-PDZ3 ( see Methods ) [1] , [17] . Unexpectedly , all the four runs showed spontaneous unbinding within the first 110 ns ( see blue and light-blue lines of Fig . 1 for two unbinding trajectories and Table S1 for specific unbinding times and simulation lengths ) . Weak affinity was a somewhat surprising result , suggesting that canonical class-I interactions alone are not sufficient for binding . Interestingly , one of the runs showed rebinding from a partially unbound state . This event was mediated by the interaction of on the peptide with on the loop following the binding site as shown in Fig . S1 . The same peptide with a charged N-terminus ( CRIPT5* ) , which can reinforce this type of electrostatic interactions , remained anchored to the binding site for the total simulation time [21] . However , the peptide canonical contacts were only partially formed ( see Fig . S2 ) . These observations suggested that interactions beyond the canonical class-I motif are needed to achieve stable binding in native conditions ( i . e . without an artificially charged N-terminal peptide ) , possibly with a major role of the loop . To elucidate this point , four simulations with a longer 9-mer CRIPT peptide ( -TKNYKQTSV-COOH , CRIPT9 ) were performed for a total of roughly 700 ns . The peptide remained bound to the original X-ray configuration in all runs ( see red curve in Fig . 1 for a typical RMSD time trace ) . Strikingly , the four extra amino acids strongly influenced binding . The two lysines at peptide positions −4 and −7 transiently formed specific salt-bridges with two negatively charged loop residues , and . These contacts are dynamic , interconverting between each other on the ns time scale . On the other hand , their cumulative contribution is large: the loop and the ligand are in contact via salt-bridges for 44% of the time . These results indicate an unexpected and biologically relevant role of this loop , going beyond class-I interactions . Structural cluster analysis provides a quantitative classification of the non-canonical interactions ( see Methods for details ) . In Fig . 2 , structural ensembles characterizing the three most populated peptide-loop configurations are shown . We used a simplified code to classify the peptide-loop interactions . At the first , second and third position there is a “1” if interactions −7∶331 , −7∶332 or −4∶331 are formed , respectively; “0” otherwise ( these three contacts are the statistically more relevant ones ) . For example , “110” indicates that peptide is in contact with both and , as shown in Fig . 2e–f . The most observed configurations are “110” , “001” and “100” , having a relative population of 13% , 10% and 8% , respectively ( see Fig . 2 for their structural characterization; the cumulative 44% is obtained by summing up the remaining peptide-loop interacting conformations ) . This scenario is represented in Fig . 3 by the transition network of the different peptide-loop configurations ( see Methods ) . Multiple pathways are present , where a quite unspecific network of conformational changes stabilizes peptide-loop interactions on a time scale which is faster than unbinding ( for example , was measured for another member of the PDZ family [22] ) . Interestingly , the presence of peptide-loop interactions strongly influence the propensity to form canonical class-I contacts ( see Fig . S2 ) . The dynamic nature of the interactions explains why peptide-loop contacts were difficult to detect by previous structural experimental investigations [1] , [13] . For instance , both the original PDZ3 X-ray structure reported by McKinnon and collaborators [1] as well as further attempts by other groups ( e . g . PDB-ID:1TP3 ) indicated that only a four residue C-terminal stretch ( positions 0 to −3 ) is directly involved in binding . However , this observation is not supported by in vitro evolution and mutagenesis studies [7] , [9] , [15] . Along the same line , titration calorimetry experiments provided evidence for the role of peptide positions beyond −3 for both affinity and specificity [23] , while water-mediated interactions were found when bound to the oncogenic E6 peptide [13] . Our observations reconcile these two views , providing a unifying picture for peptide binding to PSD95-PDZ3 . While a dominant configuration characterizing the interactions between the peptide and the loop is absent , the cumulative effect of these interactions is necessary for binding . This effect is mostly dynamical , indicating that structure alone does not suffice to understand function in this case . PSD95-PDZ3 is characterized by an extra-domain helix at the C-terminus [1] , [17] . Structural analysis of our MD data showed that the helix directly interacts with the loop as well as with a region distant from the binding site , via two salt-bridges ( red dashed lines in Fig . 4a ) . The first one involves at the end of the helix and a negatively charged amino acid on the loop , . The second ionic interaction is between helix and , which is located in a region of the domain without specific secondary structure . This region ( blue in Fig . 4a ) , in turn , is in spatial contact with the carboxylate binding loop . No specific helix-peptide interactions were found , only unstable hydrophobic contacts . Recent experiments indicated that the extra-domain helix strongly influences the dynamics of the domain [17] . Binding affinity to the 9-mer CRIPT peptide was shown to decrease by 21-folds upon helix truncation through a purely entropic effect . The truncated form of PDS95-PDZ3 is defined by residues 306–395 , and referred to as throughout the text [17] . To provide atomistic insights into this mechanism , MD simulations of bound to CRIPT5 and CRIPT9 were performed ( see Table . S1 ) . The short 5-mer peptide unbound very quickly ( ) from the domain in all the four simulation runs , while CRIPT9 remained in the binding site . As observed for the WT , binding is stabilized by a network of dynamic salt-bridges between the ligand and the loop ( see Fig . S2 ) . Analysis of the backbone root-mean-square-fluctuations ( RMSF ) in the WT showed that the flexibility of the bound form is not affected by helix truncation ( Fig . 4b ) . However , it affects the unliganded ( apo ) form , enhancing the overall domain backbone flexibility ( Fig . 4c ) . The enhanced flexibility is mainly localized in three regions: the carboxylate binding loop ( residues 318–323 ) , the loop ( residues 330–336 ) and residues 341–356 . The latter corresponds to the region where the helix is forming the salt-bridges with . In our simulations for the WT , this interaction is present 49% and 41% of the time in the apo and bound forms , respectively . Given the spatial vicinity between this region ( i . e . , 341–356 ) and the carboxylate binding loop , we assume that the peaks relative to these two regions are coupled , arising from the missing interaction with the helix . Similarly , the enhanced flexibility of the loop is induced by the missing interaction with the extra-domain helix through the salt-bridge between and . This interaction is very stable in both the apo and peptide-bound states , being formed 83% and 82% of the time , respectively . These observations have important consequences for the interpretation of the entropic penalty upon binding to . Given that the flexibility of the bound form is unaffected by helix truncation , while it is much larger in the apo form , peptide binding to requires the quenching of the three regions reported in Fig . 4c and described above . Hence , our results suggest that the quenching of both the carboxylate and loops is responsible for the entropic penalty . Nevertheless , we cannot fully exclude other effects like a contribution from side chain dynamics , since decoupling entropy into local terms is a controversial and unsolved problem [24] , [25] . The important role of backbone dynamics is in contrast with recent NMR relaxation experiments which found a negligible contribution of the backbone compared to side chains flexibility [17] . We suggest that this apparent contrast can be solved by looking at the time scales of the fluctuations reported in Fig . 4c . RMSF differences peaks vanish when the time windows used for the calculations are similar to the ones relevant for NMR measurements , i . e . of the order of 10 ns or less ( grey and black lines in Fig . 4c ) . Our data indicates that the relevant backbone fluctuations are on the 100 ns time scale . Such dynamics is , on the one hand too slow to be detected by NMR spin-relaxation techniques ( i . e . ) [19] , [25] , [26] and , on the other hand , too fast to show up as a separate subpopulation in NMR relaxation-dispersion experiments ( i . e . ) . Stabilization of the extra-domain helix is further mediated by a hydrophobic patch , formed by and on the PDZ domain , and and on the helix , as shown in Fig . 5a . Analysis of all human PDZ domains ( see Methods ) revealed that , while position 337 largely consists ( i . e . 86% ) of aliphatic or aromatic residues , position 328 is less conserved , with a large portion of aliphatic amino acids ( see Fig . S3 ) . Free-energy calculations between this helix and the PDZ domain performed with FoldX [27] ( see Methods ) predict that V328A and V328I mutants in the apo-form have a of 1 . 35 and −0 . 79 kcal/mol , respectively . Hence , mutation to ALA destabilizes the domain . MD simulations of both mutants are consistent with this scenario . Given the direct interaction between the extra-domain helix and the loop ( Fig . 5b ) , it is found that bulkier aliphatics make this loop more rigid , avoiding the peptide induced quenching upon binding described in the previous section . Reversely , loop flexibility of the V328A mutant increases , approaching the one obtained in absence of the extra-domain helix ( , blue line ) . These results suggest a correlation between bulkier aliphatics at position 328 and the presence of an extra-domain helix . To further investigate this hypothesis , we used PSIPRED [28] to compute the helical propensity of C-terminal segments in all 258 human PDZ domains ( see Methods ) . A larger helical propensity is found for domains with ILE , LEU or VAL at position 328 , compared to the ones with ALA ( see Fig . 5c ) . For instance , around 10 residues downstream of the C-terminus , an helical propensity twice as large is found ( P-value of 0 . 02 , see Methods ) . These results correlate very well with our previous findings , indicating that large aliphatic side chains at position 328 can serve as anchors for extra-domain segments , stabilizing the loop . Consequently , domains with an alanine at position 328 are less likely to have an extra-domain helix and we expect that in those cases the loop would be structured differently with respect to PSD95-PDZ3 . This is in agreement , for example , with both PDZ1 and PDZ2 of PSD95 . These domains are known to lack the C-terminal extra helix , possess an alanine at position 328 and have a different composition of the loop ( see next section ) . The PSD95-PDZ3 loop ( together with V328 ) and the extra-domain alpha-helix are remarkably well conserved in orthologs up to fly ( and even partially conserved in worm ) , as well as in human paralogs such as SAP97 ( DLG1 ) , PSD93 ( DLG2 ) or SAP102 ( DLG3 ) , see Fig . S4 . In particular , the three charged residues involved in peptide binding and helix contact are conserved in almost all cases , providing indirect evidence that the same loop-mediated protein/ligand recognition is taking place in distant organisms . This is not the case when looking at the entire PDZ family , where the loop is highly heterogeneous both in length and amino acid composition . For instance , the loop of the PSD95-PDZ2 is more rigid , making self-interactions with the main domain body in a region close to the hydrophobic patch mentioned earlier [29] . Despite these differences , there are studies suggesting a role of the loop in binding to PDZ2 . Large chemical-shifts were measured in the loop region upon binding , substantially contributing to affinity [29] . Finally , X-ray crystallography of PDZ2 from the human paralog DLG1 in complex with the oncogenic E6 peptide pointed out to an asparagine on the loop ( ) interacting with the ligand backbone at position ( using our notation ) [13] . To provide a dynamical picture of the process , we performed additional simulations of the DLG1-PDZ2:E6 complex ( see Methods ) . Our calculations reiterate the importance of for binding to PDZ2 . It is found that the E6 peptide is in contact with the loop through mainly three interactions , : , : and : , for a total of 69% of time . An example structure is shown in Fig . 6 . These contacts interconvert on a ns time scale . Together with the results obtained for PDZ3 , these observations suggest that the loop is actively involved in binding specificity: a property that would need to be consistently explored throughout the entire PDZ family .
In PDZ binding , the relatively limited information about peptide amino acids more distant from the C-terminus prevented a clear structural understanding of the effect and importance of these upstream side chains . Our work aims to fill this gap by providing calculations with both a canonical 5-mer CRIPT peptide as well as a longer 9-mer peptide in complex with PSD95-PDZ3 . Three main results emerge from our work . First , we observe in our simulations that peptide binding is mediated by ionic interactions with the loop following the binding site , referred to here as the loop . These contacts are found with the 9-mer peptide , while the shorter 5-mer unbinds spontaneously after a few tens of ns . Recent experimental results on several PDZ domains support our interpretation [23] , [30] . Strong differences between short and long peptides were found for negatively charged loops ( e . g . MAGI1-PDZ2 ) [30] . Peptide-loop contacts are dynamic , where multiple specific interactions interconvert on a fast time scale of tens of ns ( i . e . much faster than unbinding [22] , [31] ) . Such dynamic interactions are likely to characterize several other PDZ domains . Further calculations on another member of the PDZ family , the DLG1-PDZ2 , which is characterized by a different loop , support our hypothesis . Moreover , unresolved side chains away from the C-terminus are often found in other PDZ-ligand X-ray structures ( see examples in Table S2 ) , indicating that these side chains can adopt multiple conformations . We note that the presence of positively charged residues downstream of the fourth C-terminal positions of PDZ peptide ligands is well attested by recent experimental specificity profiles [7] . These charged residues are not necessarily always at the same positions , even within ligands of the same domain [30] . This is likely so because the peptide is flexible at these positions ( as shown in Fig . 2 ) . Consistently , loops display a clear over-representation of negatively charged residues compared to other regions in PDZ domains: 11 . 6% of D/E in entire PDZ domains , 15 . 2% for D/E in loops ( according to the Fisher's test the probability to have this difference by chance is as low as , see Methods ) . Many of these residues on the loop provide clusters of negatively charged side chains that are ideally suited to recruit ligands with positive charges at any position between −4 and −7 . Second , we propose a mechanistic explanation for the microscopic origin of the binding entropic penalty in absence of the extra-domain helix of PSD95-PDZ3 . In the apo form , the helix plays a crucial role in stabilizing both the carboxylate binding loop and the loop . Hence , these two loops are more flexible in the helix truncated domain . In this case , the peptide quenches the two regions upon binding , resulting in the observed entropic penalty . This quench does not take place when the extra-domain helix is present . Our findings suggest that extra-domain regions might play a more important role than mere linkers between functional domains [16] , reiterating that the reductionist approach that protein domains can be studied in isolation should be always validated . This is especially important because several segments adjacent to domains show little sequence specificity ( and thus are often not included in domain definition ) , although they adopt well-defined secondary structures such as the in the third PDZ domain of PSD95 . Third , analysis of 258 human PDZ domains as well as MD simulations of single-mutants allowed for the identification of an amino acid at the beginning of the loop , VAL in PSD95 , that correlates with the presence of the extra-domain helix in other PDZ domains . Prediction of helical propensities at positions following the C-terminus of the domain showed enhanced probability for those domains presenting bulkier aliphatic side chains other than alanine at that position . This analysis suggests that a binding mechanism , indirectly involving the extra domain helix as in PSD95-PDZ3 , might be relevant for a significant portion of the PDZ domain family .
Molecular dynamics simulations were performed using the GROMACS implementation [32] of the CHARMM27 force field [33] , [34] at constant temperature and pressure with reference values equal to 300 K and 1 atm , respectively . The use of hydrogen virtual sites and fixed covalent bonds allowed a 4 fs integration time-step [35] . All systems were solvated in a dodecahedric box with an average of roughly 5000 tip3p water molecules ( see Table S1 for details of each simulation setup ) . In the case of PDZ3 , the system was equilibrated from the deposited X-ray structures 1BE9 and 1BFE [1] for the bound and apo forms , respectively , using residues 306–402 for the WT and 306–395 for . The PDZ2 starting structure is 2I0L [13] ( from DLG1/SAP97 ) . Each molecular setup was sampled by four independent runs of approximately 200 ns each for a total of ( Table S1 ) . The first 50 ns of each trajectory were neglected in the analysis to reduce the bias from the starting configuration . Snapshots were saved every ps . The peptide N-terminus was neutralized in all cases , except CRIPT5* . The sequences of the 9-mer peptides are -TKNYKQTSV-COOH and -LQRRRETQV-COOH for PDZ3 and PDZ2 , respectively . The first 5 peptide residues ( i . e . , positions from −4 to −8 ) as well as mutations at position 328 and the truncation of the extra domain helix were modeled using PyMol [36] . For each run , backbone RMSF values were calculated per residue as an average over the atoms C , and N . Final RMSF values were averaged over the four runs . Molecular trajectories were analyzed with the programs WORDOM [37] , [38] and GROMACS [39] . Hydrogen bonds were determined based on cutoffs for the angle Acceptor - Donor - Hydrogen ( ) and the distance Donor - Acceptor ( 3 . 6 Å ) . Ionic interactions are considered to occur when the two last carbons before the charged atoms are closer than 5 Å . Each protein-ligand snapshot was labeled by a four-digits code . The first three digits describe the peptide-loop interactions , e . g . “110” . The last digit represents an id , encoding the peptide structural conformation ( i . e . , the internal degrees of freedom ) . The latter was obtained by running a leader-based cluster-analysis on the ligand backbone ( atoms C , and N ) with a 2 Å cutoff , using the program WORDOM [37] , [38] . This digit distinguishes between different peptide conformations characterized by the same contacts with the loop . Each four-digit string represents a microstate of the protein-ligand complex . This decomposition is used to build a conformation-space-network [40]–[42] , where each microstate is a node and a link between two nodes is placed if there is a direct transition between them during the MD simulation . Basins of attraction are defined using a gradient-cluster analysis [43] , [44] , where multiple microstates are lumped together if they interconvert rapidly . Each gradient-cluster represents a metastable configuration , which can contain heterogeneous peptide-loop contacts . Connectivity between these metastable configurations is represented as a coarse-grained network as shown in Fig . 3 ( see also Fig . S2 in the Supp . Mat . ) . The gradient-cluster algorithm is freely available in the program PYNORAMIX ( GPL license , available at the website raolab . com ) . Predictions of free-energy differences upon mutations were done with FoldX using the BuildModel option after properly repairing the structures with the RepairPDB command [27] . The initial structure ( PDB 1BFE ) was first minimized with GROMACS in explicit water . This structure was originally crystallized with an ILE at position 328 . We mutated it both to VAL ( WT ) and ALA to compute the free-energy differences . The set of all human PDZ domains was retrieved from PFAM [45] and SMART [46] databases . A first multiple sequence alignment was generated with MUSCLE [47] . The alignment was manually curated , removing PDZ domains that could not be unambiguously aligned ( most of them are unconventional PDZ domains ) . This resulted in a total number of 258 PDZ domains ( see Table S3 ) . The loop was mapped by homology starting form the structure of PSD95-PDZ3 . Several PDZ domains are close paralogs , and this can result in strong biases when computing frequencies or correlation patterns . To account for this effect , we always grouped paralogs together ( see Table S4 ) . Groups of paralogs were defined using a cut-off of 50% on the sequence identity . The contribution of each member of a group was weighted by the inverse of the group size . For instance , to compute the amino acid frequency at a given position , residues from a group of 5 paralogs only contributed 1/5 each to the total frequencies . The helical propensity of C-terminal extensions of PDZ domains was computed with PSIPRED [28] for up to 20 residues downstream of the domains . If the protein C-terminus was reached before the 20 residues , a helix propensity of 0 was used . Here again , the contribution of paralogs was weighted to prevent purely phylogenetic correlations . P-values were computed by reshuffling the amino acid composition at position 328 in all PDZ domains of Table S3 . The Fisher's test was used to compute the probability to have a given number of negative residues within all loop residues , knowing the total number of negative residues within the sequences of all PDZ domains [48] . | Protein interactions play crucial roles in all biological processes . A common way of studying them is to focus on sub-parts of proteins , called domains , that mediate specific types of interactions . For instance , it is known that most PDZ domains mediate protein interactions by binding to the C-terminus of other proteins . Humans have more than 200 slightly different copies of these domains . At the level of the binding site , PDZ domains look quite similar . This is in apparent contradiction with their heterogeneous binding specificity . Using detailed molecular dynamics simulations in conjunction with statistical analysis , we predict that contacts outside of the canonical binding site play important roles in regulating protein interactions . Some of these contacts influence the overall dynamics of PDZ domains , providing an explanation for their allosteric effect . These interactions involve regions of the PDZ domains that are much less conserved , suggesting that they can help in differentiating selectivity in this large domain family . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"physics",
"biology",
"computational",
"biology",
"biophysics"
] | 2012 | Beyond the Binding Site: The Role of the β2 – β3 Loop and Extra-Domain Structures in PDZ Domains |
CD163 , receptor for the haptoglobin–hemoglobin complex , is expressed on monocytes/macrophages and neutrophils . A soluble form of CD163 ( sCD163 ) has been associated with the M2 macrophage phenotype , and M2 macrophages have been shown to down-modulate inflammatory responses . In particular , previous studies have shown that M2 is closely associated with the most severe clinical presentation of leprosy ( i . e . lepromatous leprosy ( LL ) ) , as well as tuberculosis . We hypothesized that sCD163 correlates with severity of diseases caused by intracellular pathogens . To assess this hypothesis , sCD163 levels were measured in the serum of leprosy and visceral leishmaniasis ( VL ) patients stratified by severity of the clinical presentation . sCD163 levels were significantly higher in patients with these diseases than those observed in healthy control individuals . Further analyses on infection and disease status of leprosy and VL patients revealed a clear association of sCD163 levels with clinical parameters of disease severity . In vitro culture assays revealed that Leishmania infection induced CD163 expression on the surface of both monocyte/macrophages and neutrophils , suggesting these cells as possible sources of sCD163 . FACS analyses shows that the cells expressing CD163 produces both TNF-α and IL-4 . Taken together , our results reveal sCD163 as a potential biomarker of severity of diseases caused by intracellular pathogens M . leprae and Leishmania spp . and have a modulatory role , with a mix of an inflammatory property induced by TNF-α release , but that potentially induces an anti-inflammatory T cell response , related to IL-4 release .
CD163 is a member of the scavenger receptor cysteine-rich family [1] . CD163 binds to hemoglobin ( Hb ) and haptoglobin ( Hp ) complex [2] and helps to coordinate the receptor-mediated endocytosis by phagocytes [3] to be processed by hemeoxygenase-1 ( HO-1 ) [4 , 5] . CD163 , largely expressed on monocytes/macrophages and neutrophils [6 , 7] , has several roles as an extracellular sensor for bacteria and modulator of immunological responses [8] . Alternatively activated macrophages , or M2 , have anti-inflammatory and tissue repair properties and have been described to express CD163 [9 , 10] . CD163 can be shed from the macrophage surface in response to inflammatory stimuli [3] , and can then be found as a soluble form ( i . e . sCD163 ) [7 , 11] . Macrophages expressing CD163 have been described in lepromatous leprosy ( LL ) , the most severe presentation of the infectious disease caused by Mycobacterium leprae , with CD163 facilitating bacterial survival by providing a source of iron for mycobacterial survival as well as triggering IL-10 production [7] . sCD163 has been identified as an indicator of disease severity in several inflammatory and infectious diseases [7 , 9 , 12–15] . Several prognostic and severity markers have been described in visceral leishmaniasis ( VL ) patients , including mucosal bleeding , jaundice , dyspnea , suspected or confirmed bacterial infections , neutrophil count <500/mm3 and platelet count <50 , 000/mm3 [16] . In addition , dos Santos et al ( 2016 ) reported that IL-6 , IL-27 and sCD14 can serve as useful biomarkers for severity of VL [17] , and that IL-6 levels greater than 200 pg/ml were strongly associated with death . Interestingly , CD163 is upregulated by interleukin-6 ( IL-6 ) and IL-10 [5 , 18 , 19] , two cytokines described to be high in VL patients [17 , 20–22] , but linkage of CD163 has not yet been reported for Leishmania infection . We therefore hypothesized that circulating sCD163 levels would correlate with severity of diseases caused by intracellular pathogens , such as leprosy and VL . We measured sCD163 levels in the sera of leprosy and VL patients to determine whether association could be made with severity of these diseases . Moreover , to determine if CD163 was simply indicative of disease state or might be involved in disease pathogenesis , we performed in vitro experiments to determine the impact of Leishmania infection on CD163 expression on macrophages and neutrophils .
The Ethics and Research Committee of the Federal University of Sergipe approved this study ( CAAE 0151 . 0 . 107 . 000–07 and CAAE 0152 . 0 . 107 . 000–07 ) and all recruits , or their legal guardians , willfully consented . All recruits provided written informed consent ( as outlined in the PLOS consent form ) to publication of their case details . Leprosy patients and their pertinent controls were enrolled at the Leprosy Clinic from the University Hospital , Federal University of Sergipe , in Sergipe State , Brazil ( HU-UFS ) . They were classified according to the Madrid ( 1953 ) criteria of clinical forms: Indeterminate Leprosy ( IL , n = 9 ) , Tuberculoid Leprosy ( TT , n = 14 ) , Borderline Leprosy ( BL , n = 14 ) and Lepromatous Leprosy ( LL , n = 10 ) [23] . The inclusion criteria were a diagnosis of leprosy confirmed by clinical aspects of the lesions and either positive bacilloscopy or histopathological confirmation in skin biopsies . Exclusion criteria were having other conditions ( pregnancy ) or diseases ( HIV , HTLV-1 , Diabetes ) that interfere in the immune response or in the clinical outcome of leprosy . After collection of blood and tissue samples , patients were treated following the standard multidrug therapy ( MDT ) , according to the Brazilian Ministry of Health and World Health Organization guidelines . Sera of household contacts of patients ( Contacts; n = 23 ) were used as controls . Contacts were individuals who lived in direct and prolonged contact with the leprosy patients and who submitted to careful dermatological exam to exclude the presentation of leprosy at the time of recruitment . As a group at elevated risk of M . leprae infection , however , we could not formally exclude the possibility that these contacts are infected or may become ill in the future . Clinical data and sera for VL patients , and their pertinent controls , were obtained from a database of the VL Reference Center at HU-UFS , Sergipe , Brazil . VL patients were divided into five groups; ( 1 ) before treatment ( D0-Classic , n = 33 ) , ( 2 ) 30 days after diagnosis with VL ( after treatment ) ( D30 , n = 19 ) , ( 3 ) severe VL at day 0 ( D0-SVL , n = 13 ) , ( 4 ) asymptomatic ( delayed type-hypersensitivity ( DTH ) -positive , n = 11 ) and ( 5 ) non-endemic health controls ( HC , n = 8 ) . DTH positive individuals are people who live with the patients and are responsive as measured by the DTH skin test positive for Leishmania soluble antigen , but do not have clinical symptoms of the disease . Patients were classified as having severe VL based on clinical features that included platelet counts <50 , 000/mm3 , bleeding , bacterial infections , neutrophil counts <500/mm3 , dyspnea and jaundice , as described by Sampaio et al . [16] . The inclusion criteria were VL diagnosis confirmed by direct observation of Leishmania in bone marrow aspirates or positive culture in NNN media ( Sigma-Aldrich ) , or a positive response in the rK39 serological test ( KalazarDetect Rapid Test: InBios International Inc ) . Patients were submitted to standard VL treatment with Antimonial ( Sbv ) [24] . Exclusion criteria were having other conditions ( pregnancy ) or diseases ( HIV , HTLV-1 , Diabetes ) that interfere in the immune response or in the clinical outcome of VL . Blood was collected from all volunteers and serum prepared . All sera samples were stored at -80°C until analyses . sCD163 quantification were performed at the same time for all sera samples by ELISA kit according to the manufacturer’s instructions ( R&D Systems ) . Haptoglobin was measured using a kit from GenWay , Heme-oxygenase I was measured using a kit from Assay Designs and Arginase-1 was measured using a kit from Hycult Biotech , following the manufacturer’s instructions . Monocytes were isolated from peripheral blood and plated in 24 well-plates at 5x105 cells/well . Differentiation of macrophages were performed as previously described by de Oliveira et al [25] . Neutrophils were isolated from peripheral blood samples of healthy donors ( with EDTA as anticoagulant ) using PolimorphPrep reagent , according to the manufacturer’s instructions ( Axis-Shield ) . The cells were washed with PBS prior to seeding into 96 well plates at a concentration of 106 neutrophils/well in RPMI 1640 supplemented with 10% FBS . L . ( L . ) amazonensis strain ( MHOM/BR/73M2269 ) [26] that constitutively expresses GFP and two different L . infantum isolates from VL patients ( MHOM/BR/2009/LVHSE17 as isolate 1 and MHOM/BR/2010/LVHSE49 as isolate 2 ) were used [27] . L . amazonensis-GFP was constructed by incorporation of the GFP gene into 18s ribosomal RNA by homologous recombination using pSSU vector , as described by Misslitz et al . ( 2000 ) [28] . L . infantum isolates were stained with CellTracker Violet BMQC dye ( Thermo Fisher ) as previously described [29] . The parasites were cultured axenically in Schneider's Drosophila medium ( Thermo Fisher ) plus 10% FBS prior to infection of cells . Infection of human macrophages was performed at a ratio of either 10 stationary-phase L . amazonensis-GFP promastigotes or 5 L . infantum per macrophage . Extracellular parasites were removed 2 hours later by washing . After 24h , the cells were then stained and subjected to flow cytometry . Neutrophils were infected at a ratio of 5 L . amazonensis parasites per neutrophil for 3h prior to staining and flow cytometry . Infection with Mycobaterium . bovis BCG ( Fundação Ataulpho de Paiva ) was performed at a ratio of 2 mycobacteria per macrophage . BCG were also stained with CellTracker Violet BMQC dye , following the same protocol used for Leishmania . Cells were washed with PBS and incubated with fluorescently-labeled antibodies according to the manufacturer’s instructions ( BD Biosciences , USA ) . Cells were incubated with anti-CD209-BV421 ( cat . 564127 ) , anti-CD163-PE ( cat . 556018 ) , anti-CD86-BV510 ( cat . 563461 ) and/or anti-CD40-APC ( cat . 555591 ) . To identify neutrophils , cells were incubated with anti-CD15-BV450 ( cat . 561584 ) and anti-CD163-PE . To assess cytokine expression , cells were incubated with BD Cytofix/Cytoperm reagent ( cat . 554722 ) , anti-TNF-alpha-PerCP-Cy5 . 5 ( cat . 560679 ) , anti-IL10-APC ( cat . 554707 ) , anti-IL-4-PerCP-Cy5 . 5 ( cat . 561234 ) and anti-IL-12-APC ( cat . 554576 ) . Cells were fixed with 4% paraformaldehyde prior to acquisition on a FACS CANTO II ( BD Biosciences ) and data was analyzed using FlowJo v10 . 0 software ( Tree Star ) . The gating strategy was to first set a gate in the FSC/SSC in the regions compatible with either macrophages morphology ( macrophage surface phenotype and cytokine analysis ) or neutrophil morphology . For surface phenotype analysis , the second step was to distinguish infected from uninfected cells using FSC vs GFP or Celltracker dot plots . The third step was to set a positive gate for each marker according to the florescence of each label versus FSC . Statistical analyses were performed using Windows GraphPad Prism version 5 . 0 ( GraphPad Software ) . Results are expressed as mean ± standard deviation ( SD ) . D’Agostinho-Pearson normality test was performed to establish if the data had a normal distribution . Differences between two groups were determined by Mann-Whitney test ( sCD163 analysis ) . Parametric t paired test was used in before/after treatment analysis . Friedman paired test with Dunn´s post test ( Surface Phenotype and Median of Fluorescence Intensity analysis ) and Wilcoxon paired test ( Intracellular cytokine analysis ) were used for non-parametric paired analyses . Correlation analysis was performed using Spearman correlation test . A p-value ≤ 0 . 05 was considered significant .
Patients with the most heavily infected and severe manifestation of leprosy , LL , had higher serum sCD163 levels than both households contacts of patients ( contacts ) ( p<0 . 001 , Mann-Whitney test ) and patients presenting with tuberculoid leprosy ( TT ) ( p = 0 , 001 , Mann-Whitney test ) , indeterminate leprosy ( IL ) ( p = 0 . 01 , Mann-Whitney test ) or borderline leprosy ( BL ) ( p = 0 . 0009 , Mann-Whitney test ) ( Fig 1A ) . Receiver Operating Characteristic ( ROC ) curves were constructed and area under the curve ( AUC ) analysis highlighted the utility of sCD163 levels for distinguishing between LL and either TT ( AUC = 0 . 8571 , 95% confidence interval ( CI ) [0 . 69–1 . 02] , p = 0 . 0034 ) or contacts ( AUC = 0 . 8439 , 95% CI [0 . 72–1 . 02] , p = 0 . 0008 ) ( Fig 1B and 1C ) . In contrast , levels of haptoglobin , heme-oxygenase-1 and arginase-1 were not different between the groups ( p>0 . 05 for all comparisons , Student t test or Mann-Whitney test ) . The mean ± SD of haptoglobin levels in these groups were: LL ( 46 . 4 ± 21 . 00 ) , TT ( 45 . 7 ± 39 . 50 ) and Contacts ( 47 . 3 ± 31 . 78 ) ; of heme-oxygenase-1 were: LL ( 0 . 6 ± 0 . 34 ) , TT ( 0 . 4 ± 0 . 16 ) and Contacts ( 0 . 4 ± 0 . 24 ) ; and arginase-1 were: LL ( 16 . 9 ± 11 . 04 ) , TT ( 10 . 6 ± 4 . 60 ) and Contacts ( 15 . 6 ± 9 . 24 ) . We detected high serum levels of sCD163 in VL patients compared to healthy individuals from non-endemic regions ( HC ) and Leishmania-infected but healthy controls ( delayed-type hypersensitivity-positive; DTH+ ) ( p<0 . 0001 , Mann-Whitney test ) ( Fig 2A ) [17] . Fig 2F shows the distinction between control group ( HC ) versus D0-classic patients ( AUC = 0 , 9697 , CI [0 , 9112–1 , 2] , p = 0 , 0001 ) , by ROC curve , reiterating the value of sCD163 as a biomarker of disease . The highest sCD163 levels were detected in patients classified as presenting with severe VL ( D0-SVL ) ( p<0 . 004 , Mann-Whitney test ) ( Fig 2A ) . ROC analysis of D0-classic versus D0-SVL patients ( AUC = 0 , 8403 , CI [0 , 7101–0 , 97] , p = 0 , 0004 ) ( Fig 2G ) provides further support for the use of sCD163 in determining disease severity and clinical improvement , respectively . A direct correlation was observed between serum sCD163 levels and both spleen size ( Spearman r = 0 . 3915 ) ( Fig 2B ) and liver size ( r = 0 . 4353 ) ( Fig 2C ) , while an inverse correlation was observed between sCD163 concentration and neutrophil counts of VL patients ( r = -0 . 4918 ) ( Fig 2D ) . These measures represent standard clinical parameters used for determining VL severity [17] , and our results therefore indicate the potential of using serum sCD163 levels as an indicator of VL severity . Having demonstrated the linkage of sCD163 levels with severity of VL , we speculated that levels would be reduced as disease resolved upon treatment . Accordingly , paired analysis of samples collected before and after treatment ( D0-Classic and D30 , respectively ) demonstrated a reduction of sCD163 serum levels in 10 of 15 patients after the completion of treatment ( Fig 2E; p = 0 . 0455 , paired t test ) . ROC analysis of D0-classic versus D30 group ( AUC = 0 , 7376 , CI [0 , 5832–0 , 89] , p = 0 , 0046 ) ( Fig 2H ) provides further support for the use of sCD163 in monitoring clinical improvement . These data support our hypothesis that the decrease of sCD163 levels can be used as an indicator of treatment success . In contrast to the sCD163 data , significant differences were not observed between these groups in terms of haptoglobin , HO-1 or arginase-1 levels ( p>0 . 05 for all comparisons , Student t test or Mann-Whitney test ) . The mean ± SD of haptoglobin levels in D0-Classic ( 25 . 8 ± 29 . 82 ) , D30 ( 10 . 2 ± 6 . 46 ) , D0-SVL ( 11 . 6 ± 9 . 72 ) and DTH+ ( 14 . 5 ± 6 . 75 ) ; heme-oxygenase-1 in D0-Classic ( 1 . 6 ± 2 . 13 ) , D30 ( 0 . 1 ± 0 . 03 ) , D0-SVL ( 2 . 1 ± 3 . 98 ) and DTH+ ( 0 . 2 ± 0 . 003 ) ; and arginase-1 in D0-Classic ( 7 . 2 ± 6 . 38 ) , D30 ( 7 . 1 ± 6 . 23 ) , D0-SVL ( 12 . 2 ± 14 . 33 ) and DTH+ ( 3 . 5 ± 3 . 09 ) . To assess the impact of infection on CD163 we infected cells in vitro with various Leishmania species . CD163 , CD40 , CD209 and CD86 expression were evaluated in macrophages incubated with a Leishmania amazonensis strain expressing GFP , two L . infantum isolates stained with Violet Celltracker or BCG also stained with Celltracker ( Fig 3A ) . While L . infantum and L . amazonensis infection induces CD163 expression in macrophage surface , BCG infection did not ( Fig 3B ) . Macrophages exposed to L . amazonensis for 24 hours yielded two cells subpopulations [30] , with GFP+ cells considered infected while GFP- cells were assumed to be non-infected . The infected cells had a higher percentage of CD86+ ( 57 . 11% ) and CD163+ ( 33 . 6% ) cells compared to both non-infected ( CD86 44 , 78% , CD163 3 , 22% ) and unstimulated cells ( CD86 36 , 93% , CD163 12 , 62% ) ( Fig 3C ) . The median fluorescence intensity ( MFI ) of GFP was evaluated in CD40 , CD86 , CD209 and/or CD163 positive populations to assess the relationship of parasite load with these surface molecules . CD163+ cells showed higher infection ( MFI = 913 . 167 ) than CD40+CD163- ( MFI = 715 . 143 ) cells ( p = 0 . 0044 , Friedman paired test ) ( Fig 3D ) . A direct correlation was observed between the CD163 expression levels and Leishmania infection ( GFP ) levels ( r = 0 , 67 , p<0 , 0001 ) ( Fig 3E ) . These results support the hypothesis that highly infected macrophages could be a source of sCD163 . To assess the effector function of CD163 expressing macrophages and its implications in the regulation of the immune response , flow cytometry was performed to evaluate the cytokine profile of these cells . A greater frequency of cells expressing TNF-α and IL-4 was detected in the CD163+ population ( MFI and integrated MFI; Fig 4B–4D and Fig 4K–4M ) . We did not observe any differences in IL-12 and IL-10 expression between CD163+ and CD163- cells ( Fig 4E–4J ) . To identify if infected neutrophils also expressed CD163 , neutrophils from healthy individuals were exposed to L . amazonensis ( Fig 5A ) . Flow cytometry reveals a higher percentage of CD15+CD163+ cells within the infected , GFP+ population ( 25 . 08% ) relative to the uninfected , GFP- ( 14 . 84% ) and unstimulated control populations ( 6 . 90% ) ( p = 0 . 0008 , Friedman paired test ) ( Fig 5B ) . These findings parallel those obtained with macrophages . Taken together , our in vitro infection data suggest that both macrophages and neutrophils are sources of sCD163 during Leishmania infection and that the quantity of CD163 is directly correlated with infection level .
CD163 is a scavenger receptor that has previously been identified as an indicator of disease severity in several inflammatory and infectious diseases [7 , 9 , 12–15] . Consistent with this , our data reveal a strong correlation of serum sCD163 levels with the severest clinical presentations of leprosy and , for the first time , VL . Moreover , we observed that Leishmania infection induces CD163 expression on monocyte/macrophages and neutrophils , suggesting that these cells might be a source of sCD163 during VL . The infected CD163+ macrophages preferentially produced TNF-α and IL-4 . Collectively , our data indicate the induction of CD163 expression during infection with M . leprae and Leishmania species , likely modulating the immune response to permit high levels of infection and the most severe clinical presentations of these diseases . Our data examining leprosy patients corroborate the previously reported association of serum sCD163 with the lepromatous leprosy ( LL ) presentation [7] . Moreover , our study extends and refines the previous data by showing that measurement of serum sCD163 can distinguish LL from the other clinical forms of leprosy , including intermediate forms of leprosy ( IL and BL ) . Moura et al ( 2012 ) demonstrated also the presence of CD163+ macrophages in lesions of LL patients and therefore attributed the association of the serum levels of sCD163 and severity of leprosy to the differentiation of macrophages to the M2 phenotype [7] . Sousa et al ( 2016 ) found a direct correlation between CD163 expression and a variety of inflammatory cytokines in lesions of LL patients , including arginase-1 enzyme expression which is characteristic of M2 macrophages [31] . Interestingly , we did not find differences in the levels of haptoglobin and the enzyme that degrades the hemoglobin-haptoglobin complex ( Hb-Hp ) , heme-oxygenase-1 ( HO-1 ) and arginase-1 among the groups , suggesting that sCD163 levels are a more robust physiologic alteration . We also detected higher levels of sCD163 in the serum of VL patients , with these levels correlating with multiple clinical parameters of disease severity . The direct correlation of serum sCD163 concentration with liver , spleen size and an inverse correlation with neutrophil counts supports the use of sCD163 as a surrogate indicator of disease severity in VL patients . In addition , a decrease in the levels of sCD163 was observed between the start and end of treatment ( D0 versus D30 ) , suggesting that sCD163 measurements could be used to monitor response to treatment . Declines were not observed in all patients , however , with 5 of the 15 patients presenting with similar or even higher levels of sCD163 at the end of treatment . At D30 the VL patients are still recovering from various symptoms of this disease , and a longer follow-up may be beneficial and further studies are required to more strenuously determine the utility of this biomarker in case management . Based on the observations from patient samples , we evaluated how CD163 expression is induced and what the likely cellular sources of this molecule are . We observed that infection by Leishmania induced CD163 expression on the surface of both macrophages and neutrophils , identifying these cells as potential sources of the sCD163 detected in serum . This was true for two Leishmania species ( observed with L . amazonensis—GFP and two isolates of L . infantum ) and this is the first study demonstrating that Leishmania parasites can induce this macrophage phenotype . In leprosy , a similar relationship has previously been suggested by the observation in lesion biopsies of macrophages expressing CD163 that are heavily infected with M . leprae [31] . Moreover , Moura et al ( 2012 ) showed CD163 expression on monocytes can be induced by M . leprae infection in vitro [7] . During M . leprae infection co-expression of CD209 and CD163 is indicative of a permissive and phagocytic programming of macrophages characteristic of heavily infected lesions [32] . Interestingly , BCG is a potent pro-inflammatory stimulus with as a previously demonstrated macrophage polarization to a M1-like phenotype , [33] and our data shows that BCG down regulates CD163 expression . Thus , we observed striking differences in the response to pathogenic M . leprae and nonpathogenic M . bovis BCG . In our model Leishmania infection did not lead to increases in CD209 positive cells with uninfected and infected populations having the same frequency of CD209+ macrophages . Under the in vitro conditions we used , only CD163+ cells was more frequent in the infected than in the uninfected population . CD209 is a marker expressed in earlier stages of activation and it is therefore possible that infection kinetics may have an impact . Regardless , these data denote some differences in the responses to these intracellular pathogens . Flow cytometry analysis found that not only is the CD163+ population more prone to infection with Leishmania parasites but produces more IL-4 and TNF-α than CD163- macrophages . No differences were observed in IL-12 or IL-10 production between these phenotypes . Saha , et al . ( 2016 ) also found in hepatitis C infection that macrophages expressing surface markers of M2 can produce both anti- and pro-inflammatory cytokines [34] . In the complex evolution of VL , these responses might interfere in instructing T cells and inhibit the killing of these intracellular pathogens . Moura et al . ( 2012 ) has also suggested that these macrophages are characteristic of M2 subtype that is more permissive for M . leprae infection [7] . In addition to macrophages , neutrophils infected with L . amazonensis also express CD163 and may be another source for sCD163 in the serum of VL patients . Groselj-Grenc et al ( 2008 ) found expression of CD163 on neutrophils in systemic inflammatory response syndrome [6] and an immunomodulatory role of polymorphonuclear leukocytes has been described during the early phase of L . major infection [35] . We suggest that CD163 may be related to differentiation of these neutrophils toward an anti-inflammatory N2 subtype , such as has recently described [36 , 37] . Further studies are needed to confirm the presence of CD163 positive neutrophils in VL patients during active disease .
In conclusion , our data indicate the potential use of serum levels of sCD163 in indicating severity of diseases caused by intracellular pathogens . Our results corroborate and expand previous findings for leprosy and , for the first time , demonstrate high levels of sCD163 correlate with severe clinical symptom observed in VL patients . This study also suggests that infected macrophages and neutrophils are possible sources of sCD163 , and show that both L . amazonensis and L . infantum can polarize macrophages to produce both pro- ( and anti- inflammatory cytokines ( TNF-α and IL-4 , respectively ) . This would appear to favor parasite multiplication and exacerbate clinical presentation . | Visceral leishmaniasis ( VL ) is a systemic , and most severe form of leishmaniasis . Soluble CD163 ( sCD163 ) levels can serve as biomarker for disease severity in several inflammatory disorders . However , no linkage has been reported for its relationship with Leishmania infections . We now demonstrate , for the first time , that sCD163 is increased in VL patients , and its presence is directly correlated to clinical parameters of disease severity . In vitro infection of monocyte-derived macrophages and neutrophils with L . infantum and L . amazonensis induces , while BCG reduce the expression of CD163 on macrophage surface Furthermore , presence of sCD163 is reduced during clinical improvements . Taken together , results reveal an important role for sCD163 in immune modulation during disease progression , and suggest a potential role as biomarker for determining disease severity and clearance . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion",
"Conclusion"
] | [
"blood",
"cells",
"flow",
"cytometry",
"innate",
"immune",
"system",
"medicine",
"and",
"health",
"sciences",
"immune",
"cells",
"immune",
"physiology",
"cytokines",
"immunology",
"tropical",
"diseases",
"parasitic",
"diseases",
"parasitic",
"protozoans",
"bacterial",
... | 2017 | sCD163 levels as a biomarker of disease severity in leprosy and visceral leishmaniasis |
The heterosexual risk group has become the largest HIV infected group in the United Kingdom during the last 10 years , but little is known of the network structure and dynamics of viral transmission in this group . The overwhelming majority of UK heterosexual infections are of non-B HIV subtypes , indicating viruses originating among immigrants from sub-Saharan Africa . The high rate of HIV evolution , combined with the availability of a very high density sample of viral sequences from routine clinical care has allowed the phylodynamics of the epidemic to be investigated for the first time . Sequences of the viral protease and partial reverse transcriptase coding regions from 11 , 071 patients infected with HIV of non-B subtypes were studied . Of these , 2774 were closely linked to at least one other sequence by nucleotide distance . Including the closest sequences from the global HIV database identified 296 individuals that were in UK-based groups of 3 or more individuals . There were a total of 8 UK-based clusters of 10 or more , comprising 143/2774 ( 5% ) individuals , much lower than the figure of 25% obtained earlier for men who have sex with men ( MSM ) . Sample dates were incorporated into relaxed clock phylogenetic analyses to estimate the dates of internal nodes . From the resulting time-resolved phylogenies , the internode lengths , used as estimates of maximum transmission intervals , had a median of 27 months overall , over twice as long as obtained for MSM ( 14 months ) , with only 2% of transmissions occurring in the first 6 months after infection . This phylodynamic analysis of non-B subtype HIV sequences representing over 40% of the estimated UK HIV-infected heterosexual population has revealed heterosexual HIV transmission in the UK is clustered , but on average in smaller groups and is transmitted with slower dynamics than among MSM . More effective intervention to restrict the epidemic may therefore be feasible , given effective diagnosis programmes .
HIV infection was first detected in the United Kingdom ( as AIDS ) in 1981–2 [1] among MSM . Early outbreaks with UK sources include Scottish IDUs dated to 1983 [2] and haemophiliacs to 1984 [3] . All strains isolated initially were of the B subtype , both in MSM and IDUs [4] and also in the small number of individuals infected through heterosexual contact during that decade [5] . However within 10 years , multiple subtypes had been detected within the UK [6] . From the mid 1990s increasing numbers of HIV infections in the UK were being found in heterosexuals , until the current situation was attained whereby this risk group comprises the majority of new HIV diagnoses [7] . This increase coincided with increasing immigration from southern and Eastern Africa , particularly from South Africa , Uganda and Zimbabwe [8] . Genetic characterisation of viruses from infected heterosexuals revealed that while subtype B was still observed in the majority of samples obtained during 1996/7 [9] , by the year 2000 , subtype C was most common ( 35% ) with subtype A at 15% , reflecting the main subtypes in those countries . Subtype B was present in only 25% of individuals [10] . Thus , the heterosexual risk group in the UK has become strongly associated with non-B HIV subtypes . Recently there has been some evidence of limited crossover among risk groups with a study of over 5000 patients from London reporting 2 small clusters of subtype A ( n = 21 ) among MSM , of whom approximately 50% of individuals were white [11] . We have applied recently developed methods of molecular phylodynamics to the analysis of partial HIV pol gene sequences obtained during routine clinical care from over 2000 MSM attending a single large clinic in London [12] . We showed that 25% of individuals whose virus showed a link to at least one other individual in the study were in fact linked to 10 or more others . Using relaxed clock approaches [13] we found that 25% of transmissions within these clusters took place within a maximum of 6 months after infection . This suggested that the elevated risks of transmission associated with acute HIV infection could be important for driving a significant component of the HIV epidemic among MSM . In this study we have analysed the entire dataset of individuals infected with non-B subtypes of HIV and receiving clinical care within the UK who are represented in the UK HIV Drug Resistance Database . The overwhelming majority ( 95% ) of non-B subtype HIV in this dataset is associated with heterosexual transmission and 83% with Black-African ethnicity [14] . Since 2003 in the UK , a baseline HIV genotyping assay has been recommended when antiretroviral therapy is initiated and accordingly a large proportion of sequences within the database have been obtained prior to therapy . Non-B subtype HIV pol sequences were available from over 11 , 000 individuals for this study: for comparison the estimated number of HIV-infected Black African and Caribbean individuals in the UK was 24 , 000 in 2007 [7] . We therefore estimate we have analysed almost 40% of the UK heterosexual HIV-infected population .
From the sequence dataset representing over 25 , 000 subjects , non-B subtypes were identified mainly using the REGA method [15] , with additional information from ad hoc phylogenetic analysis ( see Methods ) . Due to the limited number of subtypes other than A and C , these other non-B subtypes were grouped for analysis . This gave datasets of the following sizes: for subtype A , N = 1581; for C , N = 6096 and for other non-B subtypes , N = 3394 . Within these groups , the initial subset of sequences linked to at least one other was selected from all pairwise comparisons using the threshold of 4 . 5% nucleotide similarity at third codon positions [12] . This identified sequences from 367 patients infected with subtype A , 1372 infected with subtype C and 1035 infected with other non-B subtypes , a total of 2774 individuals . The datasets were then modified by removal of codons associated with drug resistance ( see Methods ) and Bayesian MCMC phylogenetic analysis was performed on subtype A and subtype C separately . In the resulting trees , 4 subtype A and 14 subtype C phylogenetic clades of ≥10 individuals were identified with a posterior probability of 1 ( Figures S1 & 2 ) . This corresponds to 25% of the subtype A closely-related sequences and 21% of the subtype C closely-related sequences . A similar analysis was performed on the 1035 sequences from other non-B subtypes . In the last case , the main fully supported clades reflected subtype divisions and were unrelated to transmission patterns . However , from within the main subtype splits we were able to identify 7 fully supported subtrees of ≥10 individuals for further analysis ( Figure S3 ) . Unlike the case for the subtype B sequences previously studied [12] , the clustering of non-subtype B sequences includes patient linkage outside of the UK . We therefore performed further analyses in which the nearest sequences to each cluster from the global HIV database were included . This leads to the breakdown of a number of clades through the inclusion of sequences from outside the UK within what were previously monophyletic groups ( Figure 1A , 1B & S4 ) . The resulting distribution of cluster size is shown in Figure 2 . Including the closest sequences from the global HIV database left 296 individuals that were in UK-based groups of 3 or more individuals . Large clusters still comprise a significant proportion of patients with a link to at least one other . The largest for subtype A was a cluster with 24 individuals and that for subtype C was one of 33 individuals . The percentage of sequences found in clusters ≥10 individuals was 14% ( subtype A ) ; 6% ( subtype C ) and 1% ( others ) , respectively . A total of 143 of the original 2774 ( 5% ) individuals were found in large clusters , although these comprised 48% of individuals within UK-based groups of 3 or more . In this and our previous study of subtype B sequences , the distribution of individuals in clusters strongly suggested a power law relationship indicative of a scale-free network . With the additional data available we have examined the fit of a power law to the non-B subtype data . The goodness of fit to a power law varies with the maximum time depth allowed for clusters . We have used the date of sampling to limit the time depth and having considered a range of values ( Figure S5 ) , find that restricting the analysis to subclusters with a maximum depth of 5 years reveals a very good fit ( Figure 3; R2 = 0 . 95; p<10−6; α = 2 . 1 ) . We previously made use of the statistically rigorous approach of relaxed-clock phylogenetics implemented in BEAST to obtain estimates , and highest posterior density distributions , of dated nodes within clusters [12] . Each sequence is obtained from a different patient so from the internode interval we can infer maximum , estimates of inter-transmission intervals . Any missing data in the form of additional individuals in the network would lead to shorter average transmission intervals . For all time-scaled subtrees of all UK-based groups containing ≥3 individuals ( 296 individuals in total ) , determined as described above , the internode distances were estimated ( Figures S6 , S7 , S8 ) . These yielded maximum estimates of transmission intervals for UK-based non-B clusters whose medians were 32 months ( subtype A ) and 25 months ( C ) , respectively and 22 months for other subtypes ( Figure 4A ) . Overall , for non-B HIV the median transmission interval for UK-based groups was 27 months . The proportion of transmission intervals in the first 6 months of infection was 0% , 2% and 5% for subtypes A , C and others , respectively , giving 2% overall . The proportion of transmission intervals between 6–36 months after infection for the non-subtype B clusters were: 53% ( subtype A ) , 68% ( subtype C ) and 56% ( others ) with an overall proportion of 62% . In this population therefore , the possible heightened risk of transmission associated with acute infection appears not to play a significant role in the epidemic ( Figure 4 ) .
We have retrospectively investigated the dynamics of the developing heterosexual HIV epidemic in the UK by applying Bayesian phylogenetic analysis to anonymised viral sequences obtained in the course of routine clinical treatment . The high level of representation in the UK HIV Drug Resistance Database ( over 40% of the estimate of the relevant risk group ) has permitted a detailed analysis of the level of clustering , the distribution of cluster size and the distribution of the interval between transmissions for non-B subtype sequences . After screening out non-UK associations , we have found that among probable UK-based infections , 14% of subtype A sequences were found in clusters ≥10 individuals , with 6% of subtype C and 1% for others , although these percentages increase sharply ( to a total of 48% ) if the denominator is restricted to the 293 individuals within UK-based clusters of 3 or more . That this would suggest that individuals within a UK-based cluster of any size are very likely to be in a large one is a striking conclusion as all likely confounding factors ( such as immigration of concordant families ) might increase the numbers of pairs , and perhaps clusters of 3 individuals but not of clusters of 10 or more , and therefore would decrease the proportion in large clusters . Despite the different geographical origin of HIV-1 subtypes , large clusters were observed in both subtype C ( 33 members ) whose primary origin would be southern Africa , and subtype A ( 24 members ) which is primarily associated with East Africa , suggesting no major distinction in the structure of the epidemic among communities from different countries . We explored the epidemic in these groups in greater detail by using time-resolved phylogenies to analyse the dynamics of transmission within clusters , adopting a relaxed molecular clock [13] . As each sequence is obtained from a different infected individual we take the internode interval as a maximum estimate of the time between transmissions [12]: missing data , in the form of individuals within the transmission network who were not sampled , would always reduce this estimate . Taking this approach a median estimate of the time between transmissions of 27 months was observed overall for non-B subtypes ( 32 , 25 and 22 months for subtypes A , C and other , respectively ) . This approach also allowed the estimation of the proportion of transmissions within defined intervals after infection: overall just 2% of transmissions in this population were estimated to occurr within 6 months or less ( 0% , 2% and 5% for A , C and other subtypes , respectively ) . In an earlier study of the phylodynamics of HIV in an MSM population attending a large clinic in London we observed a much higher frequency of linkage between individuals with 25% of those with a connection to at least one other being found in large clusters [12] . Among these MSM the median transmission interval within clusters , estimated in the same way , was almost half that for the heterosexual population studied here , at 14 months , and 25% of transmissions within clusters occurred within 6 months of infection . Nevertheless , the shape of the distribution of cluster size was similar between the two groups . The overall proportion of transmission intervals between 6–36 months after infection for the heterosexual clusters , 62% , is very similar to that estimated for the MSM dataset ( 63%; Figure 2B ) . While there is an extended right-hand tail of the transmission interval distribution for non-subtype B UK transmission clusters ( Figure 2A ) this is likely to be due in part to the inclusion of a residue of non-UK based distantly linked sequences which were not identified by the global diversity screen . In a recent study of patients selected in primary transmission in Quebec , Brenner et al . [16] indicated that while 28% of MSM diagnosed early in infection were part of transmission clusters involving 5 or more individuals , only 13% of non-B subtype infections ( mostly heterosexual ) were in clusters . The observed differences between MSM and heterosexuals in inter-transmission intervals could reflect real differences in the dynamics of the epidemics in different risk groups . In this study a possible cause of such a distinction could have been a systematic difference between them , for example in the sampling of the population if there were many more missing individuals from the heterosexual clusters . At the most basic level this would appear to work in the opposite way , as the earlier MSM study was restricted to individuals attending a single clinic in London [12] , while the analysis presented here derived from population surveillance of all HIV-infected individuals receiving treatment in the United Kingdom . As indicated earlier ( see Introduction ) , these results reflect approximately 40% of the HIV-infected Black African population . In contrast , the earlier study analysed 2126 individuals sampled from approximately 11 , 000 MSM receiving care in London ( www . hpa . org . uk ) , i . e . ∼20% of those receiving care and perhaps 10–15% of all MSM in London . We therefore have approximately 3–4 fold greater coverage of the of African-derived HIV in the UK in this study than of MSM in London previously . Another possible source of bias could lie in the frequency of testing . The possibility of higher awareness and/or access to HIV-related care among MSM than among the predominantly immigrant HIV-infected heterosexual group could in principle have led to a shorter time between infection and diagnosis . If this also led to a shorter time between infection and initiation of antiretroviral therapy then the period of opportunity for transmission could be reduced . Time of infection is unknown for most of the patients studied so we investigated this possibility by using CD4 counts at the time of diagnosis as a proxy for the average time since infection ( Text S1 , Table S1 ) . In agreement with Stöhr et al . [17] , we conclude that there is little difference between the heterosexual and MSM groups in the UK ( Figure S9 ) : the 10% difference we observe in CD4 count at treatment between subtypes C and B cannot explain the observed 50% difference in the median inter-transmission interval . Following the observations of Liljeros et al . [18] that human sexual networks based on contacts within the last year have the properties of scale free networks , we have examined the distribution of the size of transmission clusters among heterosexuals in the UK and find an excellent fit to a power law , consistent with a scale-free network ( Figure 3 ) . Inference from viral sequence data is not direct and as discussed in detail earlier [12] , it is important to recognise that the viral transmission network and the sexual network are not the same in a chronic infection such as HIV: a series of transmissions could derive from a single individual rather than as onward transmissions from their sexual contacts . The transmission network is a subgraph of the sexual network but clearly both incorporate a time dimension; the network that fits a power law was that described in terms of sexual contacts in the last year [18] and is smaller than the lifetime network . Here we tested several time depths and found that the best fit was obtained with a limit of 5 years , and the value of the shape parameter α , was estimated at 2 . 1 ( Figure 3 and S5 ) , close to estimates obtained by Liljeros et al . [18] The greater time depth reflects the substantial delay that is usual between infection , diagnosis and the onset of antiretroviral therapy , which would have been the indication for a HIV genotype test from which our sequences are derived . While nodes in a sexual network and nodes in a transmission network cannot be directly equated , the distribution in time of the latter is clearly bounded by the former . On the other hand , the relationship of the sexual network to the transmission network is determined by the probability of transmission per contact which varies greatly and is difficult to estimate [19] . Therefore a quantitative description of the transmission network for a population can provide critical information for modelling the epidemiology of HIV transmission . The degree of clustering deduced from heterosexual population differs from that found previously for MSM and there is a substantial difference in the dynamics . While it is generally recognised that concurrent partnerships form the greatest potentiating factor for HIV and other STIs , the difference between these risk groups suggests either a longer interval between partner change , or a lower per-contact risk of transmission in heterosexuals . With very few inter-transmission intervals below 6 months it is unlikely that the elevated viral load associated with acute infection [20] plays a significant role in the UK heterosexual epidemic . The slower dynamics of the heterosexual epidemic thus offer more opportunity for successful intervention , but it is essential that diagnosis is achieved as early as possible .
The patient data derived from the 25631 patients in the UK HIV Drug Resistance Database ( www . hivrdb . org ) as at 2007 , who had been recruited over the previous 10 years ( Text S1 , Figure S10 ) . Of the patients reported on here with non-B subtype HIV 5777 ( 76% ) were recruited from London , 797 ( 11% ) from Manchester , 549 ( 7% ) from the rest of northern England and Scotland and 463 ( 6% ) from Birmingham and the Midlands . Ethical approval for this work was given by the London Multicentre Research Ethics Committee ( MREC/01/2/10; 5 April 2001 ) . Where multiple sequences were present for patients within the database the oldest sequence was selected . Sequences were aligned using the sequence alignment tool in HyPhy [21] with problematic sequences aligned manually by eye . The final alignment was 1554 nucleotides ( nt ) in length ( concatenated full-length protease [PR] and partial reverse transcriptase [RT] coding sequences ) with individual sequences ranging from 791–1536 nucleotides ( median 1269 ) , according to genotyping method . HIV-1 subtype was determined using REGA [15] and HIVdb ( http://hivdb . stanford . edu ) . Sequences for which the two methods yielded discordant results were additionally assessed phylogenetically ( using NJ trees created in PAUP* under the HKY85 nucleotide substitution model [22] ) for clustering with sequences representing the 10 major subtypes within the dataset ( A , B , C , CRF02_AG , D , G CRF06_cpx , F , H , J ) . Unique inter-subtype recombinants which might confound the phylogenetic analysis were eliminated at this stage . Within-subtype recombination , which would remain undetected , would introduce artifactually long branches and could have the effect of removing some individuals from clusters that they belonged to . Only limited numbers of subtypes other than A and C were found; these other non-B subtypes were grouped together for analysis . Clinical and epidemiological data was available for those patients recruited to the United Kingdom Collaborative HIV Cohort ( UK CHIC ) [23]; A: 620 ( 39% ) , C: 1387 ( 23% ) , other non-B: 1207 ( 36% ) . Phylogenetic analysis was performed as described earlier [12] , initially removing 39 codons associated with antiretroviral resistance creating an alignment of 1437 nucleotides in length . Identification of clades likely to contain UK transmission clusters was performed using Bayesian Monte Carlo Markov Chain ( MCMC ) approach [24] with the HKY85 model of nucleotide substitution with gamma distribution of rate variation ( Γ ) . The HKY model was used because the very large size of the trees being generated meant the more complex GTR model would lead to an excessive computation time . In this study the greatest interest lies in the most closely related sequences where the difference in performance between these two models is least . Trees were rooted using a subtype G outgroup taken from the UK HIV Drug Resistance Database . Due to the size of the alignments , it was necessary to perform multiple runs for each dataset . An initial run of 5×106 generations was performed and parameter estimates ( nucleotide frequencies , transition/transversion ratio [κ] , gamma shape parameter [α] ) taken from this run . Further runs of 5×106 generations were started using the fixed values of parameters from the first run and the highest likelihood tree from the previous run until convergence of the three separate chains was visibly seen . Posterior consensus trees were generated using the final run only ( after a burn-in of 50% ) , summarising 25000 trees . In order to isolate probable epidemiologically linked UK-restricted transmission clusters from phylogenetically-defined clusters of ≥10 individuals , clusters were included in a new phylogenetic analysis together with the 100 most closely related sequences to the cluster . These were selected from the LANL HIV database by comparing the cluster consensus sequence to all sequences for that subtype in the database by % difference . For each cluster and its most closely related global sequences , a MrBayes tree was generated ( 5×106 generations , HKY+Γ model ) . Clusters that remained monophyletic with high support ( ≥0 . 95 ) were then defined as UK transmission clusters . Dated phylogenies were obtained using a Bayesian MCMC method ( BEAST version 1 . 4 . 7; [13] using a relaxed molecular clock [25] . Clades of ≥10 individuals fully supported in MrBayes phylogenetic trees were analysed in their entirety and the results for subtrees defined as UK transmission clusters extracted from the full time-scaled trees . The date used for each sequence in any analysis was the number of days since the isolation date of the oldest sequence within the entire dataset , with sequences dated using the number of days from the earliest sequence isolation date . Analysis was performed using the SRD06 model of nucleotide substitution [26] with a lognormal distribution of rates amongst branches . Lognormal priors were placed on root height , corresponding to a median of 20 years and an upper 5% limit of 40 years . The most appropriate demographic prior on population size ( constant or exponentially increasing ) was determined using Bayes factors [27] with a log10 Bayes factor ≥5 indicative of substantial evidence of improved model fitting [28] . For each cluster , BEAST analysis used three separate MCMC runs of chain length 1×107 ( sampling parameters and trees every 1000 generations ) combined after a 10% burn-in ( leaving 30000 trees ) . Tree samples were used to generate a maximum clade credibility ( MCC ) tree with mean node heights using TreeAnnotator ( available from http://beast . bio . ed . ac . uk ) . Partial sequences of the HIV pol gene analyzed here have been deposited in GenBank under accession numbers GQ462027-GQ462532 . | Since 1995 , HIV among heterosexuals in the UK increased to the point where the total number of heterosexuals infected with HIV , predominantly of non-B subtypes , exceeds the number of HIV-positive homosexual men . To understand the dynamics of this epidemic , we have applied the novel technique of phylodynamics to the analysis of viral sequences taken in the course of routine clinical care from approximately 40% of the HIV-infected heterosexual population in the UK . Phylodynamics reconstructs the pattern of viral sequence divergence in time , revealing the size of transmission clusters and the dynamics of transmission within them . Of 11 , 071 patients studied , 296 were linked to at least two others in the UK . There were 8 clusters comprising 10 or more individuals among these , yielding a total of 143 or 5% of all individuals with links , much lower than seen earlier among homosexual men ( 25% ) . Viral transmissions within clusters also occurred less rapidly , only 2% being dated to the first 6 months of infection , compared to 25% among homosexual men . Overall , transmission clusters exist in the UK heterosexual HIV epidemic but they are generally smaller than among homosexuals; onward transmission occurs less rapidly and is not associated with acute HIV infection . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"public",
"health",
"and",
"epidemiology/infectious",
"diseases"
] | 2009 | Molecular Phylodynamics of the Heterosexual HIV Epidemic in the United Kingdom |
Robustness is a property built into biological systems to ensure stereotypical outcomes despite fluctuating inputs from gene dosage , biochemical noise , and the environment . During development , robustness safeguards embryos against structural and functional defects . Yet , our understanding of how robustness is achieved in embryos is limited . While much attention has been paid to the role of gene and signaling networks in promoting robust cell fate determination , little has been done to rigorously assay how mechanical processes like morphogenesis are designed to buffer against variable conditions . Here we show that the cell shape changes that drive morphogenesis can be made robust by mechanisms targeting the actin cytoskeleton . We identified two novel members of the Vinculin/α-Catenin Superfamily that work together to promote robustness during Drosophila cellularization , the dramatic tissue-building event that generates the primary epithelium of the embryo . We find that zygotically-expressed Serendipity-α ( Sry-α ) and maternally-loaded Spitting Image ( Spt ) share a redundant , actin-regulating activity during cellularization . Spt alone is sufficient for cellularization at an optimal temperature , but both Spt plus Sry-α are required at high temperature and when actin assembly is compromised by genetic perturbation . Our results offer a clear example of how the maternal and zygotic genomes interact to promote the robustness of early developmental events . Specifically , the Spt and Sry-α collaboration is informative when it comes to genes that show both a maternal and zygotic requirement during a given morphogenetic process . For the cellularization of Drosophilids , Sry-α and its expression profile may represent a genetic adaptive trait with the sole purpose of making this extreme event more reliable . Since all morphogenesis depends on cytoskeletal remodeling , both in embryos and adults , we suggest that robustness-promoting mechanisms aimed at actin could be effective at all life stages .
Every embryo develops under its own unique set of circumstances and challenges . To then ensure a reliable outcome , mechanisms are built into development to buffer against fluctuations in genetic , biochemical , and environmental inputs [1] . This buffering , called “robustness” , can be overwhelmed , ending in miscarriage , shortened gestation , and structural and functional birth defects [2] . Thus , we need to understand how developmental robustness arises in order to define an embryo's susceptibilities to genetic/epigenetic background and environment; and to ultimately promote healthy reproduction . Many mechanisms are used to buffer biological systems against fluctuating inputs , including redundant protein function [3] , [4] , secondary or “shadow” enhancers [5] , [6] , smart network design [7]–[10] , and chaperone activity [3] , [11]–[14] . Among developmental systems , a rigorous quantitative understanding of these mechanisms has been largely limited to examples where cell fate decisions are made , and robustness is fostered at the level of gene expression [5] , [6] , [10] , [15] or signaling [9] , [16] , [17] . For morphogenesis , which translates cell fate decisions into embryonic form , the detailed characterization of specific buffering mechanisms has been slower to come . Morphogenesis requires activities that span nuclei , cytoplasm , and whole tissues , and is driven by cell shape change [18] , [19] . So robustness could be promoted at many levels ( e . g . gene expression , signaling , membrane dynamics , cytoskeletal remodeling , and cell adhesion ) . But we do not know enough about the molecular and mechanical underpinnings of morphogenesis to predict where its greatest susceptibilities are , or where buffering mechanisms would be most effective . Specifically , we lack a comprehensive understanding of how cell biological steps convert gene expression into reliable tissue-building events . Actin and microtubules seem like good targets for robustness-promoting mechanisms during morphogenesis because they drive cell-shape change and modulate the mechanics of cells and tissues [20] , [21]; however , experimental support for this is lacking . In order to identify the mechanisms that promote robustness , the outcomes of the process in question must be quantifiable [1] . For example , we know of robustness-promoting mechanisms for cell fate decisions in development because the outcomes are binary , and typically happen along well-separated spatial dimensions so that fidelity can be readily tracked and quantified over a range of perturbations [5] , [6] , [9] , [10] , [15]–[17] . For morphogenesis , which is spatially complex , challenging to image , and has long been scored by qualitative rather than quantitative methods , fidelity is not easily measured . Consequently , the number of well-tested examples that show how robust morphogenesis is achieved remains low , in the context of both individual cell shape change and whole tissue remodeling [22] , [23] . To address this gap in our knowledge , we are using the first tissue-building event in the fly embryo , cellularization , as a simple , quantifiable model to study robustness . Fly embryos first develop as a syncytium , passing through 13 mitoses with no intervening cytokinesis . Then , during cell cycle 14 the embryo undergoes cellularization , during which ∼6000 cortically-anchored nuclei are simultaneously packaged into a sheet of cells that will be the primary epithelium ( Figure 1A ) [24] . Cellularization takes ∼60 minutes and plasma membrane furrows ingress 35 µm , cutting straight between adjacent nuclei to form mononucleate cells . This simple architecture allows unambiguous quantification of the fidelity of cellularization , where furrow failures or regressions show up as multinucleate cells , and hundreds to thousands of packaging events can be assayed per embryo to generate a ratio of mononucleate cells-to-nuclei ( ratio = 1 in wild-type embryos; Figure 1A ) . Fly embryos cellularize just after the maternal-to-zygotic transition ( MZT ) [24] , when transcription from the zygotic genome starts to maximally impact the developmental program [25] . Thus , the few zygotic genes that are required for cellularization have long been thought of as “switches” to control this morphogenetic event [26] , [27] . However , we now report a new role for one of these long-supposed switches , Sry-α . We identify Sry-α as a zygotic gene product that is expressed at the MZT , not to control cellularization , but rather to make it robust in the face of both environmental and genetic perturbations . We find that Sry-α acts together with its maternally provided paralog Spt to reinforce the actin cytoskeleton and so promote robust cellularization . Our data provides a clear example of how zygotic contributions , made at the MZT , not only instruct development , but also supplement the maternal machinery to ensure the fidelity of specific morphogenetic events . What's more , our data suggests that this robustness is fostered via regulation of the actin cytoskeleton .
At the start of cellularization , actin filaments ( F-actin ) accumulate at incipient furrow tips , which in surface views form furrow “canals” around the nuclei ( Figure 1A ) [24] . Furrow canals are then maintained throughout cellularization , and are required for stable furrow ingression ( Figure 1A ) [28]–[30] . We previously showed that mutations or drug treatments that reduce F-actin levels in all furrow canals , precipitate the regression of a fraction of furrows ( Figure 1A ) [29] , [30] , consistent with multinucleation phenotypes reported for actin regulators like Rho1 GTPase , RhoGEF2 , and the Formin , Diaphanous [28] , [31] , [32] . In an effort to identify other actin regulators that are required for the fidelity of cellularization , we examined a poorly characterized mutant , serendipity-α ( sry-α ) , that similarly displays a multinucleation phenotype ( Figure 1B ) [26] , [27] . The sry-α gene was previously mapped , and is expressed at the MZT just prior to cellularization [27] . Consequently , sry-α has long been thought of as a developmental cue that provides some new activity to trigger cellularization [26] , [27] . We found that all furrow canals in sry-α null mutants ( sry-α−/− ) have significantly reduced levels of F-actin compared to wild-type , throughout cellularization ( Figure 1C , 1D ) . In addition , incipient furrow canals in sry-α−/− mutants display an increased number of Amphiphysin tubules ( Figure 1E ) , which indicates promiscuous endocytosis upon F-actin reduction [29] , [33] . These results show that Sry-α regulates F-actin levels in furrow canals during cellularization . Based on remote homology searches , including PHYRE [34] and I-TASSER [35] , we found that Sry-α is a novel member of the Vinculin/α-Catenin Superfamily ( Figure 2A , 2B ) [36] . Our analysis also identified a Sry-α paralog in the D . melanogaster genome that we called Spitting Image ( Spt , CG8247 ) . Sry-α and Spt align with the middle sequences of Vinculin and α-Catenin , including the Vinculin-Homology 2 domain ( VH2; Figure 2A , Figure S1 ) [37] , [38] . Like α-Catulin , Sry-α and Spt represent a distinct clade of the Vinculin/α-Catenin Superfamily ( Figure 2B , Figure S2 ) [39] , [40] . Based on a “roll-call” analysis of orthologs in organisms with fully sequenced genomes , sry-α and spt co-exist in all Drosophilids , while spt alone is present in other insects ( see Table S1 ) . In higher metazoans , PHYRE analysis also identified other uncharacterized proteins , like Sry-α and Spt , which share remote homology with the middle sequences of Vinculin and α-Catenin ( Figure S1 , S2 ) . Members of the Vinculin/α-Catenin Superfamily peripherally associate with the plasma membrane and interact with the actin cytoskeleton [36] . Thus , this evolutionary relationship is functionally consistent with a role for Sry-α and Spt at the actin cortex during cellularization . To examine the relationship between Sry-α and Spt , we asked if these paralogs have unique or overlapping functions . Both Sry-α and HA-tagged Spt localize to F-actin rich furrow canals in cellularizing embryos ( Figure 2C , 2D ) . In addition , we used an F-actin co-sedimentation assay to show that both Sry-α and Spt bind F-actin directly . Recombinant Sry-α and Spt proteins were purified from insect cells and mixed with F-actin . Upon centrifugation , F-actin and its interacting proteins pellet ( α-Actinin; Figure 2E , 2F ) , while unbound proteins remain in the supernatant ( GST; Figure 2E , 2F ) . For both Sry-α and Spt , we detected a significant fraction of F-actin bound protein ( Figure 2E , 2F ) . Thus , the co-localization of Sry-α and Spt , as well as their biochemical activity , suggest that they could act redundantly during cellularization . To look for overlapping in vivo functions , we first used RNAi [41] and found that spt knockdown ( sptRNAi ) causes multinucleation during cellularization , which is qualitatively indistinguishable from either the sry-α−/− genetic mutant or sry-α RNAi ( sry-αRNAi; Figure 3 , Figure S3 ) . Additionally , embryos with double sptRNAi plus sry-αRNAi knockdown display a strongly enhanced multinucleation phenotype ( Figure 3B , 3C ) . Together , the localization , biochemistry , and RNAi phenotypes suggest that Sry-α and Spt share redundant functions during cellularization . To then confirm this , we tested whether Spt overexpression ( sptOE ) can rescue sry-α−/− multinucleation . We used Sry-α immunostaining to genotype embryos ( Figure S4A ) , and found that 100% of sry-α−/− mutants are rescued by sptOE ( Figure 4A , 4B , Figure S4 ) . Rescue of sry-α−/− multinucleation by sptOE is equivalent to that accomplished with a genomic construct encoding sry-α itself ( Figure S4B ) . We also confirmed that sptOE restores F-actin to wild-type levels in sry-α−/− mutants ( Figure 4C ) . Thus , Sry-α and Spt share an overlapping actin regulating function during cellularization . These results challenge a long-standing idea that sry-α is zygotically expressed at the MZT to supply some new activity that instructs cellularization to proceed [26] , [27] . In fact , the Sry-α activity may already be available via Spt . We compared the developmental expression profiles of endogenous Sry-α and Spt . As previously shown , sry-α transcript and protein are expressed at the MZT , in a pulse that just coincides with cellularization ( Figure 5A , 5B ) [27] . Conversely , spt transcript and protein are provided maternally , and Spt levels persist throughout and far beyond cellularization ( Figure 5A , 5B , Figure S5 ) . That is , a pulse of zygotically expressed Sry-α adds to a pool of maternally provided Spt during cellularization . ( i . e . Sry-α expression only boosts an already existing Spt activity in the embryo . ) Given that Spt can completely replace Sry-α during cellularization ( Figure 4 , Figure S4 ) , Sry-α does not supply some unique activity that triggers the process . Instead , the expression profiles suggest that the level of Spt plus Sry-α is somehow critical for the successful progression of cellularization . It was previously shown that the co-expression of paralogs in worms and yeast promotes biological robustness [3] , [4] . Presumably , these “gene duplicates” provide overlapping functions , and so replace or supplement each other in the face of internal and external perturbations [3] , [4] . Hence , we hypothesized that there is a threshold level of Sry-α plus Spt activity that is required for cellularization to proceed with high fidelity . This robustness hypothesis predicts that at an optimal condition , Sry-α function is dispensable and Spt alone can support successful cellularization , whereas both are needed at sub-optimal conditions . To test this , we assayed for multinucleation phenotypes in sry-α−/− null mutants at an optimal temperature . We chose 18°C , the lowest temperature at which D . melanogaster thrives . We reasoned that the lower temperature would reduce the demand on the actin cytoskeleton , because F-actin is more stable at lower temperature [42] . As predicted , we found that multinucleation is suppressed in sry-α−/− mutants that are reared at 18°C ( Figure 6A , 6B ) . At this temperature , the ratio of mononucleate cells to nuclei in sry-α−/− mutants is not significantly different than wild-type ( Figure 6B ) . Nor did we detect changes in the dimensions of the cells that formed for sry-α−/− mutants at 18°C ( data not shown ) . Thus , Spt activity is sufficient for cellularization to proceed at an optimal condition . However , multinucleation in sry-α−/− mutants was increasingly severe at higher temperatures ( 25–32°C; Figure 6A , 6B ) , showing that Spt is not enough to ensure reliable cellularization when conditions deviate from the optimal . A second prediction of the robustness hypothesis is that reducing sry-α dosage will make cellularization more likely to fail when the embryo is challenged by perturbations [3] , [4] . To test this , we reduced the Sry-α level using genetic heterozygosity , and looked for multinucleation at an extreme temperature of 32°C . This temperature marks an upper limit at which D . melanogaster embryos can survive , but developmental events are measurably impaired [5] , [6] . We found that the occurrence of multinucleation at 32°C is significantly increased for sry-α heterozygotes ( sry-α+/−; Figure 7A , 7B ) . Genotypes were confirmed by RNA FISH ( Figure S6 ) [43] . So while Spt alone is sufficient for cellularization at an optimal temperature , Spt plus two copies of the sry-α gene are required when environmental conditions are extreme . This suggests that the expression of Sry-α serves as a robustness-promoting mechanism for cellularization . A hallmark of robustness-promoting mechanisms is that they respond equally well to different kinds of perturbations ( e . g . genetic , biochemical , or environmental ) [44] . For example , shadow enhancers for the fly genes snail and shavenbaby promote robust gene expression at high temperature , and when mutations reduce input from their respective activation pathways [5] , [6] . In addition , a systems-level analysis in yeast showed that genes that promote mutational robustness also promote high fitness in a wide range of stressful environments [44] . Thus , we also tested whether the Sry-α level ensures the fidelity of cellularization upon genetic perturbation . We checked for multinucleation in sry-α+/− heterozygotes that carry only a half maternal dose of profilin ( ½profilin ) . Profilin is an actin accessory protein that promotes actin polymerization [45] . Strikingly , we saw multinucleation in sry-α+/− heterozygotes in the ½profilin background even at 18°C ( Figure 7C , 7D ) . We conclude that the fidelity of cellularization , in the face of both environmental and genetic perturbations , critically depends on the level of Sry-α .
Our data support a model wherein maternal Spt plus zygotic Sry-α work together to promote the robustness of cellularization . Our findings refute the idea that there is a clear passing off of developmental control from the maternal to the zygotic machinery at the MZT [25]–[27] , [46] . For example , the maternal genome was previously thought to provide the basic cellular machinery for cellularization , while Sry-α and other zygotic players provided the instructions [26] , [27] . More recently zygotic gene products have been shown to actively degrade maternal mRNAs , arguing that there may be a clean break from maternal to zygotic control [47]–[49] . In both cases , the zygotic contribution is largely viewed as being instructive . But our data speaks to a more collaborative interaction between the maternal and zygotic genomes: We show that cellularization can proceed with no input from zygotic Sry-α . Instead of controlling cellularization , we find that Sry-α actually adds to the activity of maternal Spt to make this morphogenetic event more reliable in the face of environmental and genetic perturbation . Sry-α is not taking over for the maternal product because it is only expressed in a pulse during the demanding event of cellularization , while maternal Spt persists far beyond cellularization . So , the relationship of Spt and Sry-α illustrates with exceptional clarity that the maternal and zygotic genomes also work together , with redundant activities , to make development robust . Certainly , this collaboration is likely to be broadly conserved ( e . g . maternal plus zygotic contributions of Rac proteins in C . elegans and Cadherins in Xenopus support the progression of specific morphogenetic events [50] , [51] ) . Since maternal RNAs and proteins are loaded into oocytes and eggs long before they act in development ( up to months ) [25] , their levels may not be very reliable [46] . Overlapping activities encoded by the zygotic genome could then buffer this variability and ensure successful progression through early embryogenesis . In our case , this relationship was revealed by the distinct expression profiles of paralogs provided from the maternal and zygotic genomes . But the same end is likely accomplished by expressing a single gene both maternally and zygotically . In fact , there is the recurrent observation that some proteins are expressed zygotically , while a significant maternal pool still persists ( e . g . Drosophila β-Catenin ) [46] , [52] . In D . melanogaster , roughly 5% of the genome displays this expression profile , with a maternal contribution supplemented by zygotic expression at the MZT [47] . But why split the contribution between the maternal and zygotic genomes ? For cellularization , why is the same level of robustness not achieved by simply expressing more maternal Spt ? We can envision at least two possibilities: either Spt and Sry-α have some distinct functions , perhaps at different developmental stages; and/or high levels of Spt/Sry-α protein are harmful . Sry-α and its expression at the MZT may be an adaptive genetic trait with its function , perhaps its sole function , serving to buffer cellularization against external and internal perturbations . According to our roll-call analysis , both spt and sry-α are present in the genomes of all Drosophilids , while other insects only encode spt ( see Table S1 ) . Some aspect of development specific to Drosophila may then stabilize the strategy of maternal Spt plus zygotic Sry-α . For example , cellularization in Drosophila builds tall columnar cells around thousands of nuclei , and so may be more demanding than cellularization in other insects where shorter cuboidal cells form around fewer nuclei [53]–[55] . Alternatively , Drosophilids share a fast rate of embryogenesis in comparison to many other insects [53] , [54] , [56] , which could make their cellularization more difficult . Thus , the pulse of zygotic Sry-α may be advantageous for meeting the challenge of a very extreme cellularization event in Drosophila . Our data suggests that zygotic Sry-α adds to the activity of maternal Spt to regulate actin , and so ensure the fidelity of cellularization . A significant future challenge will be understanding what specific actin-based activities promote robust morphogenesis . For example , F-actin is a critical determinant of tissue mechanical properties during development because it assembles with Myosin-2 and other crosslinkers , within single cells , to form a cortical network that ( 1 ) governs the rigidity and shape of the whole embryo [57]–[59]; and ( 2 ) generates the forces for and resistance to the cell shape changes that drive morphogenesis [60]–[64] . Thus , actin could promote robustness by buffering the mechanical properties of cells and tissues . In fact , Spt and Sry-α bind F-actin directly . Also , Spt and Sry-α are related to the F-actin crosslinker Vinculin , and they contain the conserved M-domain at the VH2 , which in Vinculin dimerizes to support F-actin crosslinking [38] . So Spt and Sry-α could promote robust cellularization by crosslinking F-actin and modulating tissue mechanics . Alternatively , F-actin also controls the membrane remodeling that accompanies cell shape change and morphogenesis . Specific to Drosophila cellularization , F-actin antagonizes endocytosis to favor plasma membrane growth and furrow ingression [29] , [33] . Consequently , F-actin could also promote robustness by controlling the membrane dynamics of morphogenesis . Since all morphogenesis depends on actin remodeling , both in embryos [65] and adults [66] , we believe that robustness-promoting mechanisms that target actin are likely to be ubiquitous .
The sry-α−/− and sry-α+/− embryos were collected from Df ( 3R ) X3F/TM3 , Sb [26] crossed to OreR wild-type flies . The nulloX embryos were collected from C ( 1 ) DX , ywf [29] , [30] . For sry-α−/− plus sptOE , first matαtub-GAL4 ( II ) was crossed with UASp-Spt-HA ( III ) ( this paper ) to make stock matαtub-Gal4; UASp-Spt-HA . Second , these flies were crossed with Df ( 3R ) X3F/TM3 , Sb , and finally embryos were collected from matαtub-GAL4/+; UASp-Spt-HA/Df ( 3R ) X3F mothers crossed with their sibling matαtub-GAL4/+; UASp-Spt-HA/Df ( 3R ) X3F fathers . For sry-α−/− plus sry-αrescue , Df ( 3R ) X3F/TM3 , Sb was crossed to sry-α genomic rescue stock ( II ) ( gift of E . Wieschaus ) . For ½profilin , chic221 cn1/CyO; ry506 ( Bloomington Stock Center #107932 ) was crossed with Df ( 3R ) X3F/TM3 , Sb; and embryos were collected from chic221 cn1/+; Df ( 3R ) X3F/ry506 mothers crossed with their sibling fathers chic221 cn1/+; Df ( 3R ) X3F/ry506 . For RT-PCR and Western Blotting , OreR was used . For RNAi imaging , embryos were injected from stock Gap43-mCherry/CyO; Histone-GFP/TM3 , Sb ( parental stocks gifts of A . Martin and E . Wieschaus ) . For UASp-Spt-HA flies , the coding sequence of D . melanogaster spt ( CG8247 ) was fused with a C-terminal hemagglutinin ( HA ) tag , and cloned into pP ( UASP ) vector . Transgenesis and mapping followed standard methods ( BestGene , Inc . ) . For anti-Spt antibody , the coding sequence of spt fused with a C-terminal histidine tag was cloned into pET vector , and recombinant protein was purified from E . coli . Antibodies were produced in guinea pigs according to standard methods ( Panigen , Inc . ) . According to the publicly available data on FlyBase ( flybase . org ) , spt is not expressed in adult heads . Thus , adult heads served as the negative control for antibody characterization ( Figure S5 ) . Recombinant proteins GST-Sry-α ( full length ) and GST-SptΔ ( amino acids 250–461 ) were produced in Sf9 cells ( Baylor College of Medicine Baculovirus Core/Proteomics Shared Resource ) . The Spt truncate was used because full-length protein is insoluble in both bacterial and insect expression systems . Proteins were purified with Glutathione Sepharose ( GE Healthcare ) and dialyzed in F-buffer ( 20 mM Tris-HCl pH 7 . 5 , 2 mM DTT , 2 . 5 mM MgCl2 , 75 mM KCl and 10 mM NaCl ) . α-Actinin ( Cytoskeleton , Inc . ) was used as the positive control , and GST for the negative control was purified from Sf9 cells . Rabbit muscle monomeric actin ( G-actin ) was extracted from Rabbit Muscle Acetone powder ( Pel-Freez Biologicals ) . 20 mM G-actin was maintained in G-buffer ( 2 mM Tris-HCl pH 8 . 0 , 0 . 2 mM ATP , 0 . 5 mM DTT , and 0 . 2 mM CaCl2 ) until polymerization in F-buffer by adding buffer and salt and incubating at room temperature for 1 hour . For the co-sedimentation assay , purified proteins were pre-cleared by centrifugation at 900 , 000 rpm for 30 minutes at 4°C to remove any precipitate . Pre-cleared proteins ( ∼1 µg ) were incubated with F-actin for 30 min on ice and then centrifuged at 900 , 000 rpm for 30 min at 4°C . Supernatant was removed , and replaced by an equal volume of 3X sample buffer to resuspend pellets . Equal volumes of supernatants and pellets were separated on 5–10% SDS-PAGE gels and transferred to nitrocellulose . GST-Sry-α , GST-SptΔ , and GST were detected by 1∶10 , 000 mouse anti-GST ( ab19256 , Abcam ) . α-Actinin was detected by 1∶1000 anti-α-Actinin ( A7811 , Sigma-Aldrich ) . The goat secondary antibodies were HRP conjugates used at 1∶10 , 000 ( Jackson ImmunoResearch ) . Embryo collection cups were set up on apple juice plates according to published protocols [29] , [30] , [41] . For specific temperature incubations , collection cups were housed in air incubators of 18°C , 25°C , or 32°C within a range of ±1°C for 5 , 4 , or 3 hours , respectively . Plates were harvested and embryos fixed immediately . For immunofluorescence , embryos were fixed in boiling salt buffer; or 4% paraformaldehyde , 0 . 1 M phosphate buffer ( pH 7 . 4 ) : heptane ( 1∶1 ) [29] . Antibody concentrations and fixation methods are listed in Table S2 . Due to Myosin-2 antibody incompatibility with FISH , Septin ( Peanut ) was used as the furrow canal marker for experiments where heterozygosity was scored . For F-actin staining , embryos were fixed in 8% paraformaldehyde , 0 . 1 M phosphate buffer ( pH 7 . 4 ) : heptane ( 1∶1 ) and hand-peeled for staining with 5 U ml−1 Alexa 488-phalloidin ( Invitrogen-Molecular Probes ) . For nuclear staining , Hoescht 33342 was used at either 0 . 25 µg ml−1 for standard immunofluorescence or 1 . 0 µg ml−1 for FISH ( Invitrogen-Molecular Probes ) . For FISH , 44 oligonucleotide probes , covering the sry-α coding sequence were synthesized ( Biosearch Technologies ) , and then labeled with Alexa 488 [43] . Embryos were fixed in 4% paraformaldehyde , 0 . 1 M phosphate buffer ( pH 7 . 4 ) : heptane ( 1∶1 ) for hybridization with 50 µM Alexa 488-sry-α probe followed by immunofluorescence staining . For fixed and living embryos , images were collected on a Zeiss LSM 710 confocal microscope with a 40X/1 . 2 numerical aperture water-immersion objective ( Carl Zeiss , Inc . ) . Images were collected at a zoom of two , with resolution of 104 nm per pixel . To segment images , the Image Processing Toolbox in MATLAB ( MathWorks ) was used . From raw images , the furrow canal network and nuclei masks were generated using a series of morphological operations , as follows . For furrow canal network masks: ( 1 ) A Gaussian filter was applied to the raw image , to retain only coarse features of the furrow canals . ( 2 ) An intensity threshold was selected manually and applied to generate a preliminary mask . ( 3 ) The mask was thinned iteratively until a 1-pixel width network was produced . For nuclei masks: ( 4 ) A Gaussian filter was applied to the raw image to smooth out noise , yet retain fine features of the image . A low intensity threshold was selected manually and applied to capture weak furrow canal links present in multinucleated cells . The resulting mask was closed to join disconnected furrow canal links , and thinned iteratively to capture all nuclei separations . Remaining disconnected furrow canal links were removed . ( 5 ) The preliminary furrow canal network mask resulting from step 2 above was thinned and added to the mask obtained in step 4 . This mask was dilated and inverted to generate a preliminary nuclei mask . Holes inside the mask were filled . ( 6 ) Finally , a high threshold was applied to the result of step 1 above; the resulting mask was inverted and then multiplied by the result of step 5 to eliminate boundary artifacts produced by thinning operations . For quantifications , levels of F-actin in furrow canals and numbers of Amphiphysin tubules were scored as previously reported [29] . The ratio of mononucleate cells to nuclei was generated by manually counting in two quadrants from a raw , single plane , surface view image ( quadrant size = 2835 µm2 ) collected at the furrow canals; and the mean was calculated per embryo . The percent of embryos displaying multinucleation was counted manually using raw , single plane , surface view images collected at the furrow canals , where an entire embryo side was visible ( ≥1500 nuclei assayed per embryo ) . For genotyping by FISH , maximum intensity projections were generated from image stacks , comprising ∼4 µm depth and ≥150 nuclei; and active sry-α transcription sites were manually counted . Embryos with a maximum of 0 , 1 , or 2 spots were scored as sry-α−/− , sry-α+/− , or sry-α+/+ ( wild-type , WT ) , respectively . For genotyping by immunofluorescence , images were collected at the same microscope settings , and Sry-α signal was scored as either present or absent . For RT-PCR , embryos were staged in halocarbon oil 27 under a dissecting microscope . Approximately 100 embryos per stage were homogenized in Trizol ( Invitrogen Inc . ) , and total RNA was extracted in phenol: chloroform ( 1∶1 ) . Total cDNA was made ( SuperScript III First-Strand Synthesis System , Invitrogen ) , and sequences amplified by PCR . For developmental expression profiles , primers were: actin42a-F , actin42a-R , sry-α-F , sry-α-R , spt-F , and spt-R ( for sequences see Table S3 ) . Following RNAi , primers were: actin42a-F , actin42a-R , sry-α-F2 , sry-α-R2 , spt-F2 , and spt-R2 ( for sequences see Table S3 ) . Samples were loaded on 1% agarose gels . For Westerns , 1 hour collections of embryos were incubated at room temperature for 0 , 1 , 2 , 3 , 4 , or 5 hours respectively . Per stage , 50–100 µl embryos were homogenized in 200 µl 0 . 05 M Tris pH 8 . 0 , 0 . 15 M KCl , 0 . 05M EDTA , 0 . 5% NP-40 , 1X protease inhibitor cocktail ( Halt Protease Inhibitor Cocktail , Thermo Scientific ) . For antibody characterization , 20 adult heads were homogenized in the same buffer . Following a 15 minute spin at 15 , 000 rpm at 4°C , the cytoplasmic fraction was collected and quantified ( Pierce BCA Protein Assay Kit , Thermo Scientific ) . Equal amounts of total protein were separated on 5–10% SDS-PAGE gels and transferred to nitrocellulose . Sry-α , Spt , HA and α-Tubulin were detected by 1∶5 mouse anti-Sry-α ( 1G10 , Developmental Studies Hybridoma Bank ) , 1∶500 guinea pig anti-Spt ( this paper ) , 1∶100 rat anti-HA ( Roche ) , and 1∶1000 rat anti-α-Tubulin ( T9026 , Sigma-Aldrich ) , respectively . The goat secondary antibodies were HRP conjugates used at 1∶10 , 000 ( Jackson ImmunoResearch ) . Approximately 50 pl of sry-α and spt double-stranded RNA was prepared as previously described [41] , with primers: sry-α-RNAi-F , sry-α-RNAi-R , spt-RNAi-F , and spt-RNAi-R ( for sequences see Table S3 ) , and injected into freshly laid embryos . Following incubation , mitotic cycle 13 embryos were mounted for imaging . RNAi controls were PBS buffer-injected embryos . Protein sequences were retrieved via UniProt ( uniprot . org ) , and alignments were generated using PROMALS3D ( prodata . swmed . edu/promals3d ) [37] . Phylogenetic trees were built with PhyML v3 . 0 . 1 ( pbil . univ-lyon1 . fr/software/seaview ) [67] , and branch supports were tested with the default aLRT SH-like option . In addition , for the small and large trees , bootstrap statistics were determined from 1000 and 500 iterations , respectively . Vinculin from the pre-metazoan M . brevicollis served as the outgroup [36] . For both alignments and trees , the same results were generated using either the entire Sry-α and Spt sequences , or the VH2 domain alone . For the roll call analysis , the D . melanogaster sequences for sry-α and spt were used as input for ( 1 ) a PSI-BLAST search of the refseq protein database on NCBI ( blast . ncbi . nlm . nih . gov ) ; and ( 2 ) a BLAST search of the UniProtKB database on UniProt . For those insects where only one paralog was present , we could not assign them as sry-α or spt based on sequence identity alone because the identity values were the same . Instead , we used the presence or absence of introns as the distinguishing factor: sry-α is intronless , whereas spt has introns in all Drosophilids . Thus , all insect proteins including introns were assigned as spt . | Every embryo develops under its own unique set of circumstances , with variable inputs coming from mother , father , and the environment . To then ensure a reliable outcome , mechanisms are built into development to buffer against challenges like genetic deficiency , maternal fever , alcohol exposure , etc . This buffering , called “robustness” , can be overwhelmed , ending in miscarriage , pre-mature birth , and structural and functional birth defects . Thus , we need to understand how robustness arises in order to define an embryo's susceptibilities to genetic background and environment; and to ultimately promote healthy reproduction . In this work we provide new insight into how morphogenesis , the process of tissue building in embryos , is made more robust . First , we show that early gene expression by the embryo , or zygote , supplements the stockpile of proteins already supplied by the mother to ensure the robustness of early morphogenesis . Specifically , our data suggests that a specific gene , sry-α , and its expression by the embryo at the maternal-to-zygotic transition , is a genetic adaptation with the sole function of making the first tissue building event in the fruit fly more robust . In addition , we show that the robustness of this morphogenetic event is promoted by mechanisms regulating the actin cytoskeleton . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2013 | The Maternal-to-Zygotic Transition Targets Actin to Promote Robustness during Morphogenesis |
The host's immune system plays a key role in modulating growth of pathogens and the intestinal microbiota in the gut . In particular , inflammatory bowel disorders and pathogen infections induce shifts of the resident commensal microbiota which can result in overgrowth of Enterobacteriaceae ( “inflammation-inflicted blooms” ) . Here , we investigated competition of the human pathogenic Salmonella enterica serovar Typhimurium strain SL1344 ( S . Tm ) and commensal E . coli in inflammation-inflicted blooms . S . Tm produces colicin Ib ( ColIb ) , which is a narrow-spectrum protein toxin active against related Enterobacteriaceae . Production of ColIb conferred a competitive advantage to S . Tm over sensitive E . coli strains in the inflamed gut . In contrast , an avirulent S . Tm mutant strain defective in triggering gut inflammation did not benefit from ColIb . Expression of ColIb ( cib ) is regulated by iron limitation and the SOS response . CirA , the cognate outer membrane receptor of ColIb on colicin-sensitive E . coli , is induced upon iron limitation . We demonstrate that growth in inflammation-induced blooms favours expression of both S . Tm ColIb and the receptor CirA , thereby fuelling ColIb dependent competition of S . Tm and commensal E . coli in the gut . In conclusion , this study uncovers a so-far unappreciated role of inflammation-inflicted blooms as an environment favouring ColIb-dependent competition of pathogenic and commensal representatives of the Enterobacteriaceae family .
Enteric Salmonella enterica serovar Typhimurium ( S . Tm ) infection is antagonized by a highly complex intestinal microbiota , a condition termed colonization resistance . Disruption of colonization resistance by oral antibiotic therapy , germfree state or an immature microbiota of low complexity results in increased susceptibility to oral infection with pathogens of the Enterobacteriaceae family [1] , [2] , [3] . In addition to the named conditions , inflammatory changes in the intestine induce dysbiosis and favour Enterobacteriaceae overgrowth [4] , [5] , [6] . Recently , we have shown that S . Tm-induced gut inflammation mediates parallel blooms of S . Tm and host-intrinsic commensal E . coli [7] , [8] . In these blooms , both bacteria can reach high densities ( >108 cfu/ml ) while the residual microbiota is outgrown [7] . Therefore , under inflammatory conditions , commensal E . coli are one of the main competitors of S . Tm . In the inflamed gut , environmental conditions encountered by bacteria vastly differ from the situation in the absence of inflammation . By resisting antimicrobial defences , utilizing iron acquisition systems and exploiting anaerobic electron acceptors ( e . g . NO3− , tetrathionate ) , Enterobacteriaceae can capitalize on inflammatory conditions [9] , [10] , [11] , [12] , [13] , [14] . Besides exploitative competition for resources , bacteria can directly antagonize one another by producing antimicrobials , such as bacteriocins . Bacteriocins produced by Enterobacteriaceae ( E . coli , Salmonella and relatives ) are termed colicins . They have a narrow spectrum of activity and act only against phylogenetically close relatives . Colicins kill by one of three general mechanisms: pore formation in the inner membrane , nuclease activity or interference with cell wall synthesis [15] . We have shown that growth of S . Tm in inflammation-induced blooms promotes horizontal transfer of the conjugative pCol1B9-plasmid ( in the following termed P2 ) to commensal E . coli strains [7] . P2 encodes the locus for colicin Ib ( ColIb ) production ( cib ) and immunity ( imm ) . In blooms , ColIb was shown to increase the fitness of S . Tm in competition with commensal , colicin-sensitive E . coli . However , it has remained unclear if ColIb only affords a benefit for S . Tm in inflammation-inflicted blooms or also in the absence of gut inflammation . ColIb belongs to the group of pore-forming colicins [16] . ColIb is closely related to ColIa and most of its structural/functional properties can be inferred from ColIa [17] . Free ColIa/b binds to the outer-membrane receptor CirA , a catecholate siderophore receptor , and traverses the outer membrane in a TonB-dependent way [18] , [19] , [20] . The exact mode of outer-membrane translocation is still unclear but a recent study suggests that two molecules of CirA are required , one for ColIa/ColIb binding , the second one for its translocation [18] , [21] . ColIa/ColIb forms a pore in the inner membrane of sensitive bacteria which leads to disruption of the electrochemical membrane gradient and consequent bacterial death [22] . ColIa/ColIb producers protect themselves by expression of the immunity protein ( imm ) which interferes with ColIa/ColIb action in the inner membrane [23] . In addition , resistance to ColIb can be gained by alterations of the cell-surface receptor CirA or mutations in the TonB-dependent import pathway [24] . ColIa/ColIb expression is regulated in a Fur- and LexA-dependent way [25] . The Fur protein binds FeII and thereby measures the intracellular FeII pool [26] . In the iron-bound state , Fur acts as a transcriptional repressor of iron-regulated promoters which is released under iron-limiting conditions . In addition , the ColIb gene cib is also repressed by the LexA protein , the regulator of the SOS response . The SOS response is launched when bacteria sense DNA damage ( e . g . DNA double strands breaks ) [27] . As a consequence , the RecA protease is activated and cleaves the LexA repressor protein which in turn activates an array of genes involved in DNA repair , survival , prophages and also colicins [15] , [28] . Interestingly , ColIa and ColIb are the only reported colicins which are controlled by both , Fur and LexA . This suggests that maximal expression of ColIa/b requires iron limitation and stress conditions . Interestingly , the majority of other colicins are only regulated by LexA and usually contain two LexA-binding sites in their promoter to ensure tight repression . Theoretical and experimental studies suggested that colicin producers have a competitive advantage over non-producers when colonizing the same ecological niche [29] . When directly tested in competition experiments with sensitive strains in animal models , colicin production only conferred a competitive advantage after weeks [30] , [31] . In some cases , colicin-production even conferred no detectable benefit despite colicin-dependent killing could be demonstrated under in vitro conditions [32] , [33] , [34] . The underlying reasons were attributed to colicin inactivation by proteases [33] or reduced colicin activity under anaerobic conditions [35] , [36] . In the S . Tm mouse colitis model ColIb conferred an overt fitness benefit for S . Tm in competition against a colicin-sensitive E . coli strain [7] . This was somewhat surprising , considering the lack of an apparent competitive advantage of colicin producers in the reports mentioned above . Although different experimental setups were used in previous studies , a key difference to our study was the lack of concomitant gut inflammation . Therefore we reasoned that the inflammatory response may somehow promote ColIb dependent competition of S . Tm and E . coli . We tested this idea and analyzed ColIb-dependent S . Tm and E . coli competition under normal and inflammatory conditions in the mouse colitis model . Our experiments revealed that the inflammatory response creates conditions that potentiate the effects of colicins , both by increasing their production and by mediating susceptibility of the competitor . This finding has implications for colicin ecology and points out the importance of colicins as fitness factor for bacterial competition in Enterobacterial blooms in the inflamed gut .
ColIb confers a fitness benefit to Salmonella Typhimurium ( S . Tm ) over colicin-sensitive E . coli strains [7] . In vitro , the E . coli Nissle ( EcNissle ) strain shows intermediate susceptibility to ColIb ( turbid inhibition zone ) , while the K-12 strain E . coli MG1655 ( EcMG1655 ) is highly sensitive ( clear inhibition zone; Figure S1A ) . For this reason we selected EcMG1655 for our initial experiments . First we tested , if the growth benefit of S . Tm over EcMG1655 is ColIb-dependent . We performed the co-infection experiments with S . Tm strains and EcMG1655 in the streptomycin Salmonella mouse colitis model [2] . Here , we used gnotobiotic mice colonized with a low-complexity microbiota ( LCM ) lacking any kind of Enterobacteriaceae which may interfere with the experiment ( i . e . by the production of other colicins ) . Further , since the S . Tm P2-plasmid is highly mobile and rapidly transferred to co-colonizing E . coli strains in the gut [7] , all S . Tm strains carried an additional mutation in the origin of transfer of P2 ( ΔoriT ) to prevent conjugation ( Table 1 ) . LCM mice pre-treated with streptomycin were co-infected with 1∶1 mixtures of EcMG1655 and either ColIb-producing ( S . TmΔoriT ) or ColIb-deficient strains ( S . TmΔoriT Δcib ) . EcMG1655 was efficiently outcompeted by S . TmΔoriT but not by the ColIb-deficient mutant ( Figure 1A , B ) . Both S . Tm strains induced similar degrees of gut inflammation by day 4 p . i . ( Figure 1CD ) . This confirms that the competitive advantage of S . Tm over EcMG1655 in the inflamed intestine is for the most part ColIb-dependent . Next , we tested if an avirulent strain ( S . TmΔoriT avir ) defective in triggering an inflammatory response due to the absence of functional type three secretion systems [37] would also benefit from ColIb . Interestingly , in the absence of gut inflammation , S . TmΔoriT avir did not out-compete EcMG1655 ( Figure 1 ) . Previous studies on colicin-dependent bacterial competition were using conventional mice . To this end we aimed to verify that our key findings in gnotobiotic mice were also reproducible in a more “natural” mouse model . To this end , we used the streptomycin-treated mice with a conventional complex microbiota . We used EcNissle for the co-infection experiments as in contrast to LCM mice , EcMG1655 only poorly colonized conventional streptomycin-treated , S . TmΔoriT infected mice ( not shown ) . We set up four different experimental groups of streptomycin-treated mice ( Figure 2 ) . One group was co-infected with a virulent colicin-producing , the other with a virulent colicin–deficient S . Tm strain ( S . TmΔoriT or S . TmΔoriT Δcib , respectively ) and EcNissle . Both groups developed strong Salmonella-induced gut inflammation upon infection by day 4 p . i . ( Figure 2E , F ) . The other two groups were either co-infected with avirulent ColIb-producing S . Tm ( S . TmΔoriT avir ) and EcNissle , or avirulent ColIb-deficient S . Tm ( S . TmΔP2 avir ) and EcNissle . The latter two groups did not develop gut inflammation within 4 days of infection ( Figure 2E , F ) . In the presence of inflammation , virulent ColIb-producing S . TmΔoriT grew up to similar numbers as EcNissle ( to ∼108 cfu/g ) while ColIb-deficient S . TmΔoriT Δcib was outcompeted by EcNissle ( Figure 2B ) . This difference is also reflected in the competitive index ( Figure 2D ) . In the absence of inflammation , both co-infecting strains ( ColIb-producing S . TmΔoriT avir and EcNissle ) and ( ColIb-deficient S . TmΔP2 avir and EcNissle ) colonized well at day 1 post infection ( Figure 2AC ) but were out-competed ( to ∼107 cfu/g ) by the complex conventional microbiota , which recovers by day 5 after streptomycin-treatment ( Figure 2B ) . No benefit of ColIb-production was observed for the avirulent Salmonella strain . The absolute ratio of Salmonella/EcNissle is different when compared to Salmonella/EcMG1655 in LCM mice ( Figure 1 ) . We reason that this is due to strain-specific differences between EcMG1655 and EcNissle as well as due to differences between the gnotobiotic and complex gut microbiota . This data verified that S . Tm/EcNissle competition is only ColIb-dependent in the presence of gut inflammation . In conclusion , these experiments prompted us to hypothesize that , in the normal non-inflamed gut , either ColIb expression by S . Tm was down-regulated or susceptibility of the Ec strains to ColIb was decreased . To further address the mechanism of colicin-dependent competition in the inflammation-induced blooms we investigated the regulation of cib expression in S . Tm as well as the susceptibility of Ec to ColIb in detail . The cib promoter region contains binding sites for the transcriptional repressors Fur and LexA ( Figure 3A; Figure S2 ) . We generated a cib promoter firefly-luciferase ( luc ) -reporter construct as well as an affinity-purified polyclonal rabbit-α-ColIb antiserum to analyze the regulation of ColIb expression . ColIb expression was strongly up-regulated upon induction of the SOS response by the antibiotic mitomycin C ( 0 . 25 µg/ml ) as confirmed by luc-reporter assays and immunoblot ( Figure 3B–D ) . Depletion of Fe ( III ) from culture media by chelation with 100 µM diethylenetriaminepentaacetic acid ( DTPA ) [38] had a comparable inductive effect on ColIb production . Supplementation of both , mitomycin C and DTPA lead to maximal induction of ColIb production and secretion . This result confirmed that ColIb is de-repressed in response to SOS signals and iron starvation . The outer membrane protein CirA is the receptor ColIa and ColIb [18] , [19] , [20] . Expression of cirA is under negative control of Fur [39] ( Figure S3 ) . To confirm FeIII-dependent regulation of the cirA expression in EcMG1655 , we generated a cirA promoter-luc-reporter as well as a polyclonal rabbit-α-CirA antiserum . CirA expression was strongly upregulated in LB media upon addition of 100 µM DTPA but not by mitomycin C as confirmed by luciferase assay and immunoblot ( Figure 4A , B ) . This confirmed that EcMG1655 cirA was de-repressed in response to FeIII-starvation . Next , we tested if cirA-expression correlated with sensitivity to ColIb-mediated killing . As expected , EcMG1655 ΔcirA was resistant to ColIb ( Figure 5A ) . This phenotype was complemented by expressing cirA on a plasmid in EcMG1655 ΔcirA ( Figure S1C ) . We further investigated , whether successive iron depletion would increase ColIb sensitivity of EcMG1655 . To this end , we performed ColIb killing assays of EcMG1655 in M9 minimal media with different concentrations of FeCl3 using the same amounts of purified recombinant ColIb . Indeed , EcMG1655 was most sensitive to ColIb in M9 without FeCl3 addition and became less susceptible upon FeCl3 supplementation ( Figure 5A ) . This correlated with increased cirA expression under this condition , as determined by immunoblot ( Figure 5B ) . Thus , sensitivity of EcMG1655 to ColIb increases with elevated cirA expression . To underscore the importance of environmental conditions for ColIb-dependent competition of S . TmΔoriT and EcMG1655 , we performed in vitro co-culture experiments . We followed growth of S . TmΔoriT and EcMG1655 and the respective mutants ( S . TmΔoriT Δcib and EcMG1655 ΔcirA ) in co-cultures under different conditions ( Figure 6 ) . In the absence of supplements , S . TmΔoriT and EcMG1655 grew at similar rate . S . TmΔoriT outcompeted EcMG1655 by ∼7-fold after 8 h ( mean titer S . TmΔoriT: 1 . 7×109 cfu/ml and EcMG1655: 2 . 3×108 cfu/ml; Figure 6A ) . In contrast , EcMG1655 was outcompeted by several orders of magnitude after 8 h when either DTPA ( 7×107-fold ) , mitomycin C ( 1×104-fold ) or both supplements ( 6×106-fold ) were added to the co-culture ( Figure 6D , G , J ) . S . Tm overgrowth in vitro was indeed ColIb-dependent , as no killing was observed in the absence of ColIb ( S . TmΔoriT Δcib ) or CirA ( EcMG1655 ΔcirA ) ( Figure 5B , C , E–L ) . Of note , addition of 100 µM DTPA led to more pronounced killing of EcMG1655 than mitomycin C with comparable amounts of ColIb ( Figure 3D ) , suggesting that iron depletion has a greater impact on ColIb-dependent competition ( i . e . by enhancing EcMG1655 cirA expression and thereby its susceptibility to ColIb ) . The mutant phenotypes of the S . Tm cib mutant as well as the EcMG1655 cirA mutant were complemented using a plasmid-based complementation approach ( Figure S1 , Figure S4 and Figure S5 ) . The in vitro co-culture experiments of S . Tm and EcMG1655 showed that under iron excess conditions and in the absence of triggers of the SOS response , ColIb confers little to no detectable benefit to S . Tm . In contrast , under FeIII-limitation or in the presence of stressors , S . Tm outcompetes E . coli , which is dependent on ColIb production by S . Tm and cirA expression by E . coli . Based on these data , we reasoned that ColIb-dependent competition of S . Tm and EcMG1655 in the inflamed gut could indeed be due to increased production of ColIb by S . Tm , or up-regulation of the colicin receptor CirA by E . coli or both . To address this , we analyzed expression of EcMG1655 cirA and S . Tm ColIb ( cib ) in the streptomycin mouse colitis model using firefly-luciferase reporter-constructs . To generate inflammatory and non-inflammatory conditions in the intestine , LCM mice were infected either with virulent ( S . TmΔoriT or S . Tmwt; +inflammation ) or avirulent ( S . TmΔoriT avir or S . Tmavir; -inflammation ) Salmonella strains , respectively . To investigate regulation of EcMG1655 cirA expression , mice were co-infected with EcMG1655 carrying the pcirA-luc-reporter plasmid ( Figure 7A , C ) . To investigate regulation of S . Tm cib expression , mice were co-infected with S . TmΔoriT avir carrying the pcib-luc-reporter plasmid ( Figure 7B , D ) . The experiments showed that luciferase levels for both the pcirA-luc- and the pcib-luc-reporters were significantly increased in the inflamed intestine ( Figure 7A–D ) . Therefore , these data are in agreement with our initial hypothesis and suggest that in response to gut inflammation , ColIb production by S . Tm and susceptibility of EcMG1655 to ColIb are increased . This explains how inflammation fuels ColIb dependent competition of S . Tm and commensal E . coli ( Figure 8 ) .
Inflammatory conditions in the gut shape gut microbial community structure and are characterized by an increased prevalence of facultative anaerobic bacteria ( “blooms” ) , in particular members of the Enterobacteriaceae family [4] , [8] , [40] . Commensal E . coli strains can hitchhike gut inflammation induced by inflammatory bowel diseases or enteric pathogens ( S . Tm or Citrobacter rodentium ) [5] , [7] , [12] . Commensal and pathogenic representatives of the Enterobacteriaceae efficiently dwell in blooms as they can exploit resources with increased abundance in the inflamed gut ( ethanolamine , the anaerobic electron-acceptors tetrathionate and nitrate ) [11] , [12] , [41] . Up to 15 different E . coli strains can be detected in one individual human gut ecosystem [42] . Thus , successful competition for resources should be of major importance for bacteria in order to come out on top and eventually benefit from an episode of gut inflammation [8] . Here , we propose that colicins such as ColIb are effective means to fight off competing bacteria , particularly in inflammation-induced enterobacterial blooms . Which environmental cues promote ColIb-dependent bacterial competition in blooms ? As an immediate inflammatory defence reaction , the host deprives potential invading pathogens of nutritionally-required iron . Neutrophils release iron-sequestering lactoferrin upon degranulation [43] . Further , lipocalin 2 ( LCN2 ) , an antimicrobial protein binding the bacterial siderophore enterochelin [44] is abundantly expressed by neutrophils and intestinal epithelial cells . LCN2 is produced in response to infection with wildtype but not avirulent S . Tm strains in the streptomycin-colitis model [9] . S . Tm overcomes LCN2-imposed inhibition by production of salmochelin [45] , [46] , a siderophore resistant to binding by LCN2 . Salmochelin is produced in a Fur-dependent manner in response to iron limitation and upregulated by S . Tm thriving in the inflamed gut [10] . Therefore , host-mediated iron restriction in the inflamed gut poses an environmental cue for inducing ColIb production by S . Tm . Other cues for ColIb production are compounds triggering the SOS response . Yet , the exact source and identity of those compounds in the inflamed gut is ill-defined . The SOS response in Enterobacteriaceae can be triggered by antibiotics which directly affect DNA integrity ( mitomycin C ) , DNA replication via the DNA gyrase ( quinolones ) or induce membrane stress ( β-lactams ) [28] , [47] . Further , oxidative stress induced by reactive oxygen species ( ROS ) , superoxides , H2O2 or free radicals formed by UV-light induce DNA damage and thereby the SOS response [28] , [48] . Neutrophils infiltrating the intestinal lumen in response to Salmonella infection as well as activated iNOS-producing epithelial cells likely represent a source of free radicals and potential inducer of the SOS response [40] , [49] . Recently , it was demonstrated that E . coli indeed upregulates stress-induced proteins such as GroL , RecA , YggE and the Fe-S cluster repair protein NfuA in the inflamed gut [50] . Not only was the expression of cib increased in the inflamed gut , but also the corresponding surface receptor cirA . CirA is the receptor for monomers , dimers and linear trimers of 2 , 3-dihydroxybenzoylserine , breakdown products of enterochelin [51] . Further , it was shown , that cirA mutants are attenuated in the uptake of monomeric catechol and its analogues [52] . In order to avoid iron overload and its negative consequences ( e . g . the formation of hydroxyl-radicals ) , iron-uptake systems are tightly controlled at the transcriptional level and only de-repressed under iron-limiting conditions . In vitro , susceptibility to ColIb of E . coli was drastically increased under FeIII-depleted conditions , which correlated with elevated CirA protein levels . In the same way , the in vitro competition experiments suggest that S . Tm benefits most from ColIb production under iron-limiting conditions ( 7×107-fold ) . The growth advantage was lower in the presence of mitomycin C ( 1×104-fold ) or under both iron limitation and mitomycin C ( 6×106-fold ) . Of note , the amount of free S . Tm ColIb in the medium was even higher in the presence of mitomycin C , than with addition of DTPA only ( Figure 3D ) , supporting the idea , that the expression level of cirA has a higher impact on the competition , compared to that of ColIb ( Figure 6D , G ) . Hitherto , the underlying mechanism of ColIb-release or secretion triggered by the SOS response is not known . In conclusion , we reason that increased cirA expression and consequential high susceptibility of E . coli to ColIb may explain for the most part colicin-dependent competition in blooms ( Figure 8 ) . Interestingly , several other colicins parasitize siderophore outer membrane receptors , which are all under control of the Fur-regulon . ColM binds to the ferrichrome receptor FhuA while ColB and ColD bind to FepA , the receptor for enterobactin [15] . This suggests that increased sensitivity to colicin-mediated killing under iron depletion may also apply for other colicins binding to TonB-dependent outer-membrane transporters . Previously , it was shown that killing of susceptible bacteria by pyocin , a bacteriocin produced by Pseudomomas spp . , is increased under iron-limited conditions [53] . Pyocin binds the ferri-pyoverdine receptor FpvA , which is controlled by Fur . In summary , physiological changes of the murine intestine upon Salmonella-induced colitis are likely to provide the environmental cues required for upregulation of both , ColIb and its receptor CirA . Yet , we cannot rule out that other physiological parameters altered in the inflamed gut also contribute to the observed phenotype . So far , it is unclear if expression of other types of colicins as well as other Enterobacteriaceae-derived bacteriocins ( microcins , pyocins , klebicins ) would be upregulated in the inflamed gut . The majority of these bacteriocins is only under control of the SOS response and repressed by LexA and not regulated in a Fur-dependent fashion . Thus , it remains to be shown if the principle of colicin-colicin-receptor upregulation in the inflamed gut also applies to other bacteriocins and their respective receptors . Supposedly , colicins play a major role in mediating bacterial population dynamics [54] . Superior fitness of colicinogenic over sensitive strains in vivo could so far only be demonstrated in few studies performing long-term competitive infection experiments ( ≥12 weeks ) [30] , [31] . In contrast , a number of other studies reported that high bacteriocidal activity against closely related , sensitive strains observed in vitro could not be recapitulated in in vivo experiments ( see below ) . Our data presented in this paper might explain this puzzling observation: we suggest that the fitness benefit of colicin production strongly depends on the environmental conditions prevailing in the gut . Under normal conditions , colicin expression and expression of their cognate receptors may not be stimulated enough to induce colicin-dependent inhibition of the sensitive strain . In the absence of gut inflammation , S . Tm did not benefit from ColIb in competition with E . coli . Likewise , competition experiments in germfree mice with a colicin-producing E . coli and a sensitive strain resulted in equal colonization levels of both strains over weeks ( no inflammation induced ) [32] , [33] , [34] and similar results were obtained for other strains and colicins [55] , [56] . The underlying reasons for the absence of an overt fitness-benefit of colicin production were attributed to colicin inactivation by intestinal proteases [33] , acquisition of colicin-resistances [55] or absent colicin activity under anaerobic conditions [35] , [36] . Our study suggests that absence of inflammatory conditions might be an additional explanation . Colicin production is a common trait in E . coli populations [54] . On average , 30% of natural E . coli populations produce one or more colicins [57] . Many experimental and theoretical studies have addressed the ecological consequences of colicin production in bacterial populations [58] , [59] . In general , it is assumed , that colicins afford a competitive advantage to the strain producing it . However , the producer pays a fitness cost due to the higher metabolic load of colicin synthesis as well as lethality of production ( e . g . lysis-mediated colicin release ) . Bacteria have partially overcome this limitation by applying the principle of ‘division of labor’ . In a population of colicin producers , only a small fraction of bacteria are induced to produce the colicin ( = phenotypic heterogeneity ) [60] , [61] . In recA negative strains decreased frequencies of colicin producers were observed , suggesting that the rate of colicin production is regulated by the SOS response [60] . This strategy seems to be evolutionary successful , as colicins released by the subpopulation serve as a “common good” for the whole population and secure propagation of the shared genotype . Nevertheless , colicin expression needs to be tightly controlled to ensure , that the fraction of producers is kept at low rates under conditions , when colicin is not required . Those conditions include the lack of stress , nutrient starvation but also the absence of any direct competitors . Thus , we assume that colicin production of a bacterial population should be confined to environmental niches which are characterized by high density and diversity of competing E . coli and close relatives . E . coli titers in the mammalian gut lumen are rather low as the intestinal tract is dominated by strictly anaerobic bacteria [62] . E . coli predominantly colonizes the mucus layer of the large intestine where it thrives on mucin-derived sugars [63] . Thus , the intestinal mucus layer is one highly competitive environment for E . coli and colicins may be beneficial for competing for the preferred limiting substrates [64] . In contrast , we identify inflammation-induced blooms as an alternative niche for colicin-dependent Enterobacterial competition ( Figure 8 ) . Enterobacterial blooms can contain multiple closely-related species at high concentrations which likely compete for the same resources . Under this highly competitive situation , the chances are increased that colicin-sensitive competitors are present at high numbers . Thus , the bacteria may benefit hugely from colicin production under this environmental condition . Moreover , the population size of the colicin-producer is large enough to tolerate loss of a fraction of the population due to suicidal colicin production . In summary , the results presented here provide evidence that intestinal inflammation drives colicin-dependent competition by bacteria of the Enterobacteriaceae family . These findings shed new light on the role of colicins as important fitness factors providing a competitive advantage for growth in Enterobacterial blooms .
All animal experiments were approved by the Regierung von Oberbayern and the Kantonales Veterinäramt Zürich performed according to national German and Swiss guidelines ( Deutsches TschG; Schweizer Kantonale TschV ) . The permit no . 55 . 2-1-54-2532-49-11 ( Germany ) and 201/2007 ( Switzerland ) . All mice used in the study were on C57BL/6J background and bred at the Rodent Center , ETH Zürich and the Max-von-Pettenkofer Institute , LMU Munich under SPF conditions in individually ventilated cages . Low-complexity microbiota ( LCM ) mice were generated by associating germfree mice with members of the Altered Schaedler flora [65] as described previously [66] . Conventional SPF C57BL/6J mice were purchased from Janvier , Le Genest Saint Isle . For infections , conventional and LCM mice were pretreated with streptomycin and infected by gavage with 5×107 cfu S . Tm or mixtures of S . Tm and E . coli as described [7] . For in vivo luciferase-assays , LCM mice were pretreated with ampicillin ( 20 mg/animal 24 h prior to infection ) . Live bacterial loads in the cecal content were determined by plating on MacConkey-agar ( Roth ) with respective supplements ( streptomycin 100 µg/ml; kanamycin 30 µg/ml; chloramphenicol 30 µg/ml; ampicillin 100 µg/ml and tetracycline 12 , 5 µg/ml ) . Histology of the cecum was done at necropsy . Cecum tissue was embedded in O . C . T . ( Sakura , Torrance ) and flash frozen . Cryosections ( 5 µm ) of the cecal tissue were H&E-stained and scored as described in detail in [6] . The parameters submucosal edema , PMN infiltration , loss of goblet cells and epithelial damage were scored according to the severity of inflammatory symptoms yielding a total score of 0–13 points . For infections , E . coli and S . Tm strains were grown as described [67] . Briefly , cultures in LB supplemented with 0 . 3M NaCl were inoculated with 2–3 bacterial colonies from plates . Bacteria were grown over night for 12 h and subcultures ( 1∶20 ) for an additional 4 h . Bacteria were mixed ( as indicated ) washed in PBS and applied to the mice in a total volume of 50 µl by gavage . Bacterial strains and plasmids used in this study are listed in Table 1 . EcMG1655 ΔcirA ( LPN2 ) was constructed using the lambda Red recombinase system as described using pKD4 as template for the kanamycin-resistance gene including the FRT-sites [68] . Briefly , EcMG1655 was transformed with the plasmid pKD46 . The kanamycin resistance cassette from plasmid pKD4 was amplified by PCR using primers K12ΔcirA_Fwd/K12ΔcirA_Re ( Table S1 ) . Correct recombination was verified by PCR using primers cirA-up/cirA-down and cirA-up/cirA-d1 ( Table S1 ) . S . Tmavir ΔoriT ( LPN5 ) was generated by P22-transduction of the ΔoriTnikA::cat allele from M1407 into M557 [37] . Correct insertion was verified by PCR using primers ΔoriTnikArev_val , ΔoriTnikA_val . The P2 plasmid was cured from S . Tm ΔinvG; sseD::aphT cured as described previously [7] . For the generation of c-terminal CirA-His-tag fusion , the open reading frame of cirA was amplified from E coli Nissle genomic DNA by PCR , using Fow_cirA_NheI and Re_cirA_XhoI primers ( Table S1 ) and cloned into pET-24c ( Novagen ) via NheI and XhoI to yield pLPN13 . For the generation of c-terminal ColIb-His-tag fusion , the ColIb gene cib was amplified from S . Tmwt genomic DNA by PCR , using primers Fow_colicin_NheI and Re_colicin_XhoI ( Table S1 ) and cloned into pET-24c via NheI and XhoI to yield pLPN14 . To generate pLPN1 , the cirA promoter was amplified from E . coli Nissle using pcirA-BamHI , pcirA-XbaI ( Table S1 ) and inserted in BamHI and XbaI digested pM979 [69] . For generation of pM1437 , the cib promoter from E . coli8178 was amplified using pColIb-XbaI , pColIb-BamHI ( Table S1 ) and inserted in pM968 [69] via restriction with XbaI and BamHI . To generate pLPN15 and pLPN16 , the firefly-luciferase gene luc from pLB02 [70] was amplified with Luc-for-BamHI and Luc-rev-HindIII primers ( Table S1 ) and inserted into BamHI/HindIII digested pLPN1 or pM1437 , respectively . Primers pWSK29-Gbs-for and pWSK29-Gbs-rev were used in a PCR with pWSK29 [71] as a template to amplify the low-copy-number plasmid ( Table S1 ) . Primers CirA-pWSK29-Gbs-for and CirA-pWSK29-Gbs-rev were used in a PCR with chromosomal DNA of EcMG1655 as a template to amplify cirA including its natural promoter . Primers Cib-Imm-pWSK29-Gbs-for and Cib-Imm-pWSK29-Gbs-rev were used in a PCR with S . Typhimurium strain SL1344 plasmid pCol1B9_SL1344 as a template to amplify the cib/imm locus including both natural promoters . The pWSK29 PCR fragment was combined with the cirA or cib/imm fragment in a Gibson assembly reaction [72] . Four microliters of the Gibson assembly mix were transformed into chemically competent E . coli Mach1 T1 cells ( Life Technologies ) . Constructs were verified using colony PCR , restriction analysis and sequencing . For annotation of transcription factor binding sites ( Fur and LexA regulon ) , all known transcription factor binding sites of each family one were taken from RegulonDB ( version 8 . 0 ) [73] and a binding motif was created using MEME [74] . The nucleotide sequences of the cib ( S . Tm SL1344; EMBL accession no . FQ312003 ) and cirA promoter regions ( E . coli MG1655 genome accession no . NC_000913 . 2 ) were searched for the computed MEME binding site motifs using MAST [74] . For expression of ColIb-His , we used E . coli BL21 transformed with pC831-2 ( expression of the ColIb immunity protein gene imm [7] ) and pLPN14 . For expression of cirA-His we used E . coli BL21 transformed with pLPN13 ( Table S1 ) . Over-night cultures of bacteria , grown at 37°C , 180 rpm in Luria-Bertani ( LB ) medium containing antibiotics were used for inoculation of subcultures ( dilution 1∶20 ) . At OD600 between 0 . 6–0 . 8 the subcultures was induced with 0 . 1 mM isopropyl β –D-thiogalactopyranoside ( IPTG ) and incubated for additional 4 h at 37°C , 180 rpm . Bacteria were harvested ( 4 , 500 rpm , 30 min at 4°C ) , resuspended in 40 ml 1×PBS and spun down at 5 , 000 rpm , 20 min at 4°C . The pellet was frozen at −20°C . Thereafter , the pellet was thawed and resuspended in 25 ml loading buffer ( 40 mM Na2HPO4 , 0 . 3M NaCl , 5 mM Imidazol , pH 7 . 8 ) , supplemented with 2 mM phenylmethylsulfonyl fluoride ( PMSF ) and benzonase nuclease ( Novagen ) . Bacteria were lysed in the French Press ( 1 , 000 PSI ) and the lysate was filtered ( 0 . 22 µm ) . Further , the lysate was loaded on a 5 ml HisTrap column ( GE Healthcare ) , and purified using the ÄKTA system ( GE Healthcare ) . ColIb-His was eluted with 5 mM Imidazole . The fractions containing the protein were desalted on a 5 ml HiTrap desalting column ( GE Healthcare ) , using the ÄKTA system and exchange buffer ( 20 mM Na2HPO4 , 100 mM NaCl , pH 7 . 4 ) . CirA-His was purified as outlined above for ColIb-His , but with following exceptions: loading buffer for the French Press ( 8M Urea , 40 mM Na2HPO4 , 0 . 3M NaCl , 5 mM Imidazol , pH 7 . 8 , 2 mM PMSF , and Benzonase nuclease ) : loading buffer for HisTrap column ( 6M Urea , 40 mM Na2HPO4 , 0 . 3M NaCl , 5 mM Imidazol , pH 7 . 8 ) ; exchange buffer ( 4M Urea , 20 mM Na2HPO4 , 100 mM NaCl , pH 7 . 4 ) . Rabbit antisera against ColIb-His and CirA-His were raised using standard protocols ( Pineda Antikörper-Service , Berlin , Germany ) . 6 mg/ml ColIb-His ( in 20 mM Na2HPO4 , 100 mM NaCl , pH 7 . 4 ) and 6 mg/ml CirA-His ( in 20 mM Na2HPO4 , 100 mM NaCl , 4M Urea , pH 7 . 4 ) were used for immunization . Control and immune serum were received from bleedings day 61 , 90 and 135 post immunization . Affinity purification of polyclonal rabbit α-ColIb-His antiserum was done using the Aminolink kit ( Thermo Scientific ) following the manufacturer's protocol with some minor modifications . For ColIb-His , PBS was used as binding/wash buffer and 1M Glycine , pH 2 . 7 was used as elution buffer . ColIb-His ( stored in 20 mM Na2HPO4 , 100 mM NaCl , pH 7 . 4 ) was added to the binding buffer at 1∶3 ratio . Desalting of the affinity-purified rabbit-α-ColIb-His antiserum was done using PD-10 desalting columns with PBS as exchange buffer ( GE Healthcare ) . For CirA-His , His-tagged CirA ( 20 mM Na2HPO4 , 100 mM NaCl , 4M Urea , pH 7 . 4 ) was dialyzed against PBS using 5 ml Zebra Spin desalting columns ( Thermo scientific ) shortly before coupling to the column . Coupling was done with CirA-His ( in PBS ) and coupling buffer supplemented with 4M Urea ( MP Biomedicals ) . Desalting of the affinity-purified rabbit-α-CirA-His was done with Zebra Spin desalting columns with PBS containing 0 . 05% sodium azide ( Merk ) . Purified antisera were stored at −80°C with addition of sodium azide to 0 . 01% . For measuring colicin production and -sensitivity , the colicin-producing strain was grown o . n . as small spot ( ø5 mm ) on LB agar containing 0 . 25 µg/ml mitomycin C ( Roth ) . The plate was overlaid with the tester strain in top-agar ( 0 . 75% agar ) . Growth of the tester strain was analyzed after 24 h . Formation of an inhibition zone ( halo ) around the producer indicated production of colicin and sensitivity of the tester strain . For determining ColIb sensitivity dependent on the FeIII concentration , the assay was modified accordingly . Starter cultures of E . coli in 3 ml M9 medium ( 40 mM Na2HPO4×2H2O , 20 mM KH2PO4 , 9 mM NaCl , 2 g/L NH4Cl , 1 mM MgSO4 , 100 µM CaCl2 , 2 g/l D-glucose , 10 mg/l thiamine , 500 mg/l histidine ) were grown for 10 h and used for inoculation ( 1∶20 ) of 2 ml M9 medium , supplemented with 1 µM , 10 µM , 0 . 1 mM or 1 mM FeCl3 and grown for 12 h . From each subculture , 50 µl was added to 5 ml 50°C 0 . 7% M9 top-agar ( 0 . 75% agar ) , which was used to overlay M9 agar plates . Further , antimicrobial susceptibility test discs ( Oxoid ) were laid on each plate and supplemented with 8 µl 1 . 3 mg/ml recombinant His-tagged ColIb . The plates were incubated over-night at 37°C . Bacteria were grown in a starter culture in LB or M9 media and used for inoculation of subcultures ( 1∶20 ) , except of in vitro co-cultures , where subcultures were inculcated to an OD600 of 0 . 05 for each strain . Following supplements were used: 0 . 25 µg/ml mitomycin C ( Roth ) ; 100 µM diethylenetriaminepentaacetic acid ( DTPA; Sigma ) , 1 µM , 10 µM , 0 . 1 mM or 1 mM FeCl3 ( Sigma ) . All cultures were grown at 37°C on a wheel rotor , except of in vitro co-cultures , where subcultures were grown in Erlenmeyer-flasks in a shaker at 200 rpm . Luciferase assays were performed as described [75] . Briefly , overnight cultures ( 3 ml LB , 100 µg/ml ampicillin ) were grown for 12 h and used for 1∶20 inoculation of 3 ml subcultures ( LB , 100 µg/ml ampicillin with respective supplements ) and grown for 4 h . From each subculture , 250 µl ( of an OD600 of 1 ) was spun down for 5 min , 14 , 000 rpm , 4°C . The supernatant was removed and the bacterial pellet was frozen at −80°C for 1 h . Further , the pellet was thawn and resuspended in 500 µl lysis buffer ( 100 mM K2HPO4/KH2PO4 buffer , pH 7 . 8 , 2 mM EDTA , 1% Triton X-100 , 5 mg/ml BSA , 1 mM DTT and 5 mg/ml lysozyme ) and incubated for 15 min at room temperature while vortexing every 3 minutes . Bacterial lysates ( 25 µl ) were transferred in 96-well plates ( white; Thermo scientific ) and 50 µl luciferase reagent [1 mM ( MgCO3 ) 4Mg ( OH ) 2×5H2O , 20 mM tricine , 0 . 1M EDTA , 470 µM D ( − ) luciferin ( Sigma ) , 530 µM Mg-ATP ( Sigma ) , 125 µM glycylglycine ( Sigma ) , 270 µM Li3-coenzym A ( Sigma ) , 33 mM DTT] was added to each well . Luminescence was measured using a FLUOstar Optima plate reader ( BMG Labtech ) . For luciferase assay from bacteria extracted from cecum content , the cecum content was harvested from infected mice and shortly stored on ice . The cecum content was resuspended in 500 µl PBS ( 0 . 1% tergitol ) and mixed in a tissue-lyser ( Qiagen; 5 min; 50 Hz ) . Further , the cecum content was filtered through a 40 µm cell-sieve ( Milian ) . Samples were taken to determine the cfu/ml of the reporter strain by plating on MacConkey agar with respective antibiotics . A defined volume ( i . e . 900 µl ) was pelleted at 4°C , 2 min , 14 , 000 rpm . The supernatant was removed and the pellet was frozen in dry ice and stored at −80°C . The samples were then thawn and processed as described above . Only values above detection limit ( control cecum content ) were considered . The relative luminescence units ( rlu ) per cfu luciferase-reporter strain were calculated . Overnight cultures of 3 ml M9 medium , grown for 12 h were used for inoculation ( 1∶20 ) of 2 ml M9 medium supplemented with 1 µM , 10 µM , 0 . 1 mM or 1 mM FeCl3 subculture , grown for 7 h . Starter culture of 3 ml LB , grown for 12 h was used for inoculation of 3 ml LB with supplements grown for 4 h . From each subculture , 250 µl ( for an OD600 of 1 ) was taken , spun down at 4°C , 10 min , 10 , 000 rpm . The supernatant was removed and bacterial pellet was frozen in liquid nitrogen and thawn at room temperature for 15 min ( repeated three times ) , resuspended in 100 µl lysis buffer ( 50 mM Tris , pH 7 . 5 , 150 mM NaCl , 5 mM EDTA , 0 . 25% nonidet P-40 , 1 mg/ml lysozyme ) and incubated in thermomixer at 550 rpm at 23°C , for 1 h and thereafter spun down at 4°C , 20 min , 14 000 rpm . Total protein was quantified in the lysate using protein assay reagent ( BioRad ) . Further , bacterial lysate was added to protein loading buffer ( 50 mM Tris , 100 mM DTT , 2% SDS , 0 . 1% bromphenolblue , 10% glycerol ) and incubated for 10 min at 95°C . For supernatant fractions 500 µl ( for an OD600 of 1 ) of the subculture were spun down twice , supernatant was added to 5× protein loading buffer and incubated for 10 min at 95°C . Proteins were separated by SDS gel electrophoresis [76] . Proteins were transferred onto a nitrocellulose membrane ( GE Healthcare ) at 300 mA for 2 h . The membrane was blocked in PBS ( 0 . 1% tween; 5% milk powder ) and probed with antisera ( affinity-purified α-CirA-His ( 1∶50 ) or affinity-purified α-ColIb-His ( 1∶500 ) . Goat-α-rabbit-HRP ( GE-Healthcare ) was used as secondary antibody . For detection of E . coli and S . Tm DnaK , the mouse monoclonal α-E . coli DnaK antibody ( mAb 8E2/2; Enzo Life Sciences ) and a secondary goat-α-mouse-HRP ( Sigma ) was used . Blots were developed with ECL detection system ( GE Healthcare ) . Statistical analysis was performed using the exact Mann-Whitney U Test ( Graphpad Prism Version 5 . 01 ) . P-values less than 0 . 05 ( 2-tailed ) were considered statistically significant . | Colicins are bacterial protein toxins which show potent activity against sensitive strains in vitro . Ecological models suggest that colicins play a major role in modulating dynamics of bacterial populations in the gut . However , previous studies could not readily confirm these predictions by respective in vivo experiments . In animal models , colicin-producing strains only show a minor or even absent fitness benefit over sensitive competitors . Here , we propose that the gut environment plays a crucial role in generating conditions for bacterial competition by colicin Ib ( ColIb ) . Gut inflammation favours overgrowth of Enterobacteriaceae ( “inflammation-inflicted Enterobacterial blooms” ) . We show that a pathogenic Salmonella Typhimurium ( S . Tm ) strain benefits from ColIb production in competition against commensal E . coli upon growth in inflammation-inflicted blooms . In the absence of gut inflammation , ColIb production did not confer a competitive advantage to S . Tm . In the inflamed gut , the genes for ColIb production in S . Tm and its corresponding ColIb-surface receptor CirA in E . coli were markedly induced , as compared to the non-inflamed gut . Therefore , environmental conditions in inflammation-inflicted blooms favour colicin-dependent competition of Enterobacteriaceae by triggering ColIb production and susceptibility at the same time . Our findings reveal a role of colicins as important bacterial fitness factors in inflammation-induced blooms . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2014 | Inflammation Fuels Colicin Ib-Dependent Competition of Salmonella Serovar Typhimurium and E. coli in Enterobacterial Blooms |
Quantitative modeling is useful for predicting behaviors of a system and for rationally constructing or modifying the system . The predictive power of a model relies on accurate quantification of model parameters . Here , we illustrate challenges in parameter quantification and offer means to overcome these challenges , using a case example in which we quantitatively predict the growth rate of a cooperative community . Specifically , the community consists of two Saccharomyces cerevisiae strains , each engineered to release a metabolite required and consumed by its partner . The initial model , employing parameters measured in batch monocultures with zero or excess metabolite , failed to quantitatively predict experimental results . To resolve the model–experiment discrepancy , we chemically identified the correct exchanged metabolites , but this did not improve model performance . We then remeasured strain phenotypes in chemostats mimicking the metabolite-limited community environments , while mitigating or incorporating effects of rapid evolution . Almost all phenotypes we measured , including death rate , metabolite release rate , and the amount of metabolite consumed per cell birth , varied significantly with the metabolite environment . Once we used parameters measured in a range of community-like chemostat environments , prediction quantitatively agreed with experimental results . In summary , using a simplified community , we uncovered and devised means to resolve modeling challenges that are likely general to living systems .
Successful prediction of quantitative traits of a biological system can be tremendously useful . For example , if we can quantitatively predict properties of microbial communities , then we will be empowered to design or manipulate communities to harness their activities [1–6] , ranging from fighting pathogens [7] to industrial production of vitamin C [8 , 9] . An important community-level property is community dynamics , including how species concentrations change over time [6] . Community dynamics can be predicted using statistical correlation models . For example , community dynamics observed over a period of time can be used to construct a model that correlates the concentration of one species with the growth rate of another , and the model can then be used to predict future dynamics [10–12] . However , even for two-species communities , statistical correlation models might generate false predictions on species coexistence [13] . Alternatively , mathematical models can be constructed based on species interaction mechanisms , such as how metabolites released by one species might affect the growth of another species . For example , genome-scale metabolic models use genome sequences , sometimes in conjunction with RNA and protein expression profiles , to predict metabolic fluxes within species as well as metabolic fluxes among species ( i . e . , metabolic interactions ) [14 , 15] . However , these models face multiple challenges , including unknown protein functions or metabolic fluxes [16] . When interaction mechanisms are known [14 , 17–21] , we can construct a model based on interaction mechanisms . Ideally , we would use the model to first determine which parameters are critical for the phenomenon of interest , and then directly quantify those critical parameters . However , parameter quantification can be time-consuming . Thus , in many models , a fraction of model parameters are “free parameters” ( unmeasured parameters that can be chosen to fit data ) . Sometimes , a free parameter is assigned a literature value measured in a different strain or even a different species . This is a poor practice for quantitative modeling , because literature values can vary by orders of magnitude [22] . Sometimes , a model is “calibrated” or “benchmarked” to fit experimental data [23] , and thus free parameters become “fitting parameters . ” This type of model calibration can also be problematic , because wrong models can also be calibrated to fit empirical data [23] , and , not surprisingly , the resulting model predictions are likely wrong [23 , 24] . Even when all parameters are directly measured , quantitative modeling can still be challenging . First , a parameter measured from a cell population represents the population average and ignores cell-to-cell heterogeneity [25] , which can be problematic . Second , parameter values may vary with the environment or time [26–29] . For example , the rate of acetate excretion by Escherichia coli is sensitive to the growth environment [28 , 29] . Third , during parameter measurements , cells may rapidly evolve , and thus parameters no longer correspond with the intended genotype . Fourth , in a model with multiple parameters , measurement uncertainty in each parameter can accumulate such that prediction confidence interval is too broad to be useful . Finally , the correctness or sufficiency of a particular model structure can be questionable . It is unclear how severe each of these problems can be in empirical examples , nor how to overcome these problems . As a result , it is not clear how feasible it is to perform quantitative modeling of living systems , including microbial communities . Here , using a highly simplified community of engineered yeast cells , we stress test quantitative modeling of community dynamics . Our community “Cooperation that is Synthetic and Mutually Obligatory” ( CoSMO ) [17] consists of two differentially fluorescent , non-mating haploid Saccharomyces cerevisiae strains ( Fig 1A; S1 Table ) . One strain , designated A−L+ , cannot synthesize adenine ( A ) because of a deletion mutation in the ADE8 gene , and over-activates the lysine ( L ) biosynthetic pathway due to a feedback-resistant LYS21 mutation [30] . The other strain , designated L−A+ , requires lysine because of a deletion mutation in the LYS2 gene , and over-activates the adenine biosynthetic pathway due to a feedback-resistant ADE4 mutation [31] . Overproduced metabolites in both strains are released into the environment and are consumed by the partner . In minimal medium lacking adenine and lysine supplements , the two strains engage in obligatory cooperation and stably coexist [17 , 32] . The biological relevance of CoSMO is as follows . First , simplified communities are useful for biotechnology applications [1 , 3 , 33] . For example , mutualistic communities similar to CoSMO have been engineered to divide up the labor of synthesizing complex drugs [34] . Second , cooperation and mutualisms modeled by CoSMO are widely observed in naturally occurring communities ( including those in the gut and oral microbiota [35 , 36] ) as microbes exchange essential metabolites such as amino acids and cofactors [37–42] . Indeed , principles learned from CoSMO , including how fitness effects of interactions affect the spatial patterning of community members , mechanisms that protect cooperators from non-cooperators , and how to achieve stable species composition in two-species communities , have been found to operate in communities of non-engineered microbes [32 , 43–45] . Because CoSMO has defined species interactions , and because all model parameters can be directly measured , we should be able to quantitatively predict community dynamics . Our initial model predictions significantly deviated from experimental measurements . In the process of resolving model–experiment discrepancies , we have uncovered and resolved multiple challenges in parameter quantification , a critical aspect of quantitative modeling . Because these challenges are likely general , our work serves as a road map that can be applied to quantitative modeling of other cell communities where interaction mechanisms can be inferred from genetic determinants ( see Discussion ) .
Experimentally , CoSMO growth followed a reproducible pattern: after an initial lag marked by slow growth , the two populations and thus the entire community grew at a faster rate ( Fig 1B , “Experiment” ) . Under optimized experimental conditions , post-lag growth rate reached a steady state ( Fig 7A ) . We wanted to quantitatively predict CoSMO’s post-lag steady state growth rate ( “growth rate” ) gcomm , the rate of total population increase . Community growth rate is a measure of how likely the community can survive periodic dilutions such as those in industrial fermenters [46] or during regular bowel movements . By “quantitative prediction , ” we mean that model prediction should fall within experimental error bars . We have formulated a differential equation model of the CoSMO dynamics as the following: d[L−A+]dt= ( bL ( L ) −dL ) [L−A+] ( 1 ) d[A−L+]dt= ( bA ( A ) −dA ) [A−L+] ( 2 ) dLdt=rL[A−L+]−cLbL ( L ) [L−A+] ( 3 ) dAdt=rA[L−A+]−cAbA ( A ) [A−L+] ( 4 ) Eq 1 states that the L−A+ population density ( [L−A+] ) increases at a birth rate ( bL ) dependent on the concentration of lysine ( L ) , and decreases at a fixed death rate ( dL ) . Eq 2 describes how A−L+ population density ( [A−L+] ) changes over time . Eq 3 states that the concentration of lysine ( L ) increases due to releaser A−L+ releasing at a fixed rate ( rL ) , and decreases as the cL amount is consumed per birth of consumer L−A+ . Eq 4 describes how the concentration A changes over time . To predict community growth rate , we either simulated community dynamics ( Fig 1B , dotted lines ) or calculated it from an analytical formula ( Eq 5 ) derived from Eqs 1–4 ( see Methods , “Calculating steady state community growth rate” ) : gcomm≈− ( dA+dL ) 2+rArLcAcL . ( 5 ) Eq 5 suggests that community growth rate depends on metabolite release rates ( rA; rL ) and metabolite consumption per cell birth ( cA; cL ) in a square root fashion , and depends on death rates ( dA; dL ) in a linear fashion . Simulations and analytical calculations yielded similar results ( e . g . , S1 Fig ) . Because death rates are small ( Table 1 ) compared to community growth rate gcomm ( 0 . 11 ± 0 . 01/h in Fig 7B ) , release and consumption parameters are important and should be carefully measured . Eq 5 also states that even if one parameter is free , its value can always be chosen such that the calculated community growth rate will perfectly match experiments , regardless of the accuracy of the remaining five parameters . This is the well-known danger of free parameters . Model parameters correspond to strain phenotypes and include metabolite release rate , metabolite consumption per birth , and cell birth and death rates . Even though these phenotypes reflect strain interactions ( “interaction phenotypes” in Fig 1A ) , we measured them in monocultures to eliminate partner feedback . In our earlier studies , we quantified some of these phenotypes and borrowed others from literature values [17 , 32 , 44] . Our models correctly predicted various properties of CoSMO , including the steady-state strain ratio [17] as well as qualitative features of spatial patterning [32 , 44] . Our first model ( Model i ) underestimated community growth rate . Unlike the published strains of A−L+ and L−A+ in the S288C background [17] , strains in this study were constructed in the RM11 background to reduce mitochondrial mutation rate [48] . For each RM11 strain , we measured death rate during starvation using a microscopy batch-culture assay [27] . We also quantified the amount of metabolite consumed per birth in batch cultures grown to saturation ( see Fig 4B for details; S13 Fig ) , similar to our earlier work [17] . Because release rates were more tedious to measure , we initially borrowed published release rates of L−A+ and A−L+ in the S288C background in batch starved cultures [17] . Predicted community growth rates were much slower than experimental measurements ( Fig 1B , “Model i”; Fig 7B , gray ) . A revised model ( Model ii ) without any borrowed parameters overestimated community growth rate . For this model , we directly measured the release rates of RM11 L−A+ and A−L+ in batch starved cultures ( see Fig 5B and S15B Fig for details ) . The release rates of both strains in the RM11 background were approximately 3-fold higher than those in the S288C background ( S2 Table ) . Consequently , the predicted community growth rate greatly exceeded experiments ( Fig 1B , “Model ii”; Fig 7B , blue ) . One possible cause for the model–experiment discrepancy could be that cells engineered to overproduce adenine or lysine [30 , 31] might instead release derivatives of adenine or lysine . Consequently , when we quantified phenotypes such as metabolite consumption , we could have supplemented the wrong metabolite and been misled . A genome-scale metabolic model of S . cerevisiae predicted that although A−L+ likely released lysine , L−A+ likely released hypoxanthine or adenosine- ( 3 , 5 ) -biphosphate instead of adenine [49 , 50] . Nanospray desorption electrospray ionization mass spectrometry imaging ( nanoDESI MS ) [51] performed by the Julia Laskin lab revealed a lysine gradient emanating from A−L+ and hypoxanthine and inosine gradients emanating from L−A+ , although the signals were noisy . We followed up this observation using high-pressure liquid chromatography ( HPLC ) ( Fig 2 ) . Indeed , lysine mediates the interaction from A−L+ to L−A+ . We subjected A−L+ supernatant to HPLC ( Methods , “HPLC” ) and a yield-based bioassay ( Methods , “Bioassays” ) . In HPLC , a compound in A−L+ supernatant eluted at the same time as the lysine standards ( Fig 2A ) , and its concentration could be quantified by comparing the peak area against those of lysine standards ( Fig 2A inset ) . In bioassay , we quantified the total lysine-equivalent compounds in an A−L+ supernatant by growing L−A+ in it and comparing the final turbidity with turbidities achieved in minimal medium supplemented with various known concentrations of lysine . HPLC quantification agreed with the yield bioassay ( Fig 2B ) . Thus , lysine-equivalent compounds released by A−L+ were primarily lysine . Hypoxanthine mediates the interaction from L−A+ to A−L+ . When we subjected L−A+ supernatants to HPLC , we found compounds at the elution times of hypoxanthine and inosine , but not of adenine ( Fig 2C ) . Hypoxanthine but not inosine supported A−L+ growth , and inosine did not affect how hypoxanthine stimulated A−L+ growth ( S3 Fig ) . Hypoxanthine concentration quantified by HPLC agreed with the concentration of purines consumable by A−L+ in the yield bioassay ( Fig 2D; Methods , “Bioassays” ) . Thus , A−L+ primarily consumed hypoxanthine released by L−A+ . Using phenotypes of A−L+ measured in hypoxanthine versus adenine happened to not affect model performance . Death and release rates were not affected because they were measured in the absence of purine supplements . Similar amounts of hypoxanthine and adenine were consumed to produce a new A−L+ cell ( S3 Fig ) . Although the birth rate of A−L+ was slower in the presence of hypoxanthine compared with adenine , especially at low concentrations ( S4 Fig ) , this difference did not affect community growth rate ( Eq 5 ) . Thus , distinguishing whether hypoxanthine or adenine was the interaction mediator did not make a difference in predicting community growth rate ( S1 Fig ) . Here , we continue to use A to represent the adenine precursor hypoxanthine . Model–experiment discrepancy ( Fig 1B ) could be caused by phenotypes being dependent on the environment . So far , we had measured phenotypes in batch cultures containing zero or excess metabolite . Thus , we set out to remeasure strain phenotypes in chemostats [52] that mimicked CoSMO environments . Specifically , in a chemostat , fresh medium containing the required metabolite ( lysine or hypoxanthine ) was pumped into the culturing vessel at a fixed rate ( “dilution rate” ) , while culture medium containing cells exited the culturing vessel at the same rate ( Methods , “Chemostat culturing” ) . After an initial adjustment stage , live population density reached a steady state ( Fig 3A ) , which meant that the population grew at the same rate as the dilution rate ( Eqs 6–11 in Methods , “Quantifying phenotypes in chemostats” ) [52] . By setting the chemostat dilution rate to various growth rates experienced by CoSMO ( i . e . , 5 . 5-h to 8-h doubling ) , we could mimick the CoSMO growth environments . From the population and chemical dynamics in the chemostat , we could then measure metabolite release rate , metabolite consumption per birth , and death rate ( Eqs 12–16 in Methods , “Quantifying phenotypes in chemostats” ) . During chemostat measurements , ancestral L−A+ was rapidly overtaken by mutants with dramatically improved affinity for lysine ( Fig 3C; S7 Fig; Methods , “Detecting evolved clones” ) , consistent with our earlier work [43] . These mutants , likely being present in the inoculum at a low ( on the order of 10−6 ) frequency , displayed a growth rate 3 . 6-fold that of the ancestor during lysine limitation ( S7 Fig ) . Thus , to measure ancestral L−A+ phenotypes , we terminated measurements before mutants could take over ( <10% , before magenta dashed lines in Fig 3 ) . In contrast , the evolutionary effects of A−L+ mutants on CoSMO growth were captured during phenotype measurements . Unlike L−A+ mutants , A−L+ mutants were constantly generated from ancestral cells at an extremely high rate ( on the order of 0 . 01/cell/generation; Methods , “Evolutionary dynamics of mutant A−L+” ) , presumably via frequent chromosome duplication ( S8C Fig ) . These mutants were present at a significant frequency ( 1%–10% ) , even before our measurements started , and slowly rose to 30%–40% during measurements due to their moderate fitness advantage over the ancestor under hypoxanthine limitation ( S8A Fig; S9 Fig; Methods , “Detecting evolved clones” ) . Consequently , we measured the average phenotypes of an evolving mixture of ancestors and mutants . Fortunately , these averaged phenotypes could be used to model CoSMO , because mutants accumulated in similar fashions during phenotype measurements and during CoSMO measurements so long as the two time windows were compatible ( S9B Fig; S11 Fig ) . Metabolite consumption per birth depends on the growth environment . Consistent with our previous work [17] , consumption during exponential growth in excess supplement was higher than that in a culture grown to saturation ( Fig 4; Methods , “Measuring consumption in batch cultures” ) , presumably due to exponential phase cells storing excess metabolites [54] . Consumption in chemostats ( Methods , “Quantifying phenotypes in chemostats , ” Eq 12 ) was in between exponential and saturation consumption ( Fig 4C for L−A+ and S13 Fig for A−L+ ) . For both strains , because consumption in chemostat was relatively constant across the range of doubling times encountered in CoSMO ( 5 . 5–8 h ) , we used the average value in Model iii ( dashed line in Fig 4C and S13 Fig; Table 1; S5 Table , S6 Table ) . Metabolites can be released by live cells or leaked from dead cells . We want to distinguish between live versus dead release for the following reasons . First , if death rate were to evolve to be slower , then live release would predict increased metabolite supply , whereas dead release would predict the opposite . Second , dead release would imply nonspecific release and , thus cell–cell interactions may be highly complex . Finally , leakage from dead cells is thermodynamically inevitable , whereas active release of costly molecules would require an evolutionary explanation . Hypoxanthine is likely released by live L−A+ . In the absence of lysine ( Methods , “Starvation release assay” ) , we tracked the dynamics of live and dead L−A+ ( Fig 5A , magenta and gray ) and of hypoxanthine accumulation ( Fig 5A , lavender ) . If live cells released hypoxanthine , then hypoxanthine should increase linearly with live cell density integrated over time ( i . e . , the sum of live cell density * h , Fig 5B , left ) , and the slope would represent the live release rate ( fmole/cell/h ) . If cells released hypoxanthine upon death , then hypoxanthine should increase linearly with dead cell density , and the slope would represent the amount of metabolite released per cell death ( Fig 5B , right ) . Because the live release model explained our data better than the dead release model ( Fig 5B ) , hypoxanthine was likely released by live cells during starvation . In lysine-limited chemostats , we could not use dynamics to distinguish live from dead release ( note the mathematical equivalence between Eqs 9 and 10 in Methods , “Quantifying phenotypes in chemostats” ) . Instead , we harvested cells and chemically extracted intracellular metabolites ( Methods , “Extraction of intracellular metabolites” ) . Each L−A+ cell , on average , contained 0 . 12 ( ±0 . 02 , 95% CI ) fmole of hypoxanthine ( Methods , “HPLC” ) . If hypoxanthine was released by dead cells ( about 105 dead cells/mL , Fig 3A ) , we should see 0 . 012 μM instead of the observed approximately 10 μM hypoxanthine in the supernatant ( Fig 3B ) . Thus , hypoxanthine is likely released by live L−A+ in chemostats . Hypoxanthine release rates of L−A+ are similar in lysine-limited chemostats mimicking the CoSMO environments ( Methods , “Quantifying phenotypes in chemostats , ” Eq 14 ) versus during starvation ( Fig 5C ) . Thus , we used the average hypoxanthine release rate ( Fig 5C black dashed line; Table 1 ) in Model iii . Note that release rates declined in faster-growing cultures ( ≤3-h doubling; Fig 5C ) , but we did not use these data because CoSMO did not grow that fast . Lysine is likely released by live A−L+ . When we measured lysine release from starving A−L+ cells ( S15A Fig ) , a model assuming live release and a model assuming dead release generated similar matches to experimental dynamics ( S15B and S15C Fig ) . However , after measuring intracellular lysine content , we concluded that dead release was unlikely , because each dead cell would need to release significantly more lysine than that measured inside a cell to account for supernatant lysine concentration , especially during the early stage of starvation ( S16B Fig ) . Lysine release rate of A−L+ is highly sensitive to the growth environment ( Fig 6B , details in S20 Fig ) and reaches a maximum at an intermediate growth rate . Release rates in 7–8-h doubling chemostats were about 60% more than those during starvation . Lysine release rate rapidly declined as hypoxanthine became more available ( i . e . , as growth rate increased , Fig 6B ) . Variable release rate could be due to variable intracellular lysine content: lysine content per cell increased by severalfold upon removal of hypoxanthine ( from 2 . 9 fmole/cell to about 19 fmole/cell; Fig 6A black dotted line ) and leveled off at a higher level in 8-h chemostats than during starvation ( Fig 6A ) . We incorporated a variable lysine release rate in Model iii ( Table 1 ) . Death rates , which could affect CoSMO growth rate ( Methods , Eq 5 ) , are also sensitive to the environment . We measured death rates in chemostats ( Methods , “Quantifying phenotypes in chemostats , ” Eq 13 or Eq 16 ) and found them to be distinct from the death rates in zero or excess metabolite ( S21 Fig ) . Because death rates were relatively constant in chemostats mimicking the CoSMO environments ( S21 Fig , blue lines ) , we used the averaged values in Model iii ( Table 1; S7 Table; S8 Table ) . Our chemostat-measured model parameters are internally consistent: mathematical models of L−A+ in lysine-limited chemostat ( S4 Code ) and of A−L+ in hypoxanthine-limited chemostat ( S5 Code ) captured experimental observations ( S12 Fig; S19 Fig ) . Using parameters measured in chemostats ( Table 1 ) , model prediction on CoSMO growth rate quantitatively matches experimental results . Experimentally , because L−A+ mutants quickly took over well-mixed CoSMO ( red in S22A Fig [43] ) , we grew CoSMO in a spatially-structured environment so that fast-growing mutants were spatially confined to their original locations and remained a minority ( red in S22B Fig ) . Spatial CoSMO growth rates measured under a variety of experimental setups ( e . g . , agarose geometry and initial total cell density ) remained consistent ( 0 . 11 ± 0 . 01/h; Fig 7B purple; S24 Fig ) . In Model iii , an analytical formula ( Eq 5; Methods , “Calculating steady-state community growth rate” ) and spatial CoSMO simulations based on chemostat-measured parameters ( Table 1 ) both predicted CoSMO growth rate to be 0 . 10 ± 0 . 01/h ( Fig 7B , green and brown ) . Thus , chemostat parameters allowed our model to quantitatively explain experimental CoSMO growth rate ( Fig 7 green and brown versus purple ) . This also suggests that our parameter measurements are valid . Note that although Model iii captures the steady-state growth rate of CoSMO , it fails to recapitulate quantitative details of strain dynamics during and immediately after the initial lag phase ( S26 Fig ) . This is not surprising , because strain phenotypes during starvation are complex ( e . g . , being time dependent , S18B Fig ) [27] and differ from those in chemostats ( Fig 4; Fig 6; S13 Fig; S21 Fig ) . In summary , phenotypic parameters are often sensitive to the environment . Thus , measuring phenotypes in a range of community-like environments may be required for quantitative modeling . Rapid evolution may further interfere with parameter measurements and model testing . Only after overcoming these challenges did we succeed in quantitatively predicting the steady-state growth rate of CoSMO .
Microbial communities are complex . Thus , qualitative modeling has been deployed to yield biological insights [55 , 56] . However , one would eventually like to understand how community-level properties quantitatively emerge from interactions among member species . The simplicity of CoSMO has allowed us to directly measure all parameters , uncover some of the challenges to quantitative modeling , and devise means to overcome these challenges . These challenges are likely general to other living systems . Below , we discuss what we have learned from quantitative modeling of CoSMO steady-state growth rate . Even when genetic determinants are known , interaction mediators can be nontrivial to identify . In CoSMO , we previously thought that adenine was released by L−A+ , whereas in reality , hypoxanthine and inosine are released ( Fig 2 ) . Fortuitously , hypoxanthine but not inosine affects A−L+ growth ( S3 Fig ) . Otherwise , we might be forced to quantify how hypoxanthine and inosine , in isolation and in different combinations , might affect A−L+ . A−L+ grows faster in adenine than in hypoxanthine ( S4 Fig ) , and although this does not affect our prediction of CoSMO growth rate ( S1 Fig ) , it could affect predictions on other community properties . Many mathematical models have relied on free parameters , which can be problematic when predictions are sensitive to the values of free parameters . In the case of CoSMO , release rates from two strain backgrounds differed by severalfold ( S2 Table ) , and not surprisingly , borrowing parameters affected prediction ( Fig 1B ) . A major challenge we uncovered was environment-sensitive parameters . A key assumption in modeling is invariant parameters . As we have demonstrated here , phenotypes ( e . g . , metabolite consumption per birth , metabolite release rate , and death rate ) measured in zero or excess metabolite can differ dramatically from those measured in metabolite-limited chemostats ( Fig 4C; Fig 5C; Fig 6B; S13 Fig; S21 Fig ) . Furthermore , even within the range of metabolite limitation experienced by CoSMO ( doubling times of 5 . 4–8 h ) , lysine release rate varied by as much as 2-fold ( Fig 6B ) , which could be caused by variable intracellular metabolite concentrations ( Fig 6A ) . Based on parameters measured in chemostats ( including variable lysine release rate ) , Model iii quantitatively predicts experimental results ( Fig 7 ) . Environment-sensitive parameters make quantitative modeling intrinsically difficult , because community environment often changes with time , and so will environment-sensitive model parameters . Even if we are only interested in predicting the steady-state community property , we may not know in advance what that steady state is and thus which environment to measure parameters in . For complex communities , multiple states could exist [57] . Thus , we may need to measure parameters in a range of environments that are typically encountered in a community . Another obstacle for model building and testing is rapid evolution . If we quantify phenotypes in starved batch cultures , cells do not grow and thus evolution is slow , but the environment deviates significantly from the community environment . In chemostat measurements , we can control the environment to mimic those encountered by the community . However , in addition to the time-consuming nature of constructing and calibrating chemostats to ensure accurate flow rates [58] , rapid evolution occurs . For L−A+ , mutants pre-exist at a low frequency but can grow severalfold faster than the ancestor ( S7 Fig ) . Consequently , a population will remain largely ( >90% ) ancestral only for the first 24 h in the well-mixed chemostat environment ( Fig 3 ) . A short measurement time window poses additional challenges if , for example , the released metabolite has not accumulated to a quantifiable level . For A−L+ , mutants are generated from ancestral cells at an extremely high rate before and during phenotype quantification ( Methods , “Evolutionary dynamics of mutant A−L+”; S9 Fig ) . Because mutants accumulated at a similar rate in CoSMO ( S9B Fig ) , we accounted for evolutionary effects by using similar quantification time windows for A−L+ phenotypes and for CoSMO growth rate . Note that this approximation is valid here , because our model ( without any free parameters ) matches experiments quantitatively ( Fig 7B ) . Rapid evolution also poses a problem for model testing . For example , when quantifying CoSMO growth rate , which requires several days , we were forced to use a spatially structured environment so that fast-growing L−A+ mutants could not take over ( S22 Fig ) . Thus , unless one is careful , one may not even know what one is measuring ! Rapid evolution need not be unique to our engineered community of “broken” strains . Indeed , rapid evolution has been observed in phage–bacteria communities in aquatic environments and in acidophilic biofilms [59–61] . Rapid evolution is not surprising: given the large population sizes of microbial populations , mutants can pre-exist [62] . These pre-existing mutants can quickly take over in novel environments ( e . g . , exposure to evolving predators or to man-made pollutants and drugs ) where the ancestor is ill adapted . Choosing the right level of abstraction is yet another important consideration during model building , because different levels of abstraction show trade-offs between generality , realism , and precision [63] . When the level of abstraction is chosen properly , even complex biological phenomena can be described by simple and predictive equations . For example , a simplified model considering negative feedback regulation of carbon intake in E . coli quantitatively predicted cell growth rate on two carbon sources based on growth rates on individual carbon sources using only one single parameter that is fixed by experiments [64] . For CoSMO , one could construct a complex model that , for example , considers physiological and genetic networks of each cell to account for the dependence of phenotypes on the environment and on evolution . However , this would require making numerous assumptions and measuring even more numerous parameters . In the absence of free parameters , quantitative matching between model predictions and experimental results provides strong evidence that no additional complexity is required to explain the biological phenomenon of interest . Once the right level of abstraction is chosen , a good model can serve multiple purposes [65–67] , especially when coupled with quantitative measurements . First , a model suggests which parameters need to be carefully measured . For example , for spatial CoSMO growth rate , parameters such as diffusion coefficients are not critical ( S23 Fig ) , but metabolite release and consumption parameters are ( Eq 5 ) . Second , a useful model not only explains existing data but also makes extrapolative predictions accurately . An example is the quantitative theory of the lac operon in E . coli ( [68–70] ) . Finally , model–experiment discrepancy exposes knowledge gaps . When predicting CoSMO growth rate , the missing piece was environment-sensitive phenotypes . Our approach can be applied to communities where interaction mechanisms can be inferred from genetic analysis . For example , we have applied this approach to understand an evolved metabolic interaction . Specifically , we observed that a single yeast population evolutionarily diverged into two genetically distinct subpopulations [71] . One subpopulation acquired a met− mutation that prevented the synthesis of organosulfurs and thus must rely on the MET+ subpopulation for organosulfurs ( which are essential for viability ) . Similar to this work , we first identified the released organosulfurs to be mainly glutathione and glutathione conjugates , using liquid chromatography–mass spectrometry . Because glutathione and glutathione conjugates were consumed by met− cells in a similar fashion , we “lump-summed” organosulfurs and quantified them in terms of “glutathione equivalents . ” We then determined that organosulfurs were likely released by live cells , and quantified organosulfur release rate at various MET+ growth rates . Finally , we quantified organosulfur consumption per birth of met− cell . These measurements allowed us to understand the steady-state ratio of the two subpopulations [71] . In summary , despite the many challenges , quantitative modeling of cell communities is possible . Importantly , by eliminating free parameters through direct experimental quantification , we arrive at two possibilities , both useful: quantitative matching between model predictions and experiments would provide strong evidence that no additional complexity is required to explain the biological phenomenon of interest . Significant mismatching between predictions and experiments would motivate us to look for the important missing pieces .
We constructed CoSMO in the RM11 background due to its lower rate of mitochondrial mutation [48] compared with the S288C background used in our earlier studies [17] . Thus , phenotypes measured here differed from those measured in strains of the S288C background [17 , 32] . We introduced desired genetic modifications into the ancestral RM11 background via transformation [72 , 73] ( S1 Table ) . Strains were stored at −80°C in 15% glycerol . We used rich medium YPD ( 10 g/L yeast extract , 20 g/L peptone , 20 g/L glucose ) in 2% agar plates for isolating single colonies . Saturated YPD overnight liquid cultures from these colonies were then used as inocula to grow exponential cultures . To prevent purines from being yield limiting , we supplemented YPD with 100 μM hypoxanthine for A−L+ cells . We sterilized YPD media by autoclaving . YPD overnights were stored at room temperature for no more than 4–5 d prior to experiments . We used defined minimal medium SD ( 6 . 7 g/L Difco yeast nitrogen base w/o amino acids , 20 g/L glucose ) for all experiments [74] , with supplemental metabolites as noted [73] . To achieve higher reproducibility , we sterilized SD media by filtering through 0 . 22-μm filters . To make SD plates , we autoclaved 20 g/L Bacto agar or agarose in H2O and , after autoclaving , supplemented equal volume of sterile-filtered 2-fold concentrated SD . CoSMO steady-state growth rates on agar ( which contains trace contaminants of metabolites ) and agarose generate similar results . All culturing , unless otherwise noted , was performed at 30°C in a well-mixed environment where culture tubes were inserted sideways into a roller drum ( Model TC-7 , New Brunswick Scientific , Edison , NJ ) . L−A+ cells were pregrown to exponential phase ( OD600 generally less than 0 . 4 in 13-mm culture tubes , or <2 . 8 × 107 cells/mL ) in SD supplemented with excess ( 164 μM ) lysine and washed 3–5 times with SD . In microscopy assays , when noted , we starved L−A+ cells for 4 h to deplete intracellular lysine storage . Otherwise , we did not prestarve L−A+ . A−L+ cells were pregrown to exponential phase in SD supplemented with excess hypoxanthine ( 100 μM ) or excess adenine ( 108 μM ) as noted , washed 3–5 times with SD , and prestarved in SD for 24 h to deplete cellular purine storage . We pre-starved A−L+ to reduce CoSMO growth lag ( S2 Fig ) , thus facilitating quantification of CoSMO growth rate . To be consistent , we also prestarved A−L+ during phenotype quantification . We prepared fluorescent bead stocks ( 3-μm red fluorescent beads Cat R0300 , Thermo Fisher Scientific , Waltham , MA ) . Beads were autoclaved in a factory-clean glass tube , diluted into sterile 0 . 9% NaCl , and supplemented with sterile-filtered Triton X-100 to a final 0 . 05% ( to prevent beads from clumping ) . We sonicated beads and kept them in constant rotation to prevent settling . We quantified bead concentrations by counting beads via hemacytometer and Coulter counter . Final bead stock was generally 4–8 × 106/mL . Culture samples were diluted to OD 0 . 01–0 . 1 ( 7 × 105–7 × 106/mL ) in Milli-Q H2O in unautoclaved 1 . 6-mL Eppendorf tubes . A total of 90 μL of the diluted culture sample was supplemented with 10 μL bead stock and 2 μL of 1 μM ToPro 3 ( T-3605 , Molecular Probes , Eugene , OR ) , a nucleic acid dye that only permeates compromised cell membranes ( dead cells ) . Sample preparation was done in a 96-well format for high-throughput processing . Flow cytometry of the samples was performed on Cytek ( Fremont , CA ) DxP Cytometer equipped with 4 lasers , 10 detectors , and an autosampler . Fluorescent proteins GFP , Citrine , mCherry , TagBFP-AS-N ( Evrogen , Moscow , Russia ) , and ToPro are respectively detected by a 50-mW 488-nm laser with 505/10 ( i . e . , 500–515-nm ) detector , a 50-mW 488-nm laser with 530/30 detector , a 75-mW 561-nm laser with 615/25 detector , a 50-mW 408-nm laser with 450/50 detector , and a 25-mW 637-nm laser with 660/20 detector . Each sample was run in triplicates and individually analyzed using FlowJo software to identify numbers of events of beads , dead cells , and various live fluorescent cells . Densities of various populations were calculated from the cell–bead ratios . We then calculated the mean cell density from triplicate measurements , with the coefficient of variation generally within 5%–10% . All HPLC measurements were done on a Shimadzu ( Kyoto , Japan ) Nexera X2 series ultra-performance HPLC ( UHPLC ) system . All supernatant samples were filtered ( 0 . 22-μm filter ) . For standards , we made a high-concentration solution , filtered it , and stored it at −80°C . Prior to an HPLC run , we diluted the stock to various concentrations in filtered H2O . To quantify lysine , a 100-μL sample was loaded into an Agilent ( Santa Clara , CA ) 250 μL pulled point glass small volume insert ( part number 5183–2085 ) , which was then placed inside a Shimadzu 1 . 5 mL 12 × 32 mm autosampler vial ( part number 228-45450-91 ) . This vial was then placed into an autosampler ( Nexera X2 SIL-30AC ) . Prior to injection into the column , samples were derivatized at 25°C with freshly made derivatization reagents in the autosampler using a programmed mixing method as follows . A total of 7 . 5 μL of sample was removed and placed into a separate reaction small volume insert and vial . Next , 45 μL of mercaptopropionic acid ( 10 μL per 10 mL 0 . 1 M sodium borate buffer , pH 9 . 2 ) and 22 μl of o-phthaladehyde ( 10 mg per 5 mL 0 . 1 M sodium borate buffer , pH 9 . 2 ) were added to this vial , mixed , and incubated for 1 min . A total of 10 μL of 9-fluorenyl methyl chloroformate ( 4 mg per 20 mL acetonitrile , HPLC grade ) was then added , and the sample was remixed and incubated for 2 min . Finally , 10 μL of the reaction mixture was injected onto Phenomenex ( Torrance , CA ) Kinetex 2 . 6 μm EVO C18 100 Å LC Column ( 150 × 3 . 0 mm , part number 00F-4725-Y0 ) fitted with a SecurityGuard ULTRA Holder for UHPLC Columns ( 2 . 1 to 4 . 6 mm , part number AJ0-9000 ) and a SecurityGuard ULTRA cartridge ( 3 . 0-mm internal diameter , part number AJ0-9297 ) . SecurityGuard ULTRA cartridge ( precolumn ) was periodically replaced in the event of pressure reading exceeding the manufacturer’s specifications . Compounds were eluted from the column using a gradient of HPLC-grade Solution A ( 73 mM potassium phosphate , pH 7 . 2 ) and Solution B ( 50:50 acetonitrile/methanol ) . Solution A was filtered through a 0 . 2-μm filter prior to use . The percentage of solution B follows the following program: 0–2 min , 11%; 2–4 min , 17%; 4–5 . 5 min , 31%; 5 . 5–10 min , 32 . 5%; 10–12 min , 46 . 5%; and 12–15 . 5 min , 55% . The flow rate is maintained at 0 . 1 mL/min . The column was then flushed with 100% solution B for 5 min and re-equilibrated for 5 min with 11% solution B at 0 . 8 mL/min . The column was maintained at a running temperature of 35°C in a Nexera X2 CTO-20A oven . Absorbance measurements at 338 nm were measured using a high-sensitivity flow cell for a SPD-M30A UV-Vis detector . For purines , we used the above protocol without the derivatization steps . Instead , a 5–10-μL sample was directly injected onto the column . We used a yield-based bioassay for relatively high metabolite concentrations ( ≥5 μM for lysine and ≥2 μM for hypoxanthine ) . For lower concentrations , we used a rate-based bioassay with a sensitivity of 0 . 1 μM for both lysine and hypoxanthine . When necessary , we diluted the sample to get into the assay linear range . In the yield bioassay , a 75-μL sample filtered through a 0 . 2-μm filter was mixed with an equal volume of a master mix containing 2-fold concentrated SD ( to provide fresh medium ) as well as tester cells auxotrophic for the metabolite of interest ( about 1 × 104 cells/mL , WY1335 for lysine or WY1340 for hypoxanthine ) in a flat-bottom 96-well plate . We then wrapped the plate with parafilm and allowed cells to grow to saturation at 30°C for 48 h . We resuspended cells using a Thermo Scientific Teleshake ( setting #5 for about 1 min ) and read culture turbidity using a BioTek Synergy MX plate reader . Within each assay , SD supplemented with various known concentrations of metabolite was used to establish a standard curve that related metabolite concentration to final turbidity ( e . g . , S3A Fig ) . From this standard curve , the total concentration of metabolites that can support auxotroph growth in an unknown sample could be inferred . The rate bioassay was used for samples with low metabolite concentrations . For example , to measure lysine concentration in a lysine-limited chemostat , we mixed 150 μL filtered sample with an equal volume of master mix containing 2-fold concentrated SD and L−A+ tester cells ( about 1 × 104 cells/mL ) in a flat-bottom 96-well plate . As our tester strain for lysine , we used an evolved clone ( WY 2270 ) isolated after L−A+ had grown for tens of generations under lysine limitation . This clone displayed increased affinity for lysine due to an ecm21 mutation and duplication of Chromosome 14 . Growth rates of the tester strain in SD supplemented with various known concentrations of lysine and in the unknown sample were measured using a microscopy assay ( Methods , “Microscopy quantification of growth phenotypes” ) . The growth rate of WY 2270 tester cells scaled linearly with lysine concentrations up to 1 μM ( S6A Fig ) . Similarly , for hypoxanthine , we used an evolved A−L+ strain ( WY1600 ) as the tester strain . The linear range was up to about 0 . 3 μM ( S6B Fig ) . From the standard curve , we could infer the metabolite concentration of a sample . To extract intracellular metabolites , we poured a cell culture onto a 0 . 45-μm nitrocellulose membrane filter ( Cat 162–0115 , BioRad , Hercules , CA ) in a reusable filtration device ( glass microanalysis 25-mm vacuum filter holder with a 15-mL funnel , Product FHMA25 , Southern Labware , Cumming , GA ) , applied vacuum to drain the supernatant , transferred the filter into extraction solution ( 40% acetonitrile , 40% methanol , and 20% water ) , vortexed to dislodge cells , and then removed the filter . This sequence was carried out as rapidly as possible ( <10 s ) . We then flash-froze the extraction solution in liquid nitrogen and allowed it to thaw at −20°C . After thawing , we subjected the solution to five rounds of the following: vortexing for 1 min , and incubating on ice for 5 min between each vortexing . We then spun down the solution in a refrigerated centrifuge for 10 min at 14 , 000 rpm to pellet membrane-permeabilized cells as well as any membrane filter bits that may have disintegrated into the extraction solution . We transferred the supernatant containing soluble cell extract to a new tube . In order to make sure that all soluble components were extracted , we resuspended the cell pellet in a half volume of fresh extraction solution and subjected cells to another round of the same procedure ( flash-freezing , five rounds of vortexing–ice incubation , and centrifugation ) . We then removed the supernatant and added it to the original supernatant . We then dried off the extraction solution in a centrifugal evaporator and resuspended soluble components in water . This resultant solution could then be assayed for metabolite concentrations . When properly dried , extracts did not contain inhibitors that might interfere with bioassays ( S5 Fig ) . For L−A+ , cells from 19-mL cultures ( 4 × 105–4 × 106 cells/mL ) were resuspended in 3 mL extraction buffer . One third of the sample was further processed , and extracted metabolites were resuspended in 0 . 5 mL water . For A−L+ , metabolites from 1–5-mL cultures ( 1–6 × 106 cells/mL ) were extracted and resuspended in 1 mL water . See [27] for details on microscopy experimental setup , method validation , and data analysis . Briefly , cells were diluted to low densities to minimize metabolite depletion during measurements . Dilutions were estimated from culture OD measurement to result in 1 , 000–5 , 000 cells inoculated in 300 μL SD medium supplemented with different metabolite concentrations in wells of a transparent flat-bottom microtiter plate ( e . g . , Costar 3370 ) . We filled the outermost wells with water to reduce evaporation . Microtiter plates were imaged periodically ( every 0 . 5–2 h ) under a 10× objective in a Nikon ( Melville , NY ) Eclipse TE-2000U inverted fluorescence microscope . The microscope was connected to a cooled CCD camera for fluorescence and transmitted light imaging . The microscope was enclosed in a temperature-controlled chamber set to 30°C . The microscope was equipped with motorized stages to allow z-autofocusing and systematic xy-scanning of locations in microplate wells , as well as motorized switchable filter cubes capable of detecting a variety of fluorophores . Image acquisition was done with an in-house LabVIEW program , incorporating bright-field autofocusing [27] and automatic exposure adjustment during fluorescence imaging to avoid saturation . Condensation on the plate lid sometimes interfered with autofocusing . Thus , we added a transparent “lid warmer” on top of our plate lid [27] and set it to be 0 . 5°C warmer than the plate bottom , which eliminated condensation . We used an ET DsRed filter cube ( Exciter: ET545/30x , Emitter: ET620/60m , Dichroic: T570LP ) for mCherry-expressing strains and an ET GFP filter cube ( Exciter: ET470/40x , Emitter: ET525/50m , Dichroic: T495LP ) for GFP-expressing strains . Time-lapse images were analyzed using an ImageJ plug-in , Bioact [27] . Bioact measured the total fluorescence intensity of all cells in an image frame after subtracting the background fluorescence from the total fluorescence . A script plotted background-subtracted fluorescence intensity over time for each well to allow visual inspection . If the dynamics of four positions looked similar , we randomly selected one to inspect . In rare occasions , all four positions were out of focus and were not used . In a small subset of experiments , a discontinuous jump in data appeared in all four positions for unknown reasons . We did not calculate rates across the jump . Occasionally , one or two positions deviated from the rest . This could be due to a number of reasons , including shift of focal plane , shift of field of view , black dust particles , or bright dust spots in the field of view . The outlier positions were excluded after inspecting the images for probable causes . If the dynamics of four positions differed because of cell growth heterogeneity at low concentrations of metabolites , all positions were retained . We normalized total intensity against that at time zero , and averaged across positions . We calculated growth rate over three to four consecutive time points and plotted the maximal net growth rate against metabolite concentration ( e . g . , S4 Fig ) . If maximal growth rate occurred at the end of an experiment , then the experimental duration was too short and data were not used . For L−A+ , the initial stage ( 3–4 h ) residual growth was excluded from analysis . For A−L+ , because cells had already been prestarved , fluorescence intensity did not continue to increase in the absence of supplements . For longer A−L+ imaging ( 30+ h ) , we observed two maximal growth rates at low hypoxanthine concentrations ( e . g . , about 0 . 4 μM ) , possibly due to mutant clones . We used the earlier maximal growth rate even if it was lower than the later maximal growth rate , because the latter was probably caused by faster-growing mutants . We have constructed an eight-vessel chemostat with a design modified from [75] . For details of construction , modification , calibration , and operation , see [58] . For L−A+ , due to rapid evolution , we tried to devise experiments so that live and dead populations quickly reached steady state . Two conditions seemed to work well . In both , we first calculated the expected steady-state cell density by dividing the concentration of lysine in the reservoir ( 20 μM ) by fmole lysine consumed per new cell . Condition 1 consisted of the following: wash exponentially growing cells to completely remove any extracellular lysine and inoculate the full volume ( 19 mL ) at 100% of expected steady-state density . Start chemostat to drip in lysine at the prespecified flow rate . Condition 2 consisted of the following: wash exponentially growing cells to remove extracellular lysine and inoculate 50%–75% of the volume at 1/3 of the expected steady-state density . Fill the rest of the 19-mL vessel with reservoir media ( resulting in less than the full 20 μM of starting lysine , but more than enough for maximal initial growth rate , about 10–15 μM ) . The two conditions yielded similar results ( Fig 3 ) . We predominantly used Condition 2 . We set the pump flow rate to achieve the desired doubling time T ( 19 mL * ln ( 2 ) /T ) . We collected and weighed waste media for each individual culturing vessel to ensure that the flow rate was correct ( i . e . , total waste accumulated over time t was equal to the expected flow rate * t ) . We sampled cultures periodically to track population dynamics using flow cytometry ( Methods , “Flow cytometry” ) , filtered supernatant through a 0 . 45-μm nitrocellulose filter , and froze the supernatant for metabolite quantification at the conclusion of an experiment ( Methods , “Bioassays” ) . At the conclusion of an experiment , we also tested input media for each individual culturing vessel to ensure sterility by plating a 300-μL aliquot on an YPD plate and checking for growth after 2 d of growth at 30°C . If a substantial number of colonies grew ( >5 colonies ) , the input line was considered contaminated and data from that vessel were not used . A−L+ cells exponentially growing in SD + 100 μM hypoxanthine were washed and prestarved for 24 h . We then filled the chemostat culturing vessel with starved cells in SD at 100% of the expected starting density and pumped in fresh medium ( SD + 20 μM hypoxanthine ) to achieve the desired doubling time . Cultures were otherwise treated as described above for L−A+ . For most experiments , we isolated colonies from the end time point and checked percentage evolved ( Methods , “Detecting evolved clones” ) . For L−A+ , we only analyzed time courses for which >90% of population remained ancestral . For A−L+ , significant levels of mutants were generated before and throughout quantification ( S9 Fig ) . Because quantified phenotypes did not correlate strongly with the percentage of mutants ( S11 Fig ) and because mutants accumulated similarly during chemostat measurements and during CoSMO growth rate measurements ( S9B Fig ) , we used the time window for CoSMO growth rate quantification ( about 96 h ) in A−L+ chemostat experiments . We illustrate how we quantify release rate , consumption amount per birth , and death rate in chemostats , using L−A+ as an example . In a lysine-limited chemostat , live cell density [L−A+]live is increased by birth ( at a rate bL ) and decreased by death ( at a rate dL ) and dilution ( at a rate dil ) : d[L−A+]livedt= ( bL−dL−dil ) [L−A+]live ( 6 ) Dead cell density [L−A+]dead is increased by death and decreased by dilution d[L−A+]deaddt=dL[L−A+]live−dil[L−A+]dead ( 7 ) L , lysine concentration in the culturing vessel , is increased by the supply of fresh medium ( at concentration L0 ) and decreased by dilution and consumption ( with each birth consuming cL amount of lysine ) . Finally , hypoxanthine concentration A is increased by release ( either from live cells at rA per live cell per h or from dead cells at rA , d per death ) and decreased by dilution . dAdt=rA⋅[L−A+]live−dil⋅A ( 9 , if live release ) or dAdt=rA , d⋅dL⋅[L−A+]live−dil⋅A ( 10 , if dead release ) Note that at the steady state ( denoted by subscript ss ) , net growth rate is equal to dilution rate ( setting Eq 6 to zero ) : bL−dL=dil ( 11 ) To measure metabolite consumed per cell at steady state , we set Eq 8 to zero cL=L0⋅dil−L⋅dilbL[L−A+]live , ss∼L0[L−A+]live , ss ( 12 ) Here , the approximation holds because the concentration of lysine in chemostat ( L ) is much smaller than that in reservoir ( L0 ) and because birth rate bL is similar to dilution rate dil . To measure death rate at steady state , we set Eq 7 to zero and get dL=dil[L−A+]dead , ss[L−A+]live , ss ( 13 ) Thus , we can measure death rate by measuring the steady-state dead and live population densities averaged over time . To measure release rate at steady state , we can set Eq 9 to zero and obtain rA=dil⋅Ass[L−A+]live , ss ( 14 ) Alternatively , we can use both the pre–steady-state and steady-state chemostat dynamics to quantify release rate and death rate if these rates are constant . For release rate , we multiply both sides of Eq 9 with edil∙t edil⋅tdAdt+dil⋅Aedil⋅t=rA⋅[L−A+]liveedil⋅tord ( edil⋅tA ) dt=rA⋅[L−A+]liveedil⋅t . Because the initial A is zero , we have Aedil⋅t=rA∫0t[L−A+]liveedil⋅τdτ ( 15 ) How do we calculate ∫0tf ( τ ) dτ from experimental data ? The value of integral is always zero at t = 0 . For each time point t + Δt , the integral is the integral at the previous time point t ( i . e . , ∫0tf ( τ ) dτ ) plus Δtf ( t ) +f ( t+Δt ) 2 . If we plot Aedil∙t against ∫0t[L−A+]liveedil⋅τdτ , we should get a straight line through the origin with a slope of rA ( S14 Fig , blue ) . Similarly to Eq 7 , if death rate is constant , we have [L−A+]deadedil⋅t=[L−A+]dead ( t=0 ) +dL∫0t[L−A+]liveedil⋅τdτ ( 16 ) If we plot [L−A+]deadedil⋅t against ∫0t[L−A+]liveedil⋅τdτ , we should get a straight line with a slope of dL ( S14 Fig , gray ) . The two methods ( using only the steady-state data versus performing linear regression on the entire data range ) yielded similar results . We have opted for the latter method because it takes advantage of pre–steady-state data . To detect evolved clones in an L−A+ culture , we diluted it to <1 , 000 cells/mL and plated 300 μL on a YPD plate and allowed colonies to grow for 2–3 d . We randomly picked 20–50 colonies to inoculate into YPD and grow saturated overnights . We diluted each saturated overnight 1:6 , 000 into SD + 164 μM lysine and allowed cultures to grow overnight at 30°C to exponential phase . We washed cells 3× with SD , starved them for 4–6 h to deplete vacuolar lysine stores , and diluted each culture so that a 50-μL spot had several hundred cells . We spotted 50 μL on an SD plate supplemented with 1 . 5 μM lysine ( 10 spots/plate ) and allowed these plates to grow overnight . When observed under a microscope , evolved cells with increased lysine affinity would grow into “microcolonies” of about 20–100 cells , while the ancestral genotype failed to grow ( S7C Fig ) . Occasionally an intermediate phenotype was observed where smaller microcolonies with variable sizes formed , and this phenotype was counted as evolved as well . For a high-throughput version of this assay , we diluted YPD saturated culture 10 , 000× into SD and waited for 3 h at room temperature . We then directly spotted 50 μL on SD plates supplemented with 1 . 5 μM lysine . Ancestral cells formed ≤10-cell clusters , but we could still clearly distinguish ancestor versus evolved clones . To detect evolved clones in an A−L+ culture , we took advantage of the observation that evolved clones with improved affinity for hypoxanthine grew slowly when hypoxanthine concentration was high ( S8A Fig ) . A similar fitness trade-off has been observed for L−A+ [43] and in many other examples [76–79] . From an A−L+ culture , we randomly picked colonies and made individual YPD overnights in a 96-well plate . We diluted YPD overnights 1:3 , 600-fold into SD + 100 μM hypoxanthine or 108 μM adenine , and grew for 16–24 h . Some of these cultures were not turbid while other cultures and the ancestor reached near saturation ( S8B Fig ) . We considered these low-turbidity cultures as evolved , and they generally grew faster than the ancestor in low ( 0 . 4 μM ) hypoxanthine ( S8A Fig , compare blue , gray , and green against magenta ) . For L−A+ , we washed exponential phase cells and diluted each sample to OD 0 . 1 to roughly normalize cell density . We took an initial cell density reading of each sample by flow cytometry , wrapped tube caps in parafilm to limit evaporation , and incubated in a rotator at 30°C . Prep time ( from the start of washing to the initial cell density reading ) took approximately 2 h , during which time the majority of residual growth had taken place . At each time point , we measured live and dead cell densities by flow cytometry; we froze an aliquot of supernatant where supernatant had been separated from cells by filtering through sterile nitrocellulose membrane . We concluded the assay after approximately 24 h , generally aiming for time points every 6 h . At the conclusion of the assay , we quantified hypoxanthine concentration for each sample using the yield bioassay ( Methods , “Bioassays” ) . The slope of the linear regression of integrated live cell density over time ( cells/mL * h ) versus hypoxanthine concentration ( μM ) gave us the release rate . For A−L+ , the starvation release assay was similar , except that the assay lasted longer with less frequent time points to accommodate the longer assay . Pregrowth in 108 μM Ade versus 100 μM hypoxanthine generated similar release rates , and thus we pooled the data . Mutant A−L+ clones were alike , and they grew about 50% slower than the ancestor in excess hypoxanthine ( S8A Fig ) . This has allowed us to rapidly quantify mutant abundance ( S8B Fig; Methods , “Detecting evolved clones” ) . The high abundance of mutants during exponential growth is surprising , especially given the large ( about 50% ) fitness disadvantage of mutants in excess hypoxanthine ( S8A Fig ) . Whole-genome sequencing of a randomly chosen evolved A−L+ clone ( WY2447 ) revealed evidence for aneuploidy ( S8C Fig; Methods , “Genomic analysis” ) . Assuming a chromosomal mis-segregation rate of 0 . 01/generation/cell and incorporating the fitness difference between ancestor and mutant in various hypoxanthine concentrations ( S10A Fig ) , our mathematical models ( S2 Code; S3 Code ) qualitatively captured experimental observations ( S10B and S10C Fig ) . This extraordinarily high mutation rate is possibly due to an imbalance in purine intermediates in a purine auxotroph and is in line with the highest chromosomal mis-segregation rate observed in chromosome transmission fidelity mutants ( up to 0 . 015/generation/cell ) [80] . In low concentrations of hypoxanthine ( <1 μM ) , the fitness difference between mutant and ancestral A−L+ varied from 30% to 70% ( right panel of S10A Fig ) , consistent with the dynamics of mutant A−L+ in chemostats . Ancestral and evolved clones exhibited distinct phenotypes ( S8A and S8D Fig ) . However , measured phenotype values were not significantly correlated with the percentage of mutants at the end of an experiment . This was due to the relatively narrow spread in the percentage of mutants and the relatively large measurement errors ( S11 Fig ) . To measure consumption in exponential cultures , we diluted exponentially growing cells to about 1 × 106 cell/mL in SD supplemented with about 100 μM metabolite and measured cell density ( Methods , “Flow cytometry” ) and metabolite concentration ( Methods , yield assay in “Bioassays” ) every hour over 6 h . For an exponential culture of size N ( t ) growing at a rate g while consuming metabolite M , we have dNdt=gN dMdt=−cgN . Thus , dMdt=−cdNdt . Integrating both sides , we have M ( t ) − M ( 0 ) = −c ( N ( t ) − N ( 0 ) ) . Thus , if we plot M ( t ) against N ( t ) , the slope is consumption per birth . We disregarded time points after which M had declined to less than 10 μM , even though cells could still grow at the maximal growth rate . We also measured consumption after cells fully “saturated” the culture and used intracellular stores for residual growth . We starved exponentially growing cells ( 3–6 h for L−A+ , 24 h for A−L+ ) to deplete initial intracellular stores and inoculated about 1 × 105 cells/mL into various concentrations of the cognate metabolite up to 25 μM . We incubated for 48 h and then measured cell densities by flow cytometry . We performed linear regression between input metabolite concentrations and final total cell densities within the linear range , forcing the regression line through origin . Consumption per birth in a saturated culture was quantified from 1/slope . To measure release rate in an exponentially growing population in excess metabolites , we note that dMdt=rN where M is metabolite concentration , r is the release rate , and N is live population density . Let g be growth rate; then , after integration , we have M ( T ) =∫0TrNdt=∫0TrN ( 0 ) egtdt=rgN ( 0 ) egt|0T ) ≈rgN ( T ) . The approximation holds when N ( T ) >>N ( 0 ) , which is true experimentally . We grew cells in excess metabolite ( lysine or hypoxanthine ) exponentially to time T when OD600 < 0 . 5 ( i . e . , <1 . 6 × 107/mL ) . Supernatants were assayed for released metabolite using the rate bioassay ( Methods , “Bioassays” ) . Because M ( T ) was below the sensitivity of detection ( about 0 . 1 μM; S6 Fig ) for both strains , we used 0 . 1 μM as M ( T ) , growth rate ( 0 . 47–0 . 48/h for L−A+ and 0 . 43–0 . 44/hr for A−L+ ) , and N ( T ) ( 1 . 4–1 . 6 × 107/mL ) to calculate the upper bound for release rate r . High-quality genomic DNA was extracted using the QIAGEN ( Hilden , Germany ) Genomic-tip 20G kit ( CAT Number 10223 ) or the Zymo Research ( Irvine , CA ) YeaStar Genomic DNA Kit ( CAT Number D2002 ) . DNA fragmentation and libraries were prepared [81] using a Nextera DNA Sample Preparation Kit ( Illumina , San Diego , CA ) with 96 custom bar code indices [82] and TruSeq Dual Index Sequencing Primers . Libraries were pooled and multiplexed on a HiSeq2000 lane ( Illumina ) for 150-cycle paired-end reading . A custom analysis pipeline written in Perl incorporated the bwa aligner [83] and samtools [84] for alignment file generation , GATK for SNP/indel calling [85] , and cn . MOPs for local copy number variant calling [86] . Finally , a custom Perl script using vcftools [87] was used to automate the comparison of an evolved clone with its ancestor . All called mutations were validated by visual inspection in the IGV environment [88] . Ploidy was calculated using custom python and R scripts . Read depth was counted for each base and averaged within consecutive 1 , 000-bp windows . Then , the average coverage of each 1 , 000-bp window was normalized against the median of these values from the entire genome and log2 transformed . Transformed data were plotted as box plots for each chromosome/supercontig . All code is publicly available at https://github . com/robingreen525/ShouLab_NGS_CloneSeq . We grew cells to exponential phase in SD + excess supplements . While still at a low density ( <107 cells/mL ) , we measured live and dead cell densities using flow cytometry to yield a dead/live ratio . Because the percentage of dead cells was small , we analyzed a large volume of sample via flow cytometry to ensure that at least 400 ToPro3-stained dead cells were counted so that the sampling error ( 2NN ) was no more than 10% . We also calculated growth rates using optical density readings for the 2 h before and after flow cytometry measurement to yield the net growth rate , g . In exponentially growing cells , dLivedt= ( b−d ) ⋅Live dDeaddt=d⋅Live where b and d are , respectively , birth and death rates of cells . Thus , Live = Live ( t = 0 ) e ( b−d ) t Dead∼d⋅Live ( t=0 ) b−de ( b−d ) t=db−dLive , or d = ( b−d ) ∙ Dead/Live . Thus , the ratio of dead to live cells is the ratio of death rate to net growth rate . The death rate of lys2− cells in excess lysine ranged from 10−4 to 10−3/h . This large variability persisted despite our using the same culture master mix to grow independent cultures . In all experiments , L−A+ cells were grown to exponential phase in SD plus lysine , and washed free of lysine . A−L+ cells were grown to exponential phase in SD plus hypoxanthine , washed , prestarved in SD for 24 h , and washed again . Prestarvation was intended to deplete cellular hypoxanthine storage and to shorten CoSMO growth lag ( S2 Fig ) . We grew spatial CoSMO in two configurations: “column” versus “spotting . ” In the column setting , to prevent potential metabolite cross-contamination , we overfilled non-neighboring wells ( i . e . , 24 wells per deep 96-well plate ) with 2 × SD + 2% agar and covered the surface with a sterile glass plate to form a flat agar surface with no air bubbles . After solidification , we removed the glass plate and removed extra agar between filled wells using sterile tweezers . This results in an agar depth of 3 cm . For the rest of the experiment , when not setting up or sampling , we covered the plate with a sterile lid suspended above the wells by thick toothpicks . We wrapped plates with parafilm to reduce agar drying . We mixed strains at a 1:1 ratio and filtered them through MF membrane ( HAWP04700 from Millipore , Billerica , MA ) to achieve a 3 , 000 cells/mm2 density on the filter ( see [32] for details ) . We then punched 8-mm-diameter disks and transferred one disk to each agarose well , resulting in about 1 . 5 × 105 cells/disk . For each time point , we used tweezers ( ethanol flame sterilized between samples ) to pick 2–3 disks , and suspended each in water prior to flow cytometry measurements . In the spotting setting , in an 85-mm petri dish , we poured about 25 mL 2 × SD + 2% agarose + a small amount of lysine ( generally 0 . 7 μM to minimize the lag phase during CoSMO growth ) to achieve an agar/agarose depth of 5 mm . After solidification , we used a sterile blade to cut and remove 2-mm strips out of the agar to create six similarly sized sectors on the plate , with no agarose connections between them ( S24A Fig ) . We inoculated plates by spotting 15 μL of strains at a 1:1 ratio onto plates , to result in about 4 × 104 cells/patch ( 4-mm inoculum radius ) . Cells were grown and sampled as in the column setting , except that we cut out the agarose patch containing cells , submerged it in water , vortexed for a few seconds , and discarded the agarose . For both setups , we used 9 × 107 total cells as a cutoff for the CoSMO growth rate calculation . We used this cutoff because exponential CoSMO growth rate was observed beyond 9 × 107 total cells , suggesting that no other metabolites were limiting by then . We modified our previous individual-based spatial CoSMO model [32] so that in each time step , metabolite consumption and release of each cell scaled linearly with cell’s biomass to reflect exponential growth . The model used the parameters in Table 1 . The release rate of lysine for each A−L+ cell at each time step was linearly interpolated based on the local concentration of hypoxanthine ( S8 Table ) . We simulated CoSMO growth in two different settings: ( 1 ) cells were initially uniformly distributed on the surface of an agar column and ( 2 ) cells were initially spotted in the middle of an agar pad according to the experimental setup . The simulation domain used for setting ( 1 ) was 500 × 500 μm in the lateral x and y dimensions; for setting ( 2 ) , the agarose domain was 800–960 μm on each side ( 5 μm/grid ) , and the size of the inoculation spot was 1/4 × 1/4 = 1/16 of the agarose domain . In both settings ( 1 ) and ( 2 ) , the z dimension in simulation varied according to the experimental setup ( 5 mm–3 cm ) . For metabolite diffusion within the community , we used either a single diffusion coefficient ( D = 360 μm2/s; S6 Code ) or two diffusion coefficients ( D = 360 μm2/s measured in agarose and D = 20 μm2/s measured in yeast community [32]; S7 Code ) . Both codes are for spotting inoculation , but the inoculation spot can be increased to cover the entire surface . Regardless of the simulation setup , we obtained a similar steady-state community growth rate . When CoSMO achieves the steady-state growth rate , both strains will grow at the same rate as the community ( gcomm ) . This means that L and A concentrations do not change , and Eqs 1–4 become bL−dL=gcomm bA−dA=gcomm rL[A−L+]=cLbL[L−A+]=cL ( gcomm+dL ) [L−A+] rA[L−A+]=cAbA[A−L+]=cA ( gcomm+dA ) [A−L+] Combining the last two equations , we get rArL = cAcL ( gcomm + dL ) ( gcomm + dA ) . Solving this , we get gcomm=− ( dA+dL ) 2+rArLcAcL+ ( dA−dL ) 24 . ( 17 ) For Model iii , given our parameter values ( Table 1 ) , ( dA−dL ) 24≪rArLcAcL . Thus , we obtain gcomm≈− ( dA+dL ) 2+rArLcAcL ( 18; 5 ) rL , lysine release rate of A−L+ , varies with growth rate ( Fig 7 ) . When we focus on doubling times between 5 . 5 and 8 h , a range experienced by CoSMO , then we arrive at the following correlation ( Fig 6B , green dotted line ) : rL = 1 . 853−11 . 388gA , where gA is the net growth rate of A−L+ . Because at steady state , growth rate , gA = gcomm , we have gcomm≈−dA+dL2+rArLcAcL=−0 . 015+0 . 00242+0 . 27 ( 1 . 853−11 . 388*gcomm ) 3 . 1*5 . 4 ( gcomm + 0 . 0087 ) 2 = ( 1 . 853−11 . 388*gcomm ) 0 . 0161 = 0 . 0298−0 . 1833gcomm . That is , gcomm2 + 0 . 200gcomm−0 . 030 = 0 . Thus , gcomm=0 . 10/h ( 19 ) corresponding to a doubling time of 6 . 9 h . To estimate the uncertainty in our prediction of gcomm , we use the variance formula for error propagation . Specifically , let f be a function of xi ( i = 1 , 2 , … , n ) . Then , the error of f , sf , can be expressed as sf2=∑ ( ∂f∂xisxi ) 2 where sxi is the uncertainty of xi . Thus , for each of the six parameters in Eq 18 , we divide its 95% confidence interval ( Table 1 ) by 2 to obtain error s . For lysine release rate , rL , we use the value measured in chemostats with a 7-h doubling time , which closely corresponds to CoSMO doubling time . Summing all terms and taking the square root , we have an error of 0 . 004 for gcomm . Thus , the 95% confidence interval is ±0 . 01 . We did not calculate the uncertainty of our spatial simulation prediction , because we did not solve the spatial model analytically . However , given that predicted community growth rates with or without diffusion are similar ( Fig 7 ) , we expect that the two predictions should share similar uncertainty . | A crown jewel of any scientific investigation is to make accurate and quantitative predictions based on mechanistic understanding of a system . Although quantitative prediction has been the norm and the expectation in physical sciences , living systems are notoriously difficult to predict quantitatively . One of the major challenges is obtaining model parameters . Choosing model parameters to fit data often results in a model that can explain the fitted data but not predict new data . When modeling cells , “guessing” phenotype parameters or “borrowing” parameters from a different genetic background can be highly problematic . In addition , phenotype parameters can vary significantly over time , as cell physiology changes with changing environments , or as cells evolve rapidly . Thus , although parameters are assumed to be constant in most models , this is a far cry from reality and may interfere with quantitative prediction . Here , using a simple engineered yeast community as our case example , we demonstrate that quantitative modeling is possible , but only after overcoming multiple difficulties in parameter measurements . Our approach should be generalizable for modeling communities of interacting cells for which genetic or chemical information of interaction mechanisms is available . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"death",
"rates",
"flow",
"cytometry",
"cell",
"physiology",
"cell",
"death",
"chemical",
"compounds",
"cell",
"processes",
"cloning",
"nucleotides",
"organic",
"compounds",
"cell",
"metabolism",
"metabolites",
"basic",
"amino",
"acids",
"amino",
"acids",
"molecular",... | 2019 | Uncovering and resolving challenges of quantitative modeling in a simplified community of interacting cells |
Progress in epigenetics has revealed mechanisms that can heritably regulate gene function independent of genetic alterations . Nevertheless , little is known about the role of epigenetics in evolution . This is due in part to scant data on epigenetic variation among natural populations . In plants , small interfering RNA ( siRNA ) is involved in both the initiation and maintenance of gene silencing by directing DNA methylation and/or histone methylation . Here , we report that , in the model plant Arabidopsis thaliana , a cluster of ∼24 nt siRNAs found at high levels in the ecotype Landsberg erecta ( Ler ) could direct DNA methylation and heterochromatinization at a hAT element adjacent to the promoter of FLOWERING LOCUS C ( FLC ) , a major repressor of flowering , whereas the same hAT element in ecotype Columbia ( Col ) with almost identical DNA sequence , generates a set of low abundance siRNAs that do not direct these activities . We have called this hAT element MPF for Methylated region near Promoter of FLC , although de novo methylation triggered by an inverted repeat transgene at this region in Col does not alter its FLC expression . DNA methylation of the Ler allele MPF is dependent on genes in known silencing pathways , and such methylation is transmissible to Col by genetic crosses , although with varying degrees of penetrance . A genome-wide comparison of Ler and Col small RNAs identified at least 68 loci matched by a significant level of ∼24 nt siRNAs present specifically in Ler but not Col , where nearly half of the loci are related to repeat or TE sequences . Methylation analysis revealed that 88% of the examined loci ( 37 out of 42 ) were specifically methylated in Ler but not Col , suggesting that small RNA can direct epigenetic differences between two closely related Arabidopsis ecotypes .
Epigenetics , defined as the study of heritable alteration in gene expression without changes in DNA sequence , has greatly expanded our understanding of inheritance [1] . A recent study of DNA methylation by tiling array analysis of Arabidopsis Chromosome 4 in Col and Ler showed that although transposable elements ( TEs ) are often methylated , the methylation in the transcribed regions of genes is highly polymorphic between these two ecotypes [2] . Although epigenetic differences could potentially contribute to evolution [3]–[5] , studies of evolution and natural variation have still been focused mainly on sequence variation , and little is known about the role of epigenetic machinery in these processes . This is primarily due to the lack of evidence for epigenetic natural variation between populations . Small interfering RNAs ( siRNAs ) , as a key player in the epigenetic machinery , have been well documented for their general role in gene silencing at both the transcriptional and post-transcriptional levels [6] , [7] . In Arabidopsis , ∼24 nt siRNAs can direct DNA methylation ( RNA-directed DNA methylation , RdDM ) and chromatin remodeling at their target loci [8] . In the RdDM process , ∼24 nt siRNAs are incorporated into ARGONAUTE 4 ( AGO4 ) -containing complexes and further guide the DOMAINS REARRANGED METHYLTRANSFERASE 2 ( DRM2 ) to de novo methylate their target DNA [9] , [10]; once established , the non-CG methylation could be maintained by DRM2 and/or CHROMOMETHYLASE 3 ( CMT3 ) in a locus-specific manner , and the CG methylation by METHYLTRANSFERASE 1 ( MET1 ) [11] . Recent advances in high-throughput sequencing techniques have enabled the thorough exploration of the small RNAs populations [12]–[16] . Therefore , together with the complete genome sequence , we are able to directly examine whether there are regions specifically matched by siRNAs that differ among ecotypes , a situation that could lead to epigenetic natural variation . FLC , a MADS box transcription factor , is a major repressor of the transition to flowering in Arabidopsis , and many genes coordinately function in flowering time control by regulating the amount of FLC transcript [17] . In addition , allelic variation at FLC , both genetic [18]–[21] and epigenetic [22] , [23] , contributes to the differences in flowering time and vernalization response among accessions , which makes FLC a classic locus for the study of natural variation in Arabidopsis . Previous studies have shown that in Ler , a 1224 base pair ( bp ) nonautonomous Mutator-like transposable element ( TE ) inserted in the first intron of FLC ( FLC-TE-Ler ) [19] was methylated and heterochromatic under the direction of ∼24 nt siRNAs generated by homologous TEs , and mutation of HUA ENHANCER 1 ( HEN1 ) in Ler ( hen1-1 ) , a key component in small RNA biogenesis [7] , released the transcriptional silencing of FLC-Ler [22] . In this study , we discovered a cluster of ∼24 nt siRNAs that are present at high levels in the ecotype Ler and that could direct DNA methylation and heterochromatinization adjacent to FLC promoter [24] . However siRNAs matching to the same region in Col are of low abundance and cannot direct DNA methylation . Furthermore , from comparisons between Ler and Col of small RNA data produced by high-throughput sequencing , we identified at least 68 loci that are matched by significant levels of ∼24 nt siRNAs , and 88% are methylated in Ler but not Col from a set of 42 loci that were examined . . Although siRNA clusters are often heavily methylated [25] and a large proportion of the methylation polymorphisms between Col and Ler are not associated with small RNAs [2] , our data reveal that there could still be considerable small RNA-directed epigenetic natural variation between two ecotypes of Arabidopsis .
In addition to the previously described Mutator-like transposable element ( TE ) inserted in the first intron of FLC [19] in Ler , we found that a region located adjacent to the promoter of the FLC was specifically methylated in Ler but not in Col ( Figure 1A ) . We named this region MPF ( Methylated region near Promoter of FLC ) . Restriction enzymes including AciI , HpyCH4 IV and Fnu4HI , which are sensitive to CpG methylation , were able to cut outside of the MPF but not within this region in Ler ( Figure 1 ) . Notably different from the TE inserted in FLC-Ler , the MPF of Ler and Col share almost identical sequences ( Figure S1 ) . Bisulfite sequencing of MPF ( B1 region , Figure 2A ) revealed that a small region of less than 100 bp was exhibited a very high level of asymmetric methylation ( also called CHH methylation , where H represents A , C or T ) ( Figure 2C ) . This region also demonstrated extensive CpG and CNG ( where N is any nucleotide ) methylation ( Figure 2C ) . In addition , no DNA methylation was found outside the MPF ( the B2 and B3 regions , Figure 2A ) in Ler ( data not shown ) or the MPF in Col ( Figure 3A ) by bisulfite sequencing . Since asymmetric methylation is the hallmark of RdDM [26] , we decided to verify whether there are corresponding siRNAs matching to this methylated region in Ler . Because no methylation was found at the MPF in Col , we speculated that there would be no small RNAs matching to this region . However , four 17 nt tags with very low abundances ( approximately two transcripts per quarter-million , TPQ ) were found in the Col-derived small RNA massively parallel signature sequencing ( MPSS ) datasets [12] . These small RNAs precisely matched both strands of the highly asymmetrically methylated region within MPF ( Figure 2B ) . We performed a small RNA Northern blot hybridization to verify these small RNA in Col and Ler . By using an LNA ( locked nucleic acid ) modified oligonucleotide probe ( Figure 2B ) and a large amounts of RNA enriched for small RNAs ( see materials and method for more details ) , we found that siRNAs complementary to this probe ( MPF-siRNAs ) were more abundant in Ler than in Col ( Figure 2D ) . Published high-throughput small RNA 454 sequencing datasets from Ler [15] confirmed our RNA gel blot results . In those data , six unique 23 to 24 nt small RNAs were found matching to a region of <50 bp at the MPF , in exactly the same region as the Col-derived MPF-siRNAs ( Figure 2B ) . Analyses of additional Col-derived 454 small RNA data [16] , [27] didn't identify any MPF-matching small RNAs , possibly due to lower sequencing depth compared to that of the MPSS data . We performed chromatin immunoprecipitation ( ChIP ) experiments and demonstrated that the MPF in Ler was enriched in H3K9me2 , a characteristic of heterochromatin , in comparison to Col ( Figure 2E ) . These data suggest that the high levels of MPF-siRNAs in Ler could trigger DNA methylation and heterochromatinization at MPF whereas the lower levels in Col might not be sufficient . Next , we investigated methylation at the MPF using silencing pathway mutants in either a Ler background or in lines that had been backcrossed to Ler to have the homozygote FLC-Ler allele . These mutants included hen1-1 , cmt3-7 , ago4-1 , kryptonite-2 ( kyp , a histone H3K9 methyltransferase , also known as SUVH4 , can affect the DNA methylation at some loci[28]–[30] , and drm2 5×Ler ( homozygous drm2 backcrossed five times to Ler ) . Methylation at MPF was sensitive to the deficiency in the RdDM machinery: all mutants tested , with the exception of kyp-2 , completely relieved methylation in all three sequence contexts at MPF ( Figure 3A and Figure S2A ) . Although KYP has been reported to control CNG methylation together with CMT3 [26] , [30] , the methylation at MPF was independent of its function , perhaps because MPF at several hundred base pairs is too small for KYP to maintain the positive feed back between DNA methylation and chromatin modification [30] . Alternatively , in addition to KYP , the heterochromatic feature of this region might be redundantly controlled by other two histone H3K9 methyltransferases , SUVH5 and SUVH6 [31] . In addition , methylation of the nearby TE insertion ( Figure 3B and Figure S2C ) was also sensitive to ago4-1 and hen1-1 ( Figure 3B ) . However , none of these mutants released all DNA methylation at AtSN1 , a retroelement which also undergoes RdDM [26] ( Figure 3C ) . Moreover , AGO4 complementation [15] could not restore DNA methylation at the MPF in ago4-1 ( data not shown ) . This situation resembles the FWA locus whose methylation , once lost in ddm1 ( decrease in DNA methylation 1 ) mutant , is not recovered again even in the presence of wild type DDM1 [32] . The MPF in hen1-4 , a strong hen1 allele in the Col background , had an identical methylation pattern to Col ( Figure 1 ) . Also , the identical methylation pattern of the miRNA deficient mutant dcl1-9 [7] to Ler at MPF ( Figure S2B ) ruled out the possibility that the restricted methylation at MPF is directed by miRNAs [33] . These observations were substantially different from prior analyses of silenced loci , at which DNA methylation was often affected in certain but never all sequence contexts by mutants in the RdDM pathway [26] . Since MPF is methylated and it is near to the TE insertion in FLC-Ler , it was of interest to investigate whether the methylation at MPF is induced by the TE . We examined the methylation status of MPF in several accessions that are also reported to contain transposable elements inserted in the first intron of FLC ( Figure S3A ) [19] , [20] . These were tested by McrBC-PCR [34] ( for Bd-0 , JI-1 , Stw-0 , Kin-0 ( CS1273 ) , and Gr-3 ) and bisulfite sequencing ( for Da ( 1 ) -12 ) . Although the MPF is methylated in Bd-0 , JI-1 and Kin-0 ( CS1273 ) , it remains unmethylated in Stw-0 , Gr-3 and Da ( 1 ) -12 ( Figure S3B , and data not shown for Da ( 1 ) -12 ) indicating that the TE insertions nearby are dispensable for the methylation at MPF . A previous study using 27 Arabidopsis accessions showed that the FLC-TE in Ler was also detected in Dijon-G and Di-2 ( Figure S3A ) but was absent in the closely related Landsberg-0 or Di-1 [18] . McrBC-PCR analysis showed that MPF is methylated in all four of these accessions , even in those without the FLC-TE insertion ( Figure S3C ) , which further confirmed that the methylation at MPF is independent of the TE insertion nearby . To study the origin of the MPF-siRNAs , we found that a 220 bp sequence at MPF is absent in one Kin-0 accession ( CS6755 , different from the Kin-0 ( CS1273 ) accession mentioned above that contains a methylated MPF ) . Further analysis revealed that this difference is caused by the insertion of a non-autonomous hAT element [35] with the typical 8 bp TSD ( target site duplication ) and short terminal inverted repeats ( TIRs ) ( Figure 4 and Figure S1 ) . However , MPF-siRNAs in Ler are probably not derived from other hAT elements because those MPF-siRNAs with the full length information from 454 sequencing in Ler [15] have only one match ( at MPF ) in the genome; also , genomic Southern blot hybridization revealed that Ler do not contains extra copy of this hAT element comparing to Col ( Figure S4 ) . Therefore , the MPF-siRNAs are probably generated from MPF itself . In paramutation , the silenced paramutagenic lines are able to confer the active state of the paramutable lines , and make them become paramutagenic [36] . To test whether the methylated state at MPF in Ler is transmissible , we performed bisulfite sequencing to investigate the DNA methylation status in four F1 lines from the crosses of both Col ♀×Ler ♂ and Ler ♀×Col ♂ , with the single nucleotide polymorphisms ( SNPs ) at MPF ( Figure S1 ) used to distinguish the Col and Ler derived sequencing results ( Figure 5A ) . In addition , twenty-four more lines from reciprocal crosses were tested for their MPF methylation by real-time McrBC-PCR ( Figure 5B ) . These experiments revealed extensive diversity in the methylation status of MPF in each individual line in the F1 generation . This diversity could be summarized in the following way: 1 ) in some lines , the MPF-siRNAs from Ler are able to trigger the de novo methylation at Col-derived MPF; 2 ) in some other lines , not only the Col-derived MPF remains unmethylated , the Ler-derived MPF could even lose its methylation; 3 ) there are also cases in which the Ler-derived MPF remains methylated and Col-derived MPF remains unmethylated , just like their ancestors; therefore the MPF is semi-methylated in the whole plant . The 1 . 2 kb FLC-TE , when inserted into a Col FLC genomic construct , is sufficient to cause reduced expression of FLC in the transgenic lines [19] , therefore , it is unclear whether the MPF has any functional relevance in FLC expression . Interestingly , FLC-Ler could strongly suppress the late flowering phenotype induced by FRIGIDA ( FRI ) and luminidependens ( ld ) , but remains moderately sensitive to other mutants that up-regulate FLC like fca , fve , and fpa [37] . Recently , SUPPRESSOR OF FRI4 ( SUF4 ) has been shown to bind to the promoter of FLC and directly interact with FRI and LD [38] . Moreover , FLC-Ler is again sensitive to FRI in a hen1-1 background [22] suggesting reversible epigenetic alteration might account for this weak response . To address the role of the epigenetic variation at MPF in flowering time control , we used an RNAi approach to artificially methylate MPF in Col , the ecotype in which MPF is originally unmethylated . All transgenic plants used for further analyses had been tested for their successful de novo methylation at MPF by McrBC PCR ( data not shown ) . Both flowering time and FLC expression analysis showed that de novo methylation at MPF does not alter the flowering behavior of wild type Col ( Figure S5 ) . However , since Col is an early flowering ecotype and its FLC expression level is relative low , we can not rule not the possibility that MPF may play a more prominent role in some late flowering backgrounds with higher FLC levels , like FRI or ld . The identification of MPF-siRNAs in Ler- but not Col-derived small RNA data made us wonder whether other loci are differentially and specifically matched by ∼24 nt siRNAs in these ecotypes . Because the MPSS small RNA sequencing data are not readily comparable with the 454 data ( due to length differences in the sequencing reads ) , the small RNA datasets we used for a genome-wide identification are all 454 sequencing data , derived from two recent studies: 247 , 318 unique small RNA sequences from Col [16]and 25 , 981 unique small RNA sequences from Ler [15] . Also , to balance the enrichment of longer siRNAs in the sequencing results of AGO4 precipitated pool from Ler [15] , we only selected for further analyses the siRNA reads of length no less than 23 nt , hence most of the miRNAs and short sRNAs are discarded from both the Col and Ler datasets . Since only the Col genome sequence is complete and the number of sequenced Col derived siRNAs is much greater than that of Ler , in this study , we only analyzed the regions matched by clusters of siRNAs present specifically in Ler , to exclude the interference of genetic alteration and also for higher reliability ( please see materials and methods for details about the bioinformatic analysis ) . The unique siRNA sequences over 23 nt from both Col and Ler were mapped to the genome , respectively , and hits were counted in windows of 100 bp . Although the majority of the ∼24 nt small RNA clusters are conserved between Col and Ler ( data not shown ) , after combining the overlapping regions , 68 unique loci were identified ( including the MPF , locus #57; Table S1 ) . These all shared the characteristic that they were matched by at least three distinct siRNAs within 300 bp in Ler but there were no hits in 1500 bp around the same region in Col ( see Figure 6 for an example ) . Most of these loci are MPF-like , in that the siRNA matches are restricted to a small region ( Figure S6 ) , and their distribution in the genome is quite dispersed ( Figure S7 ) . Twenty-two loci are within known genes , and the other 46 are in intergenic regions ( Table S2 ) . An search of methylation data in Col ( http://signal . salk . edu/cgi-bin/methylome ) [25] demonstrated that all of these loci except locus #60 ( located in a highly methylated region longer than several hundred kb , Table S1 ) were clearly lacking methylation; in addition , 28 loci contain repeat-associated sequences with one end beginning close to or within the small RNA matching region , and 15 loci had matching MPSS small RNA tags [12] ( Table S1 ) . We had also searched the website of DNA methylation information on the fourth chromosome in both Ler and Col background ( http://chromatin . cshl . edu/cgi-bin/gbrowse/epivariation/ ) [2] . For the 13 loci ( #44∼56 ) we identified on the fourth chromosome , six loci are found with methylation signals in their data: five loci ( #46 , 49 , 52 , 54 , 55 ) are found specifically methylated in Ler as expected; one locus ( #53 ) is methylated in both ecotypes but with a much higher methylation signal in Ler comparing to Col . Overall , our results are well supported by the two independent studies on epigenomics and epigenetic natural variation [2] , [25] . We investigated the methylation pattern of locus #10 as an example using bisulfite sequencing . Extensive methylation was found in Ler ( Figure S8 ) , whereas the same region in Col remained unmethylated ( data not shown ) . Other eight randomly selected loci were tested using methylation sensitive McrBC-PCR , and all of them , even those with the minimal number of three unique siRNAs , were methylated in Ler but not Col ( Figure S9 ) . Furthermore , we tested the methylation status of 44 loci ( in which 42 have successful amplification results ) , including all the loci on Chromosome I and II , , by real-time McrBC-PCR ( Figure 7A ) . From these analyses , 88% of the loci ( 37 out of 42 ) were found to be specifically methylated in Ler but not Col , and no locus was found only methylated in Col , strongly supporting the role of ∼24 nt siRNA in triggering epigenetic natural variation ( Figure 7B ) . For the features of these 68 loci showing evidence of small RNA-directed variation in DNA methylation , we looked at the genes either corresponding to or adjacent to these loci within less than 1 kb distance of flanking sequence . Among the 64 genes identified ( some intergenic loci did not have flanking genes within 1 kb upstream and downstream ) , 22 genes were found matched by genic siRNA clusters; 18 genes contained siRNA clusters in their 5′ region and 24 genes with clusters in 3′ regions ( Table S2 ) . Among the 22 genic regions , six were transposable elements , consistent with the role of transposable element in epigenetic regulation [39] . Moreover , many of these genes are reported or predicted to have important functions ( Table S2 ) . Therefore , additional investigation of these genes may help us to understand the role of epigenetic alteration in evolution and natural variation .
Natural variation is a fundamental aspect of biology , and the implications of natural variation for deciphering the genetics of complex agricultural traits have been widely used . Recent progress in epigenetics has revealed mechanisms that can heritably regulate gene function without alteration of primary nucleotide sequences . Although the importance of epigenetic natural variation have become more and more noticed [3] , [5] , the role of epigenetic regulation in evolution has been less well studied due in part to limitations in the techniques used for the investigation of epigenetic variation among natural populations . Recently , substantial improvements in high-throughput analysis approaches have made it possible for the effective detection of variation in DNA methylation , histone modifications and small RNA abundances [2] , [12]–[16] , [25] , [40] . Small RNAs that can target DNA methylation and chromatin modifications have been proposed as a potential source in inherited epigenetic differences [3] , and the latest techniques offer rapid and relatively inexpensive means for the profiling of small RNAs . In this study , we discovered that a hAT element adjacent to the promoter of FLC , which we named MPF , is methylated and heterochromatic in Ler but not Col because of their differences in the abundance of corresponding siRNAs . Furthermore , by comparisons between Ler and Col of publicly available small RNA data produced by high-throughput sequencing [15] , [16] , we identified at least 68 loci that are matched by significant levels of ∼24 nt siRNAs , and 88% examined loci are methylated specifically in Ler but not Col . Our data reveal that there could be a considerable amount of small RNA-directed epigenetic natural variation between two ecotypes of Arabidopsis . Although we identified dozens of loci , this analysis is still far from saturating . A Sadhu element ( At2g10410 ) , which was reported to be epigenetically silenced in Ler and other 18 strains but highly expressed in Col , did not show up among the 68 loci [41]; although bisulfite sequencing revealed that this element contains CNG and asymmetric methylation in Ler , which is presumably siRNA-directed to some extent [41] . Furthermore , hundreds of additional loci with one or two hits specifically in Ler ( data not shown ) may also be silent; these may be better characterized when additional Ler small RNA and genome sequence data become available . Two examples of siRNA-associated , naturally-occurring epigenetic variation have been well studied in plants , including the phosphoribosylanthranilate isomerase ( PAI ) gene family in Arabidopsis and paramutation in maize [36] . In some Arabidopsis ecotypes , two PAI genes form an inverted repeats that may generate siRNAs and silence related members in the same gene family [42] . Paramutation , the allele-dependent transfer of heritable silencing state from one allele to another [36] , is associated with another type of repeats , the tandem repeats . MEDIATOR OF PARAMUTATION 1 ( MOP1 ) [43] , whose deficiency disrupts paramutation , is an ortholog of the Arabidopsis RDR2 ( RNA Dependent RNA polymerase 2 ) , an essential component of RNAi machinery [6] . Notably , epigenetic variation at the MPF is quite different from these two cases: first , neither inverted- nor tandem-repeats features were found at MPF or elsewhere in the genome with similar sequence; second , the level of MPF-siRNAs is high in Ler and low in Col , instead of all-or-none; third , the restricted location of MPF-siRNAs is markedly different from the dispersed distribution of siRNAs from most inverted or tandem repeats [12] . Although paramutation phenomenon had been well documented , the details of how the silencing signal is transmitted from one allele to the other in the F1 heterozygote are still less understood . In our study , the diverse methylation status among individuals in F1 generation of the reciprocal crosses from Col×Ler indicate that there might be a reprogramming stage shortly after fertilization , in which the DNA or chromatin are open to modifiers like the MPF-siRNA containing RISC ( RNA induced silencing complex ) from Ler . However , this open stage must be very short , and when it is over , the epigenetic state , no matter active or silenced , will be maintained in the following developmental processes , so that the unmethylated state of Col-derived MPF and the methylated state of Ler-derived MPF could well maintained in Ler ♀×Col ♂line #2 ( Figure S5A ) . Thus far , the function of ∼24 nt siRNAs in plants has mainly been ascribed a role in silencing transposable elements and repeat-associated sequences [39] . Thus , it is unclear how Ler and Col , both with the functional RNAi machinery , might acquire many siRNA-directed epigenetically variable loci . One characteristic of MPF-siRNAs , their very restricted location ( all matching to a region less than 50 bp ) , may confer on them more flexibility than other , larger silent loci . Genetic variability ( due to insertion , deletion and point mutation ) occurs stochastically , at very low frequency , primarily irreversibly and is often recessive . In contrast , heritable epigenetic variability may be more appropriate to regulate , rather than disrupt or create , gene function , and thus may be an ideal or more dynamic force for evolutionary change of gene regulation .
The Bd-0 ( CS962 ) , JI-1 ( CS1248 ) , Stw-0 ( CS1538 ) , Gr-3 ( CS1202 ) , Kin-0 ( CS1273 , CS6755 ) , Da ( 1 ) -12 ( CS917 ) , Dijon-G ( CS910 ) , Di-1 ( CS1108 ) , Di-2 ( CS1110 ) , and La-0 ( CS1299 ) accessions of Arabidopsis were acquired from ABRC; hen1-1 ( Ler background ) , hen1-4 ( Col background ) , and dcl1-9 mutants were described before [22]; cmt3-7 , kyp-2 , ago4-1 , and drm2 5×Ler were generous gifts from Steve Jacobsen at UCLA . The AGO4 complementation lines were kindly provided by Gregory J . Hannon at CSHL and Yijun Qi at NIBS . RNAs were extracted from 20-day-old , soil-grown plants . 32P end-labeled LNA probe was used for hybridization . Total RNAs were extracted using Trizol solution ( Invitrogen ) from 20-d-old soil-grown plants and dissolved in RNase free water . Small sized RNAs were enriched by adding the same volume of 8M LiCl and centrifuging at 12 , 000rpm for 30 min at 4°C . RNA filter hybridizations were carried out as previously described [44] . LNA probe [45] was used for hybridization ( 5′- cgagcAgtGgcGgatCcaaga-3′; uppercases represent modified nucleotides ) . The ChIP assays were performed using 20-d-old soil-grown plants and as previously described [46] . Antibodies against H3K9me1 ( 07-450 ) , H3K9me2 ( 07-441 ) and H3K9me3 ( 07-442 ) were from Upstate Biotechnology . The genomic DNA from Col was used as a template for PCR amplification using the primer pairs ( CX2004: ctcgagATTTTTGTGGTAATATATATATA and CX2005: agatctACATCAATCCAAGTTCAAGC , carrying the XhoI and BglII sites , respectively ) . The PCR products were sequentially inserted into pUCC-RNAi vector using the XhoI/BglII and BamHI/SalI sites for both the sense and antisense orientations . The stem-loop structured fragment was cut off and further cloned into a modified pCambia1302 vector ( pCambia1302-LX-1 ) and used for plant transformation ( XF718 ) . All transgenic plants used for further analyses had been tested for their successful de novo methylation at MPF . Genomic DNA was isolated from rosette leaves of 4-week-old , soil-grown plants . Southern blots was performed as previously described [22] using PCR products amplified from FLC promoter as the probe ( Figure 1 ) . Bisulfite sequencing experiments were performed as previously described [47] . Primers with one end in FLC-TE and the other in FLC were designed to specifically amplify the FLC-TE and exclude other TEs in the genome . Only the cytosines within TE were counted for methylation analysis of FLC-TE in Figure 3 . McrBC-PCR experiments were performed as previously described [34] , [47] , Equal amounts of McrBC-digested and non-digested DNA were used for PCR amplification . Real-time McrBC-PCR was performed to quantitatively measure the methylation level . The primer information for these experiments could be found in Supporting Information ( Text S1 ) . After discarding smaller ( <23 nt ) and redundant sequences , 247 , 318 unique small RNA sequences in Col and 25 , 981 unique small RNA sequences in Ler were used for further analysis . All these siRNAs were mapped to the Col genome by BLAST [48] and PERL scripts , and the numbers of perfect matches were counted per 100 bp . Next , regions contain more than 3 hits within 300 bp in Ler but no hits in 1 . 5 kb at the same region in Col ( Figure 6 ) were filtered out and overlapping regions were artificially combined . Col derived small RNA dataset was downloaded from NCBI GEO ( GSE5228 ) , and Ler derived small RNA sequences from NCBI GenBank ( DQ927324-DQ972825 ) . The Arabidopsis genome ( Col ) information was provided by TIGR ( release version 5 ) . Gene positions were annotated according to TAIR's SeqViewer data . Tandem gene duplication information was provided by TIGR ( tandem_gene_duplicates . Arab_R5 ) . | Phenotypic variation has been mainly attributed to their differences in genetic materials , i . e . , the DNA sequence . The advances in Epigenetics in past decades has revealed it as a fundamental mechanism that could inheritably influence gene function without change in DNA sequence , but by modulating chemical modifications on DNA itself ( methylation ) , or on histone proteins , which package the DNA further into nucleosome . Nevertheless , the roles of epigenetic regulation in natural variation were not explored much because of the limitation in high-throughput analytical tools . A recent study in model plant Arabidopsis showed that there are many DNA methylation polymorphisms between the two ecotypes . In plant , a subset of RNA named small interfering RNA ( siRNA ) , is capable of triggering the epigenetic modifications on DNA or histone at their target region with complementary nucleotide sequences . Here , we took a view from the small RNA side and by applying molecular and bioinformatic approaches we showed that the same region could be led to a different epigenetic status because of the difference in their corresponding small RNA abundance and between the two closely related Arabidopsis ecotypes , suggesting that there could be small RNA-directed epigenetic differences among natural populations . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"genetics",
"and",
"genomics/epigenetics",
"plant",
"biology/plant",
"genetics",
"and",
"gene",
"expression",
"genetics",
"and",
"genomics/plant",
"genomes",
"and",
"evolution"
] | 2008 | Small RNA-Directed Epigenetic Natural Variation in Arabidopsis thaliana |
The family Polydnaviridae is of interest because it provides the best example of viruses that have evolved a mutualistic association with their animal hosts . Polydnaviruses in the genus Bracovirus are strictly associated with parasitoid wasps in the family Braconidae , and evolved ∼100 million years ago from a nudivirus . Each wasp species relies on its associated bracovirus to parasitize hosts , while each bracovirus relies on its wasp for vertical transmission . Prior studies establish that bracovirus genomes consist of proviral segments and nudivirus-like replication genes , but how these components are organized in the genomes of wasps is unknown . Here , we sequenced the genome of the wasp Microplitis demolitor to characterize the proviral genome of M . demolitor bracovirus ( MdBV ) . Unlike nudiviruses , bracoviruses produce virions that package multiple circular , double-stranded DNAs . DNA segments packaged into MdBV virions resided in eight dispersed loci in the M . demolitor genome . Each proviral segment was bounded by homologous motifs that guide processing to form mature viral DNAs . Rapid evolution of proviral segments obscured homology between other bracovirus-carrying wasps and MdBV . However , some domains flanking MdBV proviral loci were shared with other species . All MdBV genes previously identified to encode proteins required for replication were identified . Some of these genes resided in a multigene cluster but others , including subunits of the RNA polymerase that transcribes structural genes and integrases that process proviral segments , were widely dispersed in the M . demolitor genome . Overall , our results indicate that genome dispersal is a key feature in the evolution of bracoviruses into mutualists .
Long-term associations between multicellular organisms and microbes are widespread . In the case of bacteria and fungi , several taxa contain species that have evolved into vertically transmitted , obligate mutualists or pathogens [1]–[3] . Traits inherited from ancestors and acquired by horizontal gene transfer have both contributed to the initiation and maintenance of these obligate associations [4]–[6] . In contrast , small effective population size associated with vertical transmission and increased levels of genetic drift appear to differentially affect genome size and architecture . Bacteria consistently exhibit size reductions due to mutational bias that causes deletions [7]–[9] , while fungi trend toward genome expansion due to gains in mobile elements , intronic sequences , and other types of non-coding DNA [5] , [10]–[13] . All viruses require another organism to persist and propagate with most species being horizontally transmitted by virions that are produced through replication . This lifestyle results in viruses often being pathogenic and having compact genomes with evolutionary rates that are usually much higher than the host organisms they infect [14] . Obligate associations occur when a viral genome integrates into the host germline to form a vertically transmitted endogenous viral element ( EVE ) [15] , [16] . EVEs deriving from many types of viruses have been identified including some of ancient origin that have reached fixation in host populations and can no longer remobilize . Most EVEs are subject to their host's neutral rate of evolution , which results in persistence as fragments of the ancestral viral genome rendered non-functional by mutation [15] , [17] . A few single genes or regulatory domains of viral origin have also been identified where natural selection has led to new , non-viral host functions [15] , [18] , [19] . The family Polydnaviridae is of interest because it is the only known example of viruses that have evolved into vertically transmitted agents that benefit their hosts yet do so by continuing to function in many respects like the viruses they evolved from [20]–[22] . As such , polydnaviruses ( PDVs ) have evolved into mutualists and provide a study system for understanding how their obligate associations with hosts affect viral genome architecture and function . All PDVs are associated with insects called parasitoid wasps ( Hymenoptera ) , which reproduce by laying eggs into other arthropods ( hosts ) their offspring consume [23] . The genus Bracovirus ( BV ) is associated with ∼50 , 000 wasp species in the family Braconidae: forming a monophyletic assemblage called the microgastroid complex that evolved ∼100 million years ago ( Mya ) [24]–[26] . Each wasp species in the complex carries a unique BV that persists in every cell of every individual as a provirus . However , BVs only replicate in calyx cells that are located at the base of the ovaries near the lateral oviducts [27]–[30] . BV virions package multiple circular , double-stranded ( ds ) DNAs of large aggregate size , which are released by lysis of calyx cells and stored in the oviducts [31] , [32] . Females inject eggs containing the integrated provirus plus a quantity of virions into host insects they parasitize . The DNAs delivered by these virions integrate into the genome of infected host cells , while expression of virulence genes on these DNAs generates products that alter the physiology of hosts in ways that wasp offspring depend upon for survival [32]–[35] . BVs differ from other known EVEs of ancient origin because they retain the ability to replicate in wasps and produce infectious virions . However , BVs also differ from most viruses because replication in wasps produces virions that are incapable of replicating in the host insects wasps parasitize . The net result is that each BV relies on its associated wasp for vertical transmission as a provirus while each wasp relies on its BV to produce replication-defective virions for delivery of virulence genes needed to successfully parasitize hosts . The monophyly of the microgastroid complex strongly suggests BVs evolved from an ancient virus that integrated into the germline of an ancestral braconid . Insights into the identity of this ancestor come from transcriptome studies of ovaries from three wasp species ( Cotesia congregata , Chelonus inanitus , Microplitis demolitor ) , which identify more than 30 homologs of genes with predicted functions in replication from another group of insect-infecting DNA viruses called nudiviruses [27] , [28] , [36] . The family Nudiviridae is also the sister taxon to the family Baculoviridae that likewise infects insects . Most nudiviruses and baculoviruses are virulent pathogens that package a single large , circular dsDNA ( >90 kb ) genome into virions containing all genes required for infection of hosts and replication [37]–[39] . This suggests the nudivirus ancestor of BVs initially integrated into the germline of a wasp as a linear proviral DNA [20] . In contrast , BV genomes have since changed in a manner that has resulted in: 1 ) all of the nudivirus-like genes being integrated in the genomes of wasps and transcribed in calyx cells but none residing on the DNAs that are packaged into virions [27] , [28] , [36] , 2 ) the DNAs packaged into virions encoding multiple virulence genes [40]–[44] , and 3 ) almost none of these virulence genes being transcribed in wasps but most being transcribed in the hosts that wasps parasitize [33] , [45] , [46] . The circular , dsDNAs in BV virions are referred to as the encapsidated form of the genome [20] , [33] . In turn , these DNAs are referred to as proviral segments when integrated in the genome of wasps , while the proviral segments and nudivirus-like genes together constitute the BV proviral genome [41] , [47] , [48] . Screening and sequencing of BAC genomic clones from four species in two genera ( Glyptapanteles indiensis , G . flavicoxis , Cotesia congregata , C . sesamiae ) have previously shown that BV proviral segments reside in multiple loci in the genomes of wasps [41] , [47] , [48] . Sequencing of BAC clones from C . congregata further show that 10 nudivirus-like genes reside in an 18 kb domain referred to as the nudivirus gene cluster [27] . These data combined with evidence that all BVs evolved from a common nudivirus ancestor have further led to the suggestion that proviral segment loci and the nudivirus cluster are physically linked in the genomes of wasps [22] , [27] , [47] . BAC clone sequence data , however , are too limited to provide direct evidence such linkages exist . In addition , many of the nudivirus-like genes identified in transcriptome studies [27] , [28] , [36] do not reside in the nudivirus cluster identified from C . congregata . As a result , the location of most nudivirus-like genes in the genomes of wasps , including several experimentally shown to be essential for replication [49] , is also unknown . Taken together then , prior studies clearly establish that BV proviral genomes consist of two components: proviral segments organized into loci and nudivirus-like genes [20] , [22] , [27] , [28] , [32] , [41]–[43] , [47]–[49] . In contrast , how these components are organized in relation to one another and where in the genomes of wasps most nudivirus-like genes identified from transcriptome studies reside is unknown for any species . Such information is important to issues ranging from understanding how BVs function to how genome content and architecture compares to nudiviruses and baculoviruses . The only means of addressing these questions though is through whole genome sequencing . In this study , we sequenced the microgastrine braconid Microplitis demolitor , which carries M . demolitor bracovirus ( MdBV ) and diverged from wasps in the genera Cotesia and Glyptapanteles ca . 53 Mya [24] . Assembly of the M . demolitor genome shows that MdBV proviral segments reside in multiple loci and that some nudivirus-like genes are clustered as found previously in Glyptapanteles and Cotesia wasps . However , we also determined that the MdBV nudivirus-like cluster is much larger than previously found in C . congregata and that a number of nudivirus-like genes that are functionally essential for replication are widely dispersed in the M . demolitor genome . Finally , our results provide direct evidence that MdBV proviral segment loci are not closely linked physically to one another or to the nudivirus-like cluster .
We generated a draft genome sequence for M . demolitor using Illumina technology . The haploid genome size was estimated to be 241±6 Mbp by flow cytometry using wasp cell nuclei normalized to nuclei of Drosophila virilis . Based on this estimate , the M . demolitor genome was sequenced to 26× using haploid male genomic DNA from a lab culture maintained for more than 20 years with no introduction of additional field material . Sequencing of multiple small and large insert paired-end libraries ( 180 bp , 1 . 5 kb , 5 kb , and 10 kb ) produced 1 . 04 billion raw reads . SOAPdenovo v2 . 04 ( 3 ) was employed with K = 49 to assemble the 180 bp-insert library reads into 357 , 737 contigs greater than 100 bp , totaling 195 , 839 , 919 bp with a contig N50 of 1 , 585 bp . Scaffolding with iteratively longer-insert mate-pair libraries followed by GapCloser v1 . 12 resulted in 5524 scaffolds consisting of two or more contigs in appropriate order and orientation separated by regions of approximately known lengths of unknown nucleotides . Remaining sequence data consisted of 47727 contigs greater than 100 bp . The scaffold N50 was 323 , 181 bp , the singleton contig N50 173 bp , and the total assembled genome sequence including intra-scaffold gaps was 258 , 751 , 082 bp . As discussed below , a total of 40 scaffolds with a cumulative size of 19 . 3 Mb contained elements of the MdBV proviral genome . Evidence supporting gene models ( see Methods ) identified 1 , 737 genes in the 40 scaffolds containing MdBV components of which 1 , 713 were predicted protein-coding sequences and 24 were tRNAs ( Table S1 ) . In addition , the estimated aggregate size of all components of the MdBV proviral genome accounted for less than 1% of the M . demolitor genome . Our next goal was to identify regions of the M . demolitor genome that contained MdBV proviral segments . This was accomplished using a combination of previously published and newly generated data . As background , each MdBV virion produced during replication contains only one circular dsDNA segment [50] . Thus , the total complement of DNAs in the encapsidated genome of MdBV are not present in each virion but instead are distributed among the total population of virions produced during replication . In addition , the circular , dsDNAs in MdBV virions are non-equimolar in abundance , which results in delivery of a higher copy number of some DNAs to parasitized hosts than others [50] . Non-equimolar abundance of DNAs in virions from other wasp species [summarized by 20]–[23] , [32] , [33] combined with data showing that Chelonus inanitus BV also packages a single DNA per virion [51] suggests these features apply to BVs generally . The encapsidated form of the MdBV genome was previously analyzed by isolating DNA from virions followed by construction of plasmid libraries that were Sanger sequenced . These data assembled into 15 circular dsDNAs ( named A through O ) , which had an aggregate size of 190 kb [43] . However , this approach can lead to misassembly or omission of segments due to their non-equimolar abundance and the presence of repetitive DNA [40] , [47] . Segments are also easily missed in the absence of having a reference proviral genome , which was a central goal of this study . Thus , we re-sequenced the circular , dsDNAs in MdBV virions by Illumina and then mapped these reads to the assembly of M . demolitor genome . This approach produced 50 million 100 bp read pairs . After quality filtering , 99% of 37 million read pairs were successfully mapped to M . demolitor genome scaffolds . The number of MdBV reads mapped to M . demolitor scaffolds containing proviral segments ranged from 281 , 000 to 28 million . This data set recovered the 15 segments identified previously with the exception of segment O , which was only partially recovered ( Figure 1 ) . Segment O contains large repetitive regions [43] , which likely prevented its successful assembly in this study . Read mapping identified the location of these DNAs as proviral segments in the M . demolitor genome ( Figure 1 ) . Segments A , E , G , I , and K were slightly larger than previously reported [43] , while segment D was split between two scaffolds and one contig totaling 13 , 691 bp compared to the previously published size of 7 , 823 bp [43] ( Figure 1 ) . We also identified 10 previously unknown segments named K1 , owing to sequence similarity with Segment K , and P through X . Thus a total of 25 proviral segments with an aggregate size of 278 kb are amplified and packaged into MdBV virions . In comparison , BVs from wasps in the genera Cotesia and Glyptapanteles have 30 to 35 proviral segments with aggregate sizes ranging from 517 to 731 kb [40] , [41] , [47] . Comparative analyses of baculovirus genomes suggest all species share 37 core genes of which approximately half are required for replication [38] , [59] . Nudiviruses , which diverged from baculoviruses ∼300 Mya [60] , share 20 baculovirus core genes [38] , [39] , including a DNA polymerase predicted to replicate the viral genome , a DNA dependent RNA polymerase comprised of four subunits ( lef-4 , lef-8 , lef-9 , p47 ) , and several structural genes with unique promoter features that are specifically recognized and transcribed by the viral RNA polymerase . Previous transcriptome sequencing by Illumina of M . demolitor ovaries during MdBV replication identified 41 genes with homology to nudivirus genes [28] . These include the four RNA polymerase subunits , several structural genes , and multiple tyrosine recombinases named integrases ( int ) , unknown from baculoviruses , but related to a baculovirus gene named vlf-1 . A nudivirus/baculovirus-like helicase with putative roles in DNA replication was identified but a nudivirus/baculovirus-like DNA polymerase was not , which suggested that amplification of proviral segments during replication requires a wasp DNA polymerase [28] . The remaining nudivirus-like genes included 11 proteins unknown from baculoviruses [27] , [28] . MdBV replication in calyx cells is extremely high with virion production exceeding replication levels for baculoviruses [28] . Experimental studies show the predicted MdBV RNA polymerase subunits form a functional holoenzyme that transcribes the nudivirus-like structural genes [49] . Nudivirus-like genes with predicted roles in capsid and envelope formation are also required for virion formation , while vlf-1 and int-1 are required for circularization of MdBV proviral segments [49] . Proteomic analysis further shows most predicted nudivirus-like structural proteins are present in MdBV virions [49] , while studies with C . congregata and C . inanitus indicate homologs of these structural proteins are present in CcBV and CiBV virions [27] , [36] . Thus , several lines of evidence strongly support that the BV RNA polymerase and several structural genes that are nudivirus/baculovirus homologs retain ancestral functions essential for replication and virion assembly . To identify the nudivirus-like genes in the M . demolitor genome , we searched our assembly using BLASTN and TBLASTN with previously identified transcript and protein sequences as queries [28] , [36] . This resulted in identification of all previously identified nudivirus-like genes plus a few unrecognized genes of potential nudivirus origin located on 29 scaffolds ( Table S1 ) . None of the new genes were homologs of a viral DNA polymerase . In addition , none of the nudivirus-like genes contained introns or were flanked by WIMs . Annotation indicated some of these genes resided in a multigene cluster , others were duplicated genes arrayed in tandem , and the balance were single genes separated by large stretches of intervening wasp DNA . The size and number of scaffolds together with large intervening regions of wasp genes collectively indicated MdBV nudivirus-like genes were widely dispersed in the M . demolitor genome . While proviral segment loci and nudivirus-like genes have been suggested to be physically linked in the genomes of wasps [21] , [22] , [27] , our assembly of the M . demolitor genome indicated most nudivirus-like genes , including those in the nudivirus-like gene cluster , reside in locations distant from proviral segments . The only direct physical linkage we identified was in Mdem_scaffold_0157 , which contained the nudivirus-like gene HzNVorf93 and 5 . 4 kb away MdBV proviral segment T ( Table S2 ) . We identified 18 wasp gene families shared between scaffolds containing nudivirus-like genes and proviral segments ( Table S2 ) . These included members of a gene family with an EB module present in five scaffolds containing nudivirus-like genes and also the scaffold containing proviral segment U . Several members of a protein tyrosine kinase family were also present on a scaffold ( Mdem_scaffold_0407 ) containing lef-4 , and the scaffolds containing proviral locus 3 ( Mdem_scaffold_0014 ) and proviral locus 8 ( Mdem_scaffold_0025 ) . A family of histone genes was shared among the scaffolds containing proviral loci 1 , 3 , 5 and 7 , which suggested these genes may link the genomic neighborhoods of these proviral segments ( Table S2 ) . Members of 13 other wasp gene families were detected on scaffolds containing nudivirus-like genes . Among these was a family of MFS sugar transporter domain-containing proteins present on two scaffolds ( Mdem_scaffold_0004 , Mdem_scaffold_0938 ) containing nudivirus-like genes ( odv-e66 , lef-4 ) . This finding is of potential interest because two MFS sugar transporter genes reside on homologous proviral segments of GiBV and GfBV , which provides an example of duplication and transfer of wasp genes into BV proviral segments [41] .
This study advances our understanding of BVs by providing the first overall picture of proviral genome organization . Unlike the compact , circularized genomes of nudiviruses and baculoviruses , our results show that proviral segment loci and nudivirus-like genes are highly dispersed in the M . demolitor genome . Our results also show that none of the MdBV proviral segment loci are physically closely linked to the nudivirus-like gene cluster . Parasitoid wasps are among the most species-rich animal groups on Earth with estimates suggesting more than 1 , 000 , 000 species worldwide [23] , [25] yet only one parasitoid wasp genome ( Nasonia vitripennis ) has been sequenced , assembled and annotated [63] . N . vitripennis belongs to a taxon of Hymenoptera that is distantly related to microgastroid braconids and has no association with polydnaviruses . Thus more broadly our results provide a genome for a second parasitoid wasp and the first species that is a polydnavirus carrier . Illumina sequencing the DNAs in MdBV virions and mapping these reads back to the M . demolitor genome identified several DNA segments an earlier study failed to detect [43] . The 25 proviral segments now identified in 8 loci likely represent most if not all of the DNAs packaged into MdBV virions . We also identified all nudivirus-like genes found previously by transcriptome analysis of M . demolitor ovaries plus several unrecognized variants of these genes . In our view , the most important new findings from these data are: a ) the nudivirus-like gene cluster of MdBV contains many more genes and overall is much larger than recognized from earlier data generated from C . congregata [27] , b ) most of these genes are structural components of BV virions , and c ) the four nudivirus-like RNA polymerase subunits previously shown to regulate expression of BV structural genes reside outside of the nudivirus-like gene cluster as single genes that are widely dispersed in the M . demolitor genome . Like other large DNA viruses , baculoviruses and nudiviruses exhibit high diversity in gene content outside of their core gene sets [37] , [38] . It is also well known that different lineages of baculoviruses and nudiviruses have acquired many genes from their arthropod hosts and other organisms by horizontal gene transfer and other mechanisms . Given this , it is fully possible some genes from the nudivirus ancestor of BVs remain unidentified given our reliance on sequence similarity for recognition of genes of nudivirus origin [37] , [38] . Primary structure together with proximity to conserved nudivirus-like genes identified four hypotheticals in the MdBV nudivirus-like gene cluster that potentially derive from the nudivirus ancestor . In contrast , identifying genes outside the nudivirus-like gene cluster derived from the nudivirus ancestor will be very difficult in the absence of data linking a given product to replication or other virus-related activities . The dispersed architecture of the MdBV proviral genome is remarkable in light of the very high levels of replication that occur in calyx cells following pupation of female wasps . Although viral genomes are typically viewed as a contiguous stretch of DNA or RNA , our results clearly show that dispersal of BV genomes does not functionally impede high-level amplification of a portion of the proviral genome or production of virions . We suggest the physical separation of nudivirus-like genes required for virion formation from the proviral segments containing virulence genes is selectively advantageous for wasps because it assures vertical transmission of the entire proviral genome but prevents any replication machinery from escaping , which could be deleterious to wasp offspring developing in a host . Our results also suggest two trans-acting factors play critical roles in linking the physically separated components of the MdBV proviral genome together . First , the two nudivirus-like integrases ( int-1 , vlf-1 ) , once transcribed and translated in calyx cells , likely use WIMs to recognize all proviral segments for processing regardless of their location in the wasp genome . Second , the MdBV RNA polymerase holoenzyme , once made , specifically transcribes the nudivirus-like structural genes through promoter recognition , which is similar to baculovirus RNA polymerases that also specifically transcribe structural and other late viral genes [49] . Our analysis of upstream sequence indicates some MdBV structural genes have baculovirus-like late gene promoter motifs but others do not , which suggests the promoter sequences BV RNA polymerases recognize differ somewhat from those of their nudivirus/baculovirus ancestors . Identification of these recognition sequences is not amenable to computational analysis and will require experimental studies . Based on the baculovirus literature [38] , we hypothesize wasp RNA polymerase II transcribes the MdBV integrase and RNA polymerase genes , but the factors responsible for restricting transcription to only calyx cells are unknown . The absence of any baculovirus/nudivirus-like DNA polymerase in the M . demolitor genome further strengthens earlier conclusions that a wasp DNA polymerase ( s ) amplifies MdBV proviral DNAs prior to their excision , circularization , and packaging [28] , [49] . However , the specific polymerase responsible also remains unidentified . In contrast , it has long been known the DNA segments in BV virions are non-equimolar in abundance [40]–[42] , [50] , which our read mapping data indicate is due to proviral segments in different loci being differentially amplified . Similar levels of MdBV proviral segment amplification in loci 1 and 2 are also broadly similar with recent findings for CcBV where multiple adjoining proviral segments are co-amplified before processing into circularized DNAs [64] . Although most nudiviruses and baculoviruses establish systemic lytic infections that are fatal to hosts , one nudivirus has been identified that in vitro establishes long-term persistent infections associated with integration into the host genome [65] , [66] . Such latent infections can also be reactivated . This suggests the nudivirus ancestor of BVs may have established a latent infection following integration of one or more copies of its genome into the germline of the braconid ancestor of microgastroids . This integration event was then followed by a series of modifications and rearrangements to arrive at the current dispersed architecture shown here for MdBV in M . demolitor . Reconstruction of these events for BVs generally is theoretically possible through comparative data of wasp species in the microgastroid complex with different divergence times . However , with near complete data on proviral genome architecture limited to M . demolitor and partial data available for just 4 other species [41] , [47] , only a few suggestive patterns are currently possible . First , experimental studies in M . demolitor combined with conservation of these nudivirus-like genes in two other microgastroid wasps ( C . congregata , C . inanitus ) [27] , [36] , [49] strongly suggest natural selection has maintained the ancestral functions of these factors in virion formation despite their dispersal in the genomes of wasps . Second , the conserved synteny of predominantly structural genes in the nudivirus-like cluster of M . demolitor and C . congregata suggests this domain represents an initial integration site for the nudivirus ancestor , and that maintenance of these genes in a cluster is functionally important for virion formation . These data also indicate the MdBV and CcBV nudivirus-like clusters have remained stable since divergence 53 Mya , which suggests dispersal of the other nudivirus-like genes occurred relatively early in BV evolution . Comparative sequence data from additional BV-carrying wasps will reveal whether dispersal is highly variable or dispersed genes reside in similar locations in the genomes of different wasps . Third , the distribution of MdBV proviral segment loci indicates these domains are also not clustered in the M . demolitor genome , while the presence of only one nudivirus-like gene near a proviral locus indicates that proviral loci and nudivirus-like replication genes reside distantly with respect to one another . We do identify a few wasp gene families shared between scaffolds containing nudivirus-like genes and/or proviral segments but without additional comparative data it remains unclear whether these genes are indicative of physical linkages that are conserved among microgastroid wasps generally . Future assemblies of the M . demolitor genome will eventually identify linkages between at least some of the MdBV proviral segment loci and nudivirus-like genes . However , based on the assembly used for this study , we conclude these linkages will not be in near proximity to each other . The finding that all BV proviral segments are flanked by WIM sequences together with evidence that nudivirus-like tyrosine recombinases recognize these motifs to produce circularized segments suggest both of these elements derive from the ancestral nudivirus genome [35] , [41] , [49] , [52] , [53] , [64] . Despite rapid evolution obscuring proviral segment homology [20]–[22] , the similarities in architecture of proviral segment loci between MdBV and GfBV , GiBV and CcBV also suggest shared ancestry . No data currently exist regarding motifs associated with integration of nudiviruses into the genomes of insects . On the other hand , if integration motifs with homology to WIMs or HIMs were identified from nudiviruses , it could provide important insights into the relationship between BV proviral segments and the nudivirus-like genes required for virion formation , proviral segment excision from the wasp genome , or segment integration into the genome of parasitized host insects . Duplication of genes into families is a recurring theme that serves as a key source of novelty in the molecular arms races that occur between parasites and hosts [51] , [67] . For MdBV and other BVs , previous studies show the virulence genes on proviral segments have diverse origins . Previously conducted evolutionary analyses show that some gene families on proviral segments , such as the sugar transporter genes identified from BVs associated with Glyptapanteles wasps and EGF gene family in MdBV are relatively recent acquisitions from wasps [41] , [58] . In contrast , other families show evidence of acquisition from other organisms or in the case of the PTP and Ank genes are of uncertain ancestry including possibly deriving from the nudivirus ancestor [54] , [55] , [68] . Prior studies also indicate that BV virulence gene family diversification has occurred through duplications plus rearrangements within and between segments , while also showing that some gene family members exhibit signatures of positive selection in response to arms race interactions with hosts [20]–[22] , [54] , [55] , [68]–[70] . In contrast , it currently is not possible to analyse how BV proviral segments have evolved among microgastroid braconids as a group because data are available for only five species in three genera including M . demolitor . Addressing this issue will require data from far more taxa . For M . demolitor , venom glands and teratocytes secrete large amounts of proteins that females introduce together with MdBV to parasitize hosts [58] . These products exhibit almost no overlap with MdBV genes , but notably many derive from gene families that have also diversified by duplication for selective expression in venom glands or teratocytes . In the current study , we find that some nudivirus-like replication genes have also duplicated more extensively than previously recognized with odv-e66 and 35a in particular being potentially significant in parasitism of hosts because products from both families are present in virions [49] . Variation in the size and contents of bacterial and eukaryotic genomes is thought to result from differences in effective population size and degree of genetic drift [7] , [11] . Genome evolution for vertically transmitted entities like BVs is largely unexplored , but is subject to different evolutionary processes compared to bacterial and eukaryotic symbionts , which retain their own cellular architecture and whose genomes are physically separated from that of their host . What we can conclude is that first , persistence as EVEs results in BV proviral genomes being inherited like other alleles in the wasp genome , which in turn subjects each BV to the effective population size of its associated wasp species . Second , on a broad scale BV proviral genomes show clear evidence of expansion relative to baculoviruses and nudiviruses . This is especially the case for proviral segments where decreases in gene density , increases in intron frequency , and gene acquisition from different sources followed by duplication are clearly apparent [40]–[44] . Interestingly , the direct repeat boundary regions recognized by viral integrases and tRNA loci , often associated with integration events , are features of proviral segments that are shared with pathogenicity islands in the genomes of disease-causing prokaryotes [71] . Pathogenicity islands initially evolve by horizontal gene transfer followed by site-specific recombination . Such processes could in part underlie the evolution of BV proviral segments as wasps adapt to parasitism of particular host species or guilds of closely related host species , and hosts reciprocally evolve to resist parasitism . Other studies have noted similarities between duplication of BV virulence genes and amplification of genes in insects associated with resistance to insecticides [47] . Many nudivirus-like genes in contrast show extreme dispersal throughout the M . demolitor genome but only a subset of these genes , also with potential roles in parasitism of hosts by wasps , have duplicated . Dispersal itself could occur through random processes of genome flux in wasps , similar to the rapid loss of microsynteny found by comparison of 12 species genomes in the genus Drosophila . Little is known currently about rates of genome flux in hymenopterans , although sufficient comparative genome data should eventually become available for BV-carrying species to perform a microsynteny analysis to see if proviral segments and nudivirus-like genes are similarly or more dispersed in the genome than wasp genes . Besides BVs , the Polydnaviridae currently contains a second genus , the Ichnovirus ( IVs ) , associated with parasitoid wasps in the family Ichneumonidae [20] , [21] . Recent evidence strongly suggests that IVs evolved independently of BVs from a still unknown virus ancestor ( s ) [72] . Nonetheless , while IV proviral segments encode largely different virulence genes , they exhibit many of the same organizational features as BV proviral segments , which suggests convergent evolution driven by the similar roles BVs and IVs play in parasitism [20] , [21] . Thus , the Polydnaviridae was originally recognized as a family because of the similarities in how the encapsidated form of BV and IV genomes are organized and their similar functions in parasitism of hosts by their associated wasps [20]–[22] , [33] . However , current data also now indicate this is a non-natural taxon that will be revised in the future . Most EVEs in animal genomes are non-functional fragments but a few cases are known of single viral genes or regulatory elements that have been exapted by hosts for new beneficial functions . Among these are mammalian syncytins derived from endogenized retrovirus env genes which function in placental development , and the Fv4 and Fv1 genes , also of retroviral origin , which function in antiviral defense [reviewed in 15] , [73] . Such EVEs are appropriately viewed as no longer being viruses because they no longer function as such [15] , [16] , [18] . BVs have also evolved to benefit wasps yet differ from the previous examples because their beneficial roles in parasitism depend on many genes whose functions remain the same as those of their ancestor [49] . BVs also retain much of the ancestral replication machinery and produce infectious particles that are quite similar to nudiviruses and baculoviruses [20] . BVs are thus ancient EVEs that benefit wasps but do so by continuing to function in many respects like a virus . As such , BVs also share features with other microbes that are viewed as obligate mutualists in which neither the symbiont nor host can survive without the other . BV genes required for replication are clearly of nudivirus origin . Genes on proviral segments on the other hand have a mixture of origins that include acquisition by horizontal transfer from wasps or other eukaryotes plus genes and motifs like WIMs that are ancient and exhibit features at least suggestive they too originated from the nudivirus ancestor [20]–[22] . Nudivirus and baculovirus genomes likewise consist of genes that are ‘viral’ in the sense they produce products required for replication , yet also contain genes and motifs of ancient origin and uncertain ancestry , plus genes acquired from arthropod hosts or other eukaryotes that function as virulence factors [38] . So what differs between BVs and their ancestors is not so much the types of genes they encode and their origins , but rather how their genomes are organized . Nudivirus and baculovirus genes , like those of most viruses , reside on a contiguous stretch of nucleic acid that is packaged into virions , whereas BV genes are organized in a manner that prevents all of them from being packaged into virions , which in turn also prevents BVs from existing independently of wasps . Other non-viral microbes that have evolved into vertically transmitted obligate mutualists do not persist by integrating into the genome of their host , but they too often exhibit profound alterations in genome organization and function that result in them no longer being genetically independent entities . The bacterium Buchnera in aphids , for example , was acquired in a single event that occurred more than 100 Mya , yet strict co-speciation thereafter has resulted in the phylogenies of these symbionts and aphids mirroring one another [74] . Thus , similar to discussions regarding whether entities like Buchnera are organelles or bacteria [9] , [75] , BVs will be viewed by some as EVEs that have been exapted by wasps to produce a novel organelle and others , including ourselves , as viruses that have evolved into wasp mutualists . Future studies will undoubtedly further advance our understanding of these fascinating associations .
All studies were approved by the Biological Safety and Animal Care and Use Committee of the University of Georgia and were performed in compliance with relevant institutional policies , National Institutes of Health regulations , Association for the Accreditation of Laboratory Animal care guidelines , and local , state , and federal laws . M . demolitor genomic DNA was isolated from single and pooled male wasps stored at −80 C . Briefly , 50 frozen males were pooled and ground in liquid nitrogen with mortar and pestle before lysing in a SDS solution overnight with Proteinase K . The homogenate was treated with RNaseA , and proteins/debris were collected after high-salt precipitation and centrifugation . After ethanol precipitation , the DNA was resuspended in 10 mM Tris and evaluated on an agarose gel and by Qubit quantification . The W . M . Keck Center for Comparative and Functional Genomics at the University of Illinois at Urbana-Champaign generated the following libraries for sequencing: a 180 bp-insert library from a single wasp , and 1 . 5 kb , 5 kb , and 10 kb-insert mate-pair libraries from pooled wasp DNA . The 180 bp and 1 . 5 kb-insert shotgun libraries were prepared with Illumina's TruSeq DNAseq Sample Prep Kit . The 5 kb and 10 kb mate-pair libraries were prepared similarly except a custom linker was ligated between the read-ends to facilitate mate-pair recovery . All libraries were sequenced for 100 cycles on a HiSeq2000 using the TruSeq SBS Sequencing Kit v . 3 . Data were analyzed with pipeline versions 1 . 8 and 1 . 9 . An additional 5 kb mate-pair library was constructed and sequenced by the Beijing Genome Institute using pooled wasp genomic DNA with all reads trimmed to 36 bp before assembly . The custom 5 kb and 10 kb mate-pair libraries were filtered for reads containing properly-oriented reads of the appropriate insert size and uniqueness using in-house custom pipeline scripts . Raw Illumina reads were 5′- and 3′-trimmed for nucleotide-bias and low-quality bases using the FASTX Toolkit ( http://hannonlab . cshl . edu/fastx_tookit/ ) . Trimmed reads were error-corrected by library with Quake [76] counting 19-mers . SOAPdenovo v2 . 04 [76] was employed with K = 49 to assemble the 180 bp-insert library reads followed by scaffolding with iteratively longer-insert mate-pair libraries and use of GapCloser v1 . 12 to close gaps generated in the scaffolding process with short paired read data [77] . MdBV DNA was isolated from virions as described previously [28] . The DNA pellet was resuspended in 10 µl of H2O and 1 ul was used as template in four phi29 amplification reactions as performed previously [78] . To resolve amplified DNA , the amplified product was incubated with S1 nuclease ( NEB ) at 37°C for 20 min . Precipitation and re-suspension in H2O yielded a total of 5 µg for Illumina sequencing . The sequencing library was prepared by the University of Georgia Genomics Facility using the Illumina TruSeq DNA sample preparation kit and the standard low-throughput protocol , and sequenced with the Illumina HiSeq system housed at the HudsonAlpha Institute for Biotechnology ( Huntsville , AL ) . Illumina sequenced reads from MdBV virions were filtered to retain read pairs with PHRED score equivalents >30 for >90% of nucleotides . Paired reads were mapped to the M . demolitor scaffolds using the bwa sampe algorithm and samtools [79] , [80] . Tablet was used to view the mapped reads relative to the reference genome , and scaffolds with clear read coverage boundaries indicating the presence of segments were selected [81] . These scaffolds had >280 , 000 mapped reads and were all >1 kb in size . The remaining scaffolds were identified by BLAST , indicating the presence of nudivirus-like genes . The longest translated ORFs from transcripts previously identified as nudivirus-like genes expressed in M . demolitor ovaries [28] were queried against the whole genome scaffolds and contigs BLAST database . BLAST results were manually filtered to retain real hits , which were filtered to have minimum 30% identity and at least 200 amino acids in alignment length with queries . These scaffolds were selected for annotation along with those identified by MdBV read mapping . Wasp transcriptome reads were generated and assembled previously from ovaries , teratocytes , venom glands , and wasp larvae [58] , and also from Pseudoplusia includens cells infected with MdBV . The wasp derived reads were assembled de novo using Trinity with the jaccard clip option , resulting in a total of 216 , 988 transcripts from 173 , 925 loci [82] . Reads were re-mapped to the assembled transcripts using bwa bwasw , successfully mapping 89–97% of reads for each tissue type [80] . The overall reads per kilobase per million reads mapped ( RPKM ) values were used to filter out low abundance transcripts ( <5 RPKM ) , in addition to the length of transcripts ( <500 bp removed ) [83] . The resulting filtered transcripts were used as evidence in gene model predictions described below . Forty eight million P . includens read pairs were generated and also used as described below . DNA was prepared from virus extracted from ovaries as above with DNAse treatment or with whole ovaries or animals without DNAse treatment . PCR was performed in a 10 µl reaction containing segment specific primers specific for amplifying across the WIM domain in a circularized segment ( 2 . 5 pmol ) ( Table S3 ) , 0 . 25 units of Hotmaster Taq polymerase ( 5 Prime ) and 1 µl of DNA as template . Reactions were run in a Bio-Rad thermocycler for 35 cycles with the following cycling conditions: initial denaturation at 94° for 2 min , followed by 35 cycles of denaturation at 94°C for 20 s , annealing for 20 s at 58°C , and extension at 65°C for 30 s with a final extension at 65°C for 7 min . Several forms of evidence were used as input into the MAKER annotation pipeline , including previous annotation of MdBV genes on proviral segments [43] , assembled transcripts from M . demolitor and P . includens ( see above ) [28] , [51] , unassembled read mapping information , and protein sequences from other insect species [84] . The assembled transcripts were generated from M . demolitor ovaries , whole larvae , teratocytes and venom gland transcriptomes as described above . These data sets generated 113 , 106 , 107 , and 147 million 100 bp paired reads respectively , which assembled into 36 , 891 transcripts with lengths greater than 500 bp and overall abundance greater than 5 Reads Per Kilobase of exon model per Million reads mapped ( RPKM ) . Of 51 million quality filtered Illumina read pairs from ovaries , 11% were successfully mapped to the 40 scaffolds containing MdBV proviral genome elements . Unassembled read mapping information was generated via mapping wasp ovary and C . includens hemocyte transcriptome reads against the scaffolds of interest using tophat and cufflinks [85] . The tophat read junctions file ( mapping splice sites ) and cufflinks transcripts file were converted into gff3 format using scripts bundled with MAKER , tophat2gff3 and cufflinks2gff3 . Protein sequences combined all coding sequences from the Nasonia vitripennis and Apis mellifera genomes , all known MdBV coding sequences , BV nudivirus-like coding sequences from M . demolitor , Cotesia congregata , and Chelonus inanitus , as well as all predicted ORFs from the genome scaffolds ( without introns ) >500 bp in size [27] , [28] , [36] , [63] , [86] . The de novo predictors GeneMark-ES , Augustus , and SNAP were also used within MAKER [87]–[89] . A GeneMark-ES model file was made using gm_es . pl and the scaffolds of interest as input . For Augustus prediction , a set of training genes was made by running Augustus on the scaffolds of interest with the species model “Nasonia” in addition to an intron position “hints” file generated with bam2hints from ovary transcriptome reads mapped by tophat . A “Microplitis” species file was made using the Augustus training gene set and the set of assembled transcripts described above with the autoAug . pl script . SNAP training was performed by iteratively running MAKER according to the “Training ab initio Gene Predictors” section of the MAKER tutorial , using assembled transcripts in the first iteration as evidence for gene models . The final MAKER run used the GeneMark-ES model file , the Augustus species file “Microplitis” and a SNAP model file generated by two rounds of training . Repeat masking was performed with default options . The resulting MAKER gff3 files were loaded into Apollo genome browser and gene models were manually edited if necessary [90] . All gene models were exported into Genbank table format . The coding sequences were assigned putative functional roles based upon BLAST results from the NCBI nr database , the Drosophila melanogaster set of protein-coding genes , and HMMER hmmsearch results from the PFAM and TIGRFAM databases [91] , [92] . tRNAs were identified using Aragorn [93] . These results were combined into GenBank format using custom perl scripts and tbl2asn from NCBI . The resulting dataset was submitted as BioProject 195937 and assigned GenBank accession number AZMT000000000 . Altogether , we identified 1 , 737 genes in the 40 scaffolds containing MdBV components of which 1 , 713 were predicted protein-coding sequences and 24 were tRNAs ( Table S1 ) . Wasp Integration Motifs ( WIMs ) share the common sequence AGCT and identify the site at which MdBV proviral segments excise from the M . demolitor genome [35] . WIMs were previously mapped for three segments by inverse PCR [35] . The locations of WIM sites for all proviral segments in the M . demolitor genome were obvious from read mapping data , and segment beginning and end coordinates were given to AGCT sites where circularization occurs . Host Integration Motifs ( HIMs ) were also identified previously by PCR methods and sequencing as the site where MdBV DNAs in virions integrate into the genome of infected host cells [35] . The twelve HIM sequences identified by these approaches were aligned and made into a Hidden Markov Model ( hmm ) using HMMER hmmbuild ( http://hmmer . janelia . org ) . This hmm was used as a query in a HMMER hmmsearch against the scaffolds of interest to identify HIM motifs in all correctly assembled old and new segments . Sequence motifs were aligned using MUSCLE and maximum likelihood phylogenetic trees were built using phylogeny . fr with the HKY+I+G model and 100 bootstrap replicates [94] . Sets of orthologous genes were identified using orthomcl and four datasets: 1 ) all protein-coding sequences from M . demolitor , 2 ) all protein sequences from G . flavicoxis BV segments and flanking regions of the wasp genome , 3 ) same information from 2 for G . indiensis , and 4 ) protein coding sequences from nudivirus-gene containing regions of C . congregata [27] , [41] , [95] . Groups of orthologous genes ( syntenic ) in the genome were identified using the orthomcl output with orthocluster [96] . Syntenic regions were viewed with Gbrowse syn . Assembly of the M . demolitor genome initially suggested the nudivirus-like gene cluster contained two identical duplications absent from the nudivirus-like gene cluster identified in the wasp C . congregata [27]: Cc50C22 . 3/HzNVorf94-like/38K and 27b-like/Cc50C22 . 6 . Given the identical nature of these predicted duplications , we assessed whether they were correct by primer walking and Sanger sequencing these domains . Resequencing showed that these genes were not duplicated . The above information was then used to make figures that included modifications of pictorial representations of M . demolitor genes on scaffolds generated by Gbrowse [97] , [98] . | Microorganisms form obligate associations with multicellular organisms that range from antagonistic ( parasitic ) to beneficial ( mutualists ) . Although numerous examples of obligate , mutualistic bacteria , fungi , and protozoans exist , viruses are thought to usually form parasitic associations . An exception is the family Polydnaviridae , which consists of large DNA viruses that have evolved into mutualists of insects called parasitoid wasps . Each wasp species relies on its associated polydnavirus to parasitize hosts while each polydnavirus relies on its wasp for transmission . Polydnaviruses in the genus Bracovirus evolved approximately 100 million years ago from a group of viruses called nudiviruses , which are closely related to another large family of viruses called baculoviruses that are virulent pathogens of insects . Bracoviruses are of interest , therefore , because they provide a study system for examining how evolution into mutualists affects the structure and function of viral genomes . In this study , we sequenced the genome of the wasp Microplitis demolitor to characterize the proviral genome of M . demolitor bracovirus ( MdBV ) . While the viral ancestor of bracoviruses possessed an independent circular genome , the proviral genome of MdBV is widely dispersed in the genome of M . demolitor . Our results also provide new insights into how the MdBV genome functions to produce virus particles that wasps rely upon to parasitize host insects . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"evolutionary",
"biology",
"species",
"interactions",
"biology",
"and",
"life",
"sciences",
"microbiology"
] | 2014 | Widespread Genome Reorganization of an Obligate Virus Mutualist |
Terpenoid synthases create diverse carbon skeletons by catalyzing complex carbocation rearrangements , making them particularly challenging for enzyme function prediction . To begin to address this challenge , we have developed a computational approach for the systematic enumeration of terpenoid carbocations . Application of this approach allows us to systematically define a nearly complete chemical space for the potential carbon skeletons of products from monoterpenoid synthases . Specifically , 18758 carbocations were generated , which we cluster into 74 cyclic skeletons . Five of the 74 skeletons are found in known natural products; some of the others are plausible for new functions , either in nature or engineered . This work systematizes the description of function for this class of enzymes , and provides a basis for predicting functions of uncharacterized enzymes . To our knowledge , this is the first computational study to explore the complete product chemical space of this important class of enzymes .
Terpenoids , which have diverse carbon skeletons , are an important class of natural products [1–3] . To date , more than 63 , 000 different terpenoids have been reported [4] . In nature , most cyclic terpenoids are created by terpenoid synthases ( sometimes called terpenoid cyclases [5] ) , which catalyze the cyclizations of linear terpenes such as geranyl diphosphate through carbocation rearrangements [6] . The cyclized carbocationic intermediates are ultimately quenched by phosphorylation , deprotonation , or hydration to yield products ( Fig 1 ) . The intrinsic reactivity of carbocations plays an important role in the outcome of cyclization [7–9] . Terpenoids are classified as monoterpenes ( C10 ) , sesquiterpenes ( C15 ) , diterpenes ( C20 ) , sesterterpenes ( C25 ) , triterpenes ( C30 ) and sesquarterpenes ( C35 ) according to the number of C5 isoprenoid units incorporated into their carbon skeletons . Rapid advances in DNA sequencing provide an opportunity to discover enzymes involved in creating both previously characterized and novel terpenoid natural products . The gap between sequenced genes and reliable functional annotations is enormous and increasing . For example , the Structure-Function Linkage Database ( version 2014 ) [10] assigns 2778 enzyme sequences to the terpene synthase subgroup of the isoprenoid synthase 2 superfamily ( Mg-dependent ) , of which 2540 ( 91% ) are annotated as having ‘unknown’ function . Thus , the functions of the large majority of these enzymes remain uncharacterized . Inferring enzyme function from protein sequence is challenging in general [11] , and is likely to be particularly difficult for enzymes involved in terpenoid biosynthesis , because 1 ) the potential product chemical space is huge , and 2 ) single point mutations can alter product specificity [12] . In previous work , we have predicted enzyme substrates and products from protein sequence by using a combination of bioinformatics and structural modeling [13 , 14] . In order to apply similar methods to terpene synthases , a first major challenge is simply to enumerate the possible enzyme activities that could exist among the uncharacterized enzymes . Defining the possible substrates is trivial ( C5 , C10 , C15 , etc . ) , although there have been investigations into the catalytic mechanisms of a few terpene synthases [6] , no previous attempts have been made to systematically define the possible products , due to the complexity of the problem . In this work , we systematically enumerate thousands of potential monoterpenoid carbocationic intermediates , by using computer simulations . To present the complex results in a simple manner , we organize the carbocationic intermediates according to their cyclic ring structures and the locations of double bonds within the carbocycles . We identify 74 such cyclic product skeletons , among which ( at least ) 5 are represented among characterized monoterpenoid natural products . Among the remaining skeletons , several appear to be plausible albeit hypothetical monoterpene skeletons , in the sense that they can be connected to the linear substrate by a relatively small number of carbocation rearrangements known to occur in terpene synthases . Thus , although natural products with these skeletons do not appear to have been reported , they may be found among the products of the many currently uncharacterized terpene synthases , or be accessible via enzyme engineering .
Our simulations perform virtual carbocation rearrangements in the gas phase ( Figs 2 and 3 and S1 Movie ) , allowing the enumeration of all carbocations that follow from cyclization of the linear allylic monoterpene carbocation . Five reaction types are considered ( Fig 2b ) : 1 ) intramolecular alkylation of double bonds; 2 ) alkyl shifts ( excluding 1 , 2-methyl shifts ) ; 3 ) hydride shifts; 4 ) 1 , 2-methyl shifts; 5 ) proton transfers . All five types of reactions were carried out for each carbocationic intermediate ( details see Methods ) . The energies of product carbocations were evaluated by semi-empirical quantum mechanics to ensure their thermo-stability at room temperature ( 0 kcal/mol relative energy filter , see Methods ) . The ‘Simplified Molecular Input Line Entry System’ ( SMILES ) , which describes the chemical structures using ASCII strings ( Fig 2a; details see Methods ) , is used to eliminate duplicate product carbocations . The output of our simulation is a carbocationic reaction network , where nodes are intermediates and edges are reactions ( Fig 3; it should be noted that not all of the intermediates and edges are shown , for simplicity ) . To validate our code , we designed an alkane carbocation enumeration experiment for C5-C10 , where linear alkane carbocations are used as the reactants ( details see Methods ) . We expect that the output will contain all alkane carbocation isomers . We then manually drew all the carbocationic isomers for C5-C10 and compared with the output of our code . As expected , consistent results are obtained ( S2 Table ) . The total number of monoterpene carbocations obtained by our simulation is 18758 , connected by 123093 virtual reactions ( the number of edges ) . To organize the chemical space of carbocations in a simple manner , we define skeletons for the neutralized carbocation with the saturated alkyl side chains removed ( Fig 1 ) . When we group carbocations in this way , 74 cyclized skeletons are found . These cyclized skeletons can be divided into five groups: 1 ) one ring plus one double bond; 2 ) two rings containing bridged carbons; 3 ) two fused rings; 4 ) two rings linked by a spiro carbon; and 5 ) two separated rings ( Fig 4 ) . To date , only five monoterpene skeletons are associated with EC numbers ( by IUBMB; see red skeletons in Fig 4 and S3 Table ) , all of which can be found among the 74 skeletons found by our automated approach . Interestingly , none of the known skeletons belong to the groups that have two rings joined at a spiro carbon or two separated rings . More broadly , although we cannot claim to have performed an exhaustive search , we have not identified any known natural products for 69 of the skeletons . Do the 5 skeletons with EC numbers have any features that distinguish them from the 69 unobserved skeletons ? Are any of these alternative skeletons plausible , in terms of representing backbone structures that might in the future be identified among monoterpene natural products , among the many that undoubtedly remain unidentified at present; or that might be accessible by enzyme engineering ? The stability of carbocations is an important consideration . For example , secondary carbocations are avoided in most of the terpene synthase reactions . To begin to address this issue , albeit in a somewhat simplistic manner , we applied more stringent energy filters in an attempt to eliminate less stable carbocations . As desired , the fraction of secondary carbocations decreased as we made the energy cutoff more stringent ( S1 Fig and S1 Table ) . Specifically , with the original 0 kcal/mol energy cutoff ( energies are relative to the geranyl carbocation , in kcal/mol ) , 48% are secondary carbocations . With -5 and -10 kcal/mol energy cutoffs , the fraction of secondary carbocations decrease to 33% and 16% , respectively . When applying these two more stringent energy cutoffs , the number of cyclic skeletons identified decreased from 74 to 38 and 35 cyclic skeletons , respectively ( S2 Fig ) . Notably , no skeletons containing two separated rings were found , probably because they are unstable . Fig 5 maps the skeletons onto two variables , specifically the logarithm of the number of carbocations associated with each skeleton [log ( ncarbocation ) ] , versus the number of reaction steps in the shortest route to obtaining the skeleton from the linear reactant . The number of carbocations associated with a skeleton is largely related to the number of possible substitution patterns and stereoisomers associated with each skeleton . This number is also strongly correlated with the number of reaction steps . The product skeletons associated with known EC numbers ( in red ) are located primarily in the top-left corner of the plot . Monoterpene skeletons that can only be accessed through a large number of transformations ( 5 or greater ) do not appear to be represented among known natural products , although more than 5 rearrangement steps are required for the product formation of some sesquiterpenoid synthases , e . g . epi-isozizaene synthase . Seven skeletons are accessible in "step 4" of Fig 5 , the step immediately following the first cyclization step . Of these , 3 have associated EC numbers; the remaining 4 skeletons would seem to be excellent candidates for currently uncharacterized monoterpenoid natural products or for enzyme engineering , although we cannot of course prove this . It should be noted that some of the skeletons may not be accessible because high-energy intermediates and transition states are involved , e . g . the methylenecycloheptane skeleton at “step 4” ( it is not found in the simulation with -5 kcal/mol energy cutoff ) . To explore whether the predicted skeletons are stable compounds , we manually searched the chemical database PubChem [15] ( S4 Table ) . All the skeletons are found , implying that all these predicted skeletons are stable . The top 30 most populated skeletons are shown in S5 Table . To visualize the complicated carbocation reaction network , we developed a web application called ‘Search C+’ ( available at http://carbocation . jacobsonlab . org:8080/; an example query can be found in S4 Fig ) . Users can search the carbocation virtual library based on chemical similarity [16] . Once a monoterpene carbocation is found , potential reaction routes can be automatically displayed . Users can also identify the neighboring carbocations of a query carbocation in a local network view ( the complete network is too large to display ) . To predict potential reaction routes for monoterpene carbocations , we performed graph traversal on the obtained carbocation reaction network . Most carbocations can be accessed via multiple reaction routes , and we keep only the shortest route for each precursor carbocation . To predict the best route , one must obtain accurate reaction energies by performing QM/MM or QM cluster calculations in the presence of enzyme [17 , 18] , which is beyond the scope of the current work . Recently , Lobb generated ~1000 C7 carbocation intermediates and transition states by searching reaction types similar to this work , followed by geometry optimizations with DFT methods [19] . A similar approach , including explicitly identifying and optimizing transition states , would be valuable for the terpenoid carbocation intermediates considered here , but the computational cost would be rather high at the present time . Although previous theoretical studies have provided insights into the reaction mechanism for a number of known mono- , sesqui- and diterpenes [6] , this is the first computational study to systematically explore the complete chemical space of monoterpenoid carbocations . It should be noted that non-classical carbocations are not considered in our algorithm and only one conformer is retained for each carbocation .
As a critical first step towards enzymatic activity prediction for terpenoid synthases , we have created a computational algorithm to systematically enumerate plausible carbocationic intermediates and the product carbon skeletons that can formed from them . For monoterpenoid synthases ( C10 ) , we have run many iterations of the algorithm to identify intermediates and product skeletons that can result from enzymatic transformations proceeding through multiple intermediates . The results encompass all monoterpene synthase activities described by EC numbers , as well as other plausible product skeletons that we speculate could be created by one of the many uncharacterized putative terpene synthase enzymes or by engineered enzymes . It may be possible to systematically explore the chemical space of sesquiterpene cyclases ( C15 ) in an analogous manner , although clearly this will be challenging . Recently , a semi-automatic algorithm has been applied to the generation of sesquiterpene carbocations from the humulyl cation ( the 1 , 11-cyclized intermediate ) [20] . However , the computational cost of such an algorithm is high , the output of the algorithm seems to consist of less than 200 carbocations , and some of the known carbocations are not explicitly located [20] . Other algorithms [21] without using quantum mechanics may enumerate highly unstable carbocations . In our on-going work to apply the methods described here to sesquiterpene carbocations , we have already enumerated millions of possible product-precursor structures . Although the methods described here are computationally efficient , the exponential increase in the number of possible carbocations with chain length makes it unlikely that we can perform such a systematic exploration of diterpenoid or larger carbocations . In a previous study [22] , the graph-based enumeration of organic small molecules containing C , N , O , S , and halogens was performed for up to 17 heavy atoms , and 166 billion molecules were obtained ( without considering stereochemistry ) . However , an alternative approach , appropriate for product prediction of terpene cyclases with crystallographic structures ( or sufficiently accurate homology models ) , is to adapt iGen to create carbocations in the active site of an enzyme . The advantage is that one can eliminate "on the fly" those carbocations that do not fit in the site or are electrostatically incompatible , thus reducing the combinatorial explosion . Thus , in principle , the automatic enumeration algorithm may allow the prediction of novel terpenoid skeletons , which was previously impossible [13 , 14] . As a first proof-of-concept , we have recently used such an approach to facilitate discovery of a novel sesquiterpene synthase [23] .
The iGen algorithm for systematically enumerating carbocations is illustrated in Fig 2a . The reactant carbocation intermediates ( input structures ) undergo carbocation rearrangements according to a set of predefined reaction types ( Fig 2b; resonance structures are also generated ) . The input structures can be any carbocations . In the simulations for the monoterpene carbocations , we initiate the calculations with three cyclic carbocation intermediates , i . e . , two 1 , 6-cyclized intermediates , differing in stereochemistry , and a 1 , 7-cyclized intermediate ( Fig 3 shows an example starting from one of the 1 , 6-cyclized intermediates ) . The first two reaction steps , i . e . trans/cis isomerization of the linear carbocation and the cyclization of the cis linear carbocation , are not shown in Fig 3 for simplicity . We use two key iterations to generate all possible products for a given reactant carbocation ( S5 Fig ) : 1 ) iterations on atoms of the reactant; 2 ) iterations on reaction types . Atoms of the reactant carbocation are placed in a reactive atom list , except for the carbocation atom and its three bonded atoms . For each atom in the reactive atom list ( iterations on atoms ) , iGen checks whether this atom fits the features for any of the predefined reaction types; for example , if the reactive atom is a carbon atom in a double bond , it fits the reaction type 1 ( e . g . , iteration 13 in S5 Fig ) . Virtual reactions are performed by changing the connectivity of the reactant carbocation . The structure-class of the Schrӧdinger software [24] , which has built-in functions such as “addBond” , “deleteBond” and “setFormalCharge” , is used to facilitate the molecular connectivity operations . The resulting carbocations are energy-minimized using molecular mechanics ( MM ) and quantum mechanics ( QM ) calculations . The role of the MM minimization is to obtain reasonable geometries of the products after changing the molecular connectivity ( S5 Fig ) . Further semi-empirical QM minimizations , using the RM1 semi-empirical method of the MOPAC package [25] , are used to eliminate high-energy carbocations ( specific cutoffs described below ) . Duplicate carbocations are identified and eliminated by using Simplified Molecular Input Line Entry System ( SMILES strings ) , which describes chemical structures using ASCII strings . The obtained product carbocations then become reactant carbocations in the next round . This process runs repeatedly until no new carbocations can be generated , or other user-defined criteria such as the maximum round number are reached . The QM energy cutoff is set to 0 . 0 kcal/mol ( relative to the linear reactant GPP cation ) . For long-range hydride-shift and proton transfer reactions , a C-H distance-cutoff 5 . 0 Å is used for these two reaction types after Round 5 ( long-range hydride shift and proton transfer sometimes occur in enzymatic reactions , mediated by active site residues or water ) . However , such reactions normally only occur in the first few steps , e . g . , 5-epi-aristolochene synthase [26] and selina-4 ( 15 ) , 7 ( 11 ) -diene synthase [27] . iGen is written in Python and takes advantage of the Python API of the Schrӧdinger software [24] , which has many built-in functions such as a SMILES string calculator and MM minimizer . For carbocations generated in the first five rounds , conformational sampling is performed by using a Monte Carlo sampling approach implemented in the MacroModel software [24] . Each conformer undergoes virtual reactions as described above . We did not perform full conformational sampling for all the carbocations , as this significantly increases computational costs . We expect that generating more conformers may lead to larger numbers of stereo-isomers among the products but not necessarily more product skeletons . To improve chemical space sampling , we added a ‘stereochemistry module’ , which enables the generation of more stereoisomers for a given carbocation conformer . For example , for the conformer described in iteration 1 of S5 Fig , where the H1-C2-C3-C4 dihedral angle is close to zero degrees , it is not clear which stereoisomer is more favorable . In such cases , the ‘stereochemistry module’ generates both stereoisomers via Cartesian coordinate operations . We first calculate the transformation vectors: 1 ) two orthogonal vectors ( with opposite signs ) of the plane defined by the sp2 cation atom are calculated; 2 ) the final position ( Cartesian coordinates ) of the reactive atom is determined by the orthogonal vector multiplied by a default bond length; 3 ) the transformation vector is the difference between the coordinates of the final position and the current position of the reactive atom . If the reactive atom is carbon ( reaction types 1 , 2 and 4 ) , the coordinates of the atoms bonded to this reactive atom will also be changed via the same vectors as the reactive atom . In this work , the dihedral angle range to invoke the ‘stereochemistry module’ is set to be [-45°~+45°] . We performed a validation test by enumerating all possible C5-C10 alkane carbocation isomers . By running iGen with reaction types 2–4 ( alkyl shift , hydride shift and methyl shift ) on a linear alkane carbocation , all the isomers of that alkane carbocation will be generated . It should be noted that reaction types 1 and 5 do not apply to alkane carbocations . We then manually drew all possible C5-C10 alkane carbocations , and compared with the iGen output ( S2 Table ) . QM calculations are not performed in these tests , because many of the alkane carbocations containing -CH2+ are unstable in the QM calculations . | Terpenoids , as one of the largest classes of natural products , provide complex carbocycle structures for many drugs ( e . g . taxol ) and prodrugs . The diverse carbocycle structures arise from complex carbocation rearrangements catalyzed by terpenoid synthases . Many putative terpene synthase enzymes identified in genome sequencing efforts remain functionally uncharacterized , and some of these will undoubtedly have novel products , potentially including previously undiscovered carbocycles . In this work , we present a computational approach that systematically enumerates all plausible carbocations of monoterpenoid synthases in order to define and organize the potentially large product chemical space of this important class of enzymes . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"sequencing",
"techniques",
"terpenes",
"stereoisomers",
"chemical",
"compounds",
"enzymology",
"reactants",
"alkanes",
"organic",
"compounds",
"isomerism",
"molecular",
"biology",
"techniques",
"research",
"and",
"analysis",
"methods",
"hydrocarbons",
"sequence",
"analysis... | 2016 | Defining the Product Chemical Space of Monoterpenoid Synthases |
Togo has conducted annual , integrated , community-based mass drug administration ( MDA ) for soil-transmitted helminths ( STH ) and schistosomiasis since 2010 . Treatment frequency and target populations are determined by disease prevalence , as measured by baseline surveys in 2007 and 2009 , and WHO guidelines . Reported programmatic treatment coverage has averaged over 94% . Togo conducted a cross-sectional survey in 2015 to assess the impact of four to five years of MDA on these diseases . In every sub-district in the country outside the capital , the same schools were visited as at baseline and a sample of fifteen children age 6 to 9 years old was drawn . Each child submitted urine and a stool sample . Urine samples were tested by dipstick for the presence of blood as a proxy measure of Schistosoma haematobium infection . Stool samples were analyzed by the Kato-Katz method for STH and Schistosoma mansoni . At baseline , 17 , 100 children were enrolled at 1 , 129 schools in 562 sub-districts; in 2015 , 16 , 890 children were enrolled at the same schools . The overall prevalence of both STH and schistosomiasis declined significantly , from 31 . 5% to 11 . 6% for STH and from 23 . 5% to 5 . 0% for schistosomiasis ( p<0 . 001 in both instances ) . Egg counts from both years were available only for hookworm and S . mansoni; intensity of infection decreased significantly for both infections from 2009 to 2015 ( p<0 . 001 for both infections ) . In areas with high baseline prevalence , rebound of hookworm infection was noted in children who had not received albendazole in the past 6 months . After four to five years of MDA in Togo , the prevalence and intensity of STH and schistosomiasis infection were significantly reduced compared to baseline . Data on STH indicate that stopping MDA in areas with high baseline prevalence may result in significant rebound of infection . Togo’s findings may help refine treatment recommendations for these diseases .
Schistosomiasis and soil-transmitted helminths ( STH ) are parasitic diseases that cause significant morbidity worldwide , particularly in sub-Saharan Africa . STH infections are among the most prevalent infections in the world , with approximately 1 . 5 billion people infected worldwide [1 , 2] . These infections can cause weakness , malaise , anemia , and impaired physical and cognitive development [1] . Schistosomiasis can also cause hematuria , anemia , stunting , and reduced ability to learn , and in its chronic form may lead to urogenital complications with bleeding , fibrosis , kidney damage , and bladder cancer [3] . More than 200 million people worldwide are infected with schistosomiasis , and more than 85% of affected people live in sub-Saharan Africa [3] . A key public health strategy against these infections is morbidity control through mass administration of preventive chemotherapy ( PC ) for children and high-risk adults . Distribution of PC through mass drug administration ( MDA ) is a highly cost-effective public health program and integration of MDA for multiple diseases in a single campaign yields additional cost savings [4 , 5] . Through the coordinated efforts of partners worldwide , including a commitment by major pharmaceutical companies to donate the needed albendazole , mebendazole , and praziquantel , MDA has been dramatically scaled up worldwide in recent years [6] . In 2015 , more than 66 million individuals received PC for schistosomiasis , and more than 711 million for STH [7] . While Togo has adhered closely to the WHO guidelines for management of these diseases , the optimal strategies for PC are continually being researched and effectiveness may depend on local disease epidemiology , community acceptance of MDA and other factors . It is therefore incumbent upon countries to monitor the impact of their programs , both to refine control strategies and to share successes and challenges with the global audience to advance knowledge about control of these diseases . The WHO recommends that countries conduct an evaluation after five years of MDA to assess impact on the prevalence of STH and schistosomiasis , and adjust the treatment distribution plan based on the new prevalence data , if appropriate [8] . Here we report on an evaluation assessing the impact of four ( in the south of Togo ) to five ( in the north ) years of MDA on the prevalence and intensity of STH and schistosomiasis infection in school-age children in Togo .
The country of Togo has a population of more than 7 million people , all of whom are at risk for STH and schistosomiasis infection [9 , 10] . In 2009 , the Togo Ministry of Health ( MOH ) launched its Integrated Program for Neglected Tropical Disease ( NTD ) Control , initiating integrated MDA for the three PC-targeted NTDs currently endemic to Togo: schistosomiasis , STH , and onchocerciasis . Baseline disease prevalence was determined through two surveys , a pilot survey of integrated disease mapping in Binah district in 2007 and a subsequent national survey of the rest of the country in 2009 ( henceforth referred to together as “Baseline” , or simply “2009” , data ) . Together , the two integrated baseline surveys measured the prevalence of schistosomiasis and STH in school-age children in every sub-district in all 35 districts outside the capital of Lomé . Since 2010 , Togo has implemented integrated , community-based ( door-to-door ) MDA for schistosomiasis and STH , according to World Health Organization ( WHO ) guidelines [8] . School-age children ( SAC ) are treated with albendazole for STH either annually or twice per year , and with praziquantel for schistosomiasis either annually or every other year , based on the district and sub-district prevalence , respectively , of the two diseases [10] . In sub-districts with schistosomiasis prevalence ≥50% , praziquantel is also given to adults . UNICEF conducts deworming of children age 1–4 years at child health days in Togo . To compare prevalence and intensity of infection at baseline and follow-up , in 2015 we repeated the cross-sectional survey that was employed for the baseline prevalence mapping ( conducted in 2007 for Binah district and in 2009 for all other districts outside the capital ) [10] . The baseline sampling strategy was a novel approach developed by the Togo MOH and the US Centers for Disease Control and Prevention and was designed to capture the focal nature of schistosomiasis by enrolling smaller numbers of children at more sites than is proposed by WHO . We visited all 632 sub-districts of the 35 districts outside the capital . In Binah district in 2007 , three villages with expected high prevalence of schistosomiasis were selected in each sub-district , based on reports of hematuria from village health centers or proximity to water . In all other districts , two such villages were sampled per sub-district . In the absence of any risk factors for schistosomiasis the villages were chosen at random . One government-run primary school or government-assisted denominational school was sampled in each village and fifteen children were enrolled at each school . The methodology and results of these integrated baseline assessments have been previously published [10] . In 2015 we visited the same villages as at baseline and in each village , whenever possible , we sampled children from the same primary school that was surveyed at baseline . For schools that had closed or could not be located , the nearest public school was selected as a replacement . As during the baseline survey , the day prior to the arrival of the field team the sub-district nurse visited the school and instructed the headmaster to select a sample of 30 children aged 6 to 9 years old from Cours Elémentaire classes I and II ( equivalent to first and second grades in the USA ) . This age group was selected to comply with the WHO recommendation of selecting children in their third year of school , while providing an additional grade to ensure sufficient sample size if class sizes were small [8] . Consent forms were sent home with children who had verbally assented to participate in the survey . The following day , children who presented written parental consent and who could provide both urine and a stool sample were enrolled until 15 children from the school had been recruited . Each school headmaster was asked if there had been any school-based deworming activities in the past twelve months . Global positioning system ( GPS ) coordinates for each school were recorded . Immediately upon collection in the field , each urine sample was visually assessed , and results were recorded as clear , turbulent or bloody . The samples were tested immediately at the school by urine dipstick ( 1-parameter urine reagent strips; LW Scientific , Lawrenceville , GA ) for the presence of blood , as a proxy measure for Schistosoma haematobium infection , as was done at baseline [11] . The semi-quantitative result was recorded as negative , trace , small , moderate , or large amount of blood in the urine , according to the urine dipstick gradations . Any urine dipstick result , other than “negative” , was considered a positive result for S . haematobium . After collection of the field samples , the stool and urine samples were transported to the nearest health center where the field team established a mobile laboratory . All stool samples were analyzed by the Kato-Katz method; one slide per child ( as in 2009 ) was prepared and read by a laboratory technician using standard procedures [11] , and number of eggs per gram of stool was calculated for Schistosoma mansoni , Ascaris lumbricoides , Trichuris trichiura , and hookworm . The time from enrollment of the first child in the field to the final Kato-Katz analysis in the mobile laboratory was recorded . In the first school visited in each sub-district ( i . e . half the schools ) , one 10-mL volume of urine from each of the first five enrolled children was filtered through a 13mm diameter , 12-micron polycarbonate filter ( Sterlitech Corporation , Kent , WA , USA ) and examined by microscopy , and S . haematobium egg counts per 10 mL of urine were recorded . Data were double entered into Microsoft Access ( Microsoft Corporation , Bellevue , WA ) , cleaned , and analyzed in STATA 13 . 1 ( StataCorp , College Station , TX ) . Prevalence estimates between groups ( year , sex ) were compared using the chi-squared test and prevalence trends across age groups and time-to-reading of Kato-Katz slides were examined using the Cuzick non-parametric test for trend . Disease prevalence across groups was compared using the t-test for unpaired samples with equal or unequal variance ( as appropriate ) . Mixed effect logistic regression models were developed to examine factors associated with infection . The significance threshold for all statistical tests was set at 0 . 05 . Missing data were excluded from estimates of prevalence or intensity of infection . Maps were created using ArcMap 10 . 4 . 1 ( Esri , Redlands , CA ) . The protocol for this study was approved by Togo’s Bioethics Committee for Health Research ( Comité Bioéthique de Recherche en Santé; approval number 029/2014/CBRS du 04 décembre 2014 ) . Consent forms were sent home with children who had verbally assented to participate in the survey . The following day , a subset of children who presented written parental consent forms were enrolled in the study .
The results of the 2009 baseline survey have been previously reported [10] . In short , baseline data were collected from 17 , 100 children at 1129 schools , from October 28 to December 6 , 2009 , except for Binah district baseline data , which were collected from March 15 to 28 , 2007 . Only school-level summary data were available for certain measures in Binah . For the 2015 impact assessment reported on here , from February 15 to March 31 , 2015 , 16 , 890 children were enrolled from 1 , 126 schools in 562 sub-districts across Togo ( Table 1 ) . Of the 1 , 126 schools included in the impact assessment , 1 , 097 were the identical school surveyed in 2009 or 2007 and 26 were replacements for schools that no longer existed or could not be found . Three schools from the 2009 survey were no longer operating and could not be replaced because there was no school in the vicinity . Table 1 shows the age and sex of children surveyed at baseline and in 2015 . For the impact assessment , stool samples could be analyzed for 16 , 889 children and urine dipstick results were available for 16 , 783 children . The median time from enrollment of the first child in the field to examination of the final Kato-Katz slide in the mobile laboratory was 4 hours 10 minutes ( mean 4 hours 39 minutes ) . The time from collection to processing was not recorded . We did not observe a decrease in the prevalence of STH infection as time from stool collection to reading of the Kato-Katz slide increased . Fig 1 shows treatment coverage as reported by community drug distributors ( number of individuals receiving the medication divided by the number of individuals targeted to receive treatment ) by medication and year . No village had poor coverage for more than two consecutive years . We examined 2014 treatment coverage at the village level for those villages in which we had school-based STH and schistosomiasis prevalence estimates and found no correlation between albendazole or praziquantel treatment coverage in 2014 and STH or schistosomiasis prevalence , respectively , in 2015 . From 2009 to 2015 , the overall prevalence of STH decreased significantly from 31 . 5% in 2009 to 11 . 6% in 2015 ( p<0 . 001; Table 2 ) . Hookworm was the predominant STH infection at baseline and at follow-up; prevalence of hookworm decreased significantly from 31 . 0% to 11 . 1% ( p<0 . 001 ) . Among those with hookworm infection , the egg count per gram of stool was significantly lower in 2015 compared to baseline ( p<0 . 001; Table 2 ) . There were insufficient numbers of Ascaris and Trichuris infections to examine changes in intensity of infection . At baseline and in 2015 , the percent of children testing positive for STH ranged from 0% to 100% across all schools . At baseline , 33 children had mixed infections; in 2015 , 29 children had mixed infections . The prevalence of schistosomiasis decreased significantly from 23 . 5% in 2009 to 5 . 0% in 2015 ( p<0 . 001; Table 3 ) . The prevalence of microhematuria , which served as a proxy measure for S . haematobium , declined from 21 . 0% at baseline to 4 . 2% in 2015 ( p<0 . 001 ) . The prevalence of S . mansoni declined from 3 . 6% to 0 . 8% ( p<0 . 001 ) . There were 154 children with mixed infections at baseline and four children with mixed infections at follow-up , but individual-level data were not available for Binah district at baseline so the prevalence of mixed infections could not be assessed for Binah . Among those who were infected , the mean egg count was significantly lower for S . mansoni in 2015 compared to 2009 ( p<0 . 001 , Table 3 ) . In 2015 the majority of S . haematobium infections were light ( Table 3 ) ; there was no quantitative measurement of the intensity of S . haematobium infection in 2009 . Figs 2–4 show the prevalence of STH , S . haematobium and S . mansoni infection , respectively , at each school in 2009 and 2015 ( 15 children surveyed per school ) . Significant reductions in the prevalence of all three diseases are evident in all areas of the country . Figs 5–7 show locations of schools where any child had a high or moderate intensity infection with STH , S . haematobium or S . mansoni , respectively . For S . mansoni , children at schools with higher prevalence of infection were more likely to have high or moderate intensity infections . For STH , there is a significant trend toward higher prevalence of hookworm at older ages for both boys and girls , and a trend to higher mean egg counts at older ages for boys , in both 2009 and 2015 ( p<0 . 001 in each instance; Table 4 ) . For schistosomiasis , there is a significant trend toward higher S . mansoni prevalence and mean egg counts with increasing age for both boys and girls in 2015 ( p<0 . 001 in both instances ) but not in 2009 . For S . haematobium there is a trend toward higher prevalence of infection at older ages for both boys and girls in 2009 and in 2015 ( p<0 . 001 in all instances ) but no trend in intensity of infection in 2015; no data on intensity of infection are available for S . haematobium in 2009 . Boys were more likely to be infected with hookworm and had higher mean egg counts than were girls , in both 2009 and 2015 ( p<0 . 001 in each instance; Table 4 ) , but sex was not related to Ascaris or Trichuris infection . In 2009 , the prevalence of S . mansoni and S . haematobium infections was significantly higher in boys than in girls overall ( p<0 . 001 ) , but there was no significant difference in mean egg counts among those who were infected ( P = 0 . 08 ) . There was no significant difference between the sexes in prevalence or intensity of schistosomiasis infection in 2015 ( p = 0 . 55 ) . Albendazole MDA for hookworm was implemented based on the average district prevalence of hookworm , therefore , within districts with moderate STH prevalence ( 20–49% ) , which received annual treatment , there were some schools with high ( ≥50% ) STH prevalence . Conversely , within high STH prevalence districts receiving bi-annual treatment there were some schools with moderate STH prevalence . We were therefore able to compare the effect of annual versus bi-annual treatment on the prevalence of hookworm in schools with high ( ≥50% ) prevalence of hookworm at baseline , and in schools with moderate ( 20–49% ) prevalence of hookworm at baseline ( Table 5 ) . Schools with high prevalence of hookworm at baseline that received bi-annual treatment had significantly lower mean prevalence at the time of the 2015 assessment than did high prevalence schools that received only annual treatment ( Welch’s t-test , p<0 . 0001 ) , in spite of the schools that received bi-annual treatment having slightly higher baseline prevalence ( Welch’s t-test , p = 0 . 03 ) . There was also a statistically significant reduction in arithmetic mean hookworm egg count . Annual and bi-annual treatment schedules had similar impact on the prevalence of infection in schools with moderate STH prevalence at baseline; mean prevalence of hookworm in 2015 was not significantly different for Moderate-baseline-prevalence schools that received two versus one round of albendazole ( p = 0 . 55 ) . Praziquantel MDA for schistosomiasis was implemented at the sub-district level according to the WHO treatment guidelines for the higher of the two school prevalence estimates within that sub-district: annual praziquantel ( PZQ ) MDA for all persons ≥5 years of age ( in high-prevalence subdistricts ) or PZQ every other year for school-age children only . As a result , all subdistricts containing a high-baseline-prevalence school received annual PZQ treatment . Moderate-baseline-prevalence schools received PZQ yearly or every other year depending on whether the other school in their particular subdistrict had high-baseline-prevalence or not . Table 6 shows the impact of yearly or biennial ( every other year ) treatment on the prevalence of both S . haematobium and S . mansoni . Among all schools with moderate baseline prevalence of S . mansoni , there was a significantly lower prevalence of S . mansoni among those schools receiving annual treatment for persons five years and older versus those receiving treatment every other year for school-age children only ( Table 6 ) . When only those schools that were last treated in 2014 were compared , there was no difference in the 2015 prevalence of S . mansoni between moderate baseline prevalence schools treated annually versus every two years ( 3 . 8% vs 2 . 0% , respectively , p = 0 . 21 ) . Among moderate baseline prevalence schools receiving PZQ every other year , there was a trend toward higher prevalence of S . mansoni infection in those schools that were last treated in 2013 versus those schools that were last treated in 2014 ( 7 . 7% prevalence of S . mansoni in 2013 vs . 2 . 0% in 2014; p = 0 . 07 ) . We used backward stepwise selection to generate mixed effects logistic regression models , including a random intercept for schools , to identify factors associated with hookworm , S . haematobium and S . mansoni infection in 2015 ( Table 7 ) . Baseline prevalence of infection was the strongest predictor of infection in 2015 for all three parasites . For hookworm , biannual albendazole distribution resulted in half the odds of infection in 2015 .
Togo has seen a significant reduction in the prevalence of hookworm and schistosomiasis infection in school-age children after four to five years of door-to-door distribution of albendazole and praziquantel to at-risk populations . The significant reduction in the prevalence of these infections is most likely attributable to the four to five years of carefully implemented MDA that has achieved programmatic coverage averaging 94% for albendazole among SAC and 95% for praziquantel among both SAC and adults [12] . This impact assessment was implemented about nine months after the last nationwide MDA and about four months after the second round of treatment in four high-prevalence STH districts , yielding two groups of schools that differ significantly in terms of frequency of treatment and target populations in their communities , and in the time since the last antihelminth MDA in the community . Baseline prevalence of hookworm was the strongest predictor of post-treatment prevalence , and also modified the effect of biannual treatment on disease reduction . In schools with high ( ≥50% ) hookworm prevalence at baseline , we observed a lower prevalence of infection in 2015 among those children receiving bi-annual albendazole as compared to annual albendazole ( Table 5 ) . This could reflect greater reduction in prevalence with repeated rounds of MDA and/or less time for resurgence of infection in areas receiving more frequent treatment; the latter is likely an important factor as communities receiving bi-annual treatment were treated four months prior to the evaluation while those receiving annual treatment had not been treated for nine months , allowing time for rebound of infection [13–15] . Schools with moderate prevalence of hookworm at baseline demonstrated the same reduction in prevalence of infection with either annual or bi-annual treatment . The rebound of infection four to nine months after treatment in areas with high baseline prevalence , even after five years of bi-annual MDA , indicates a pressing need for additional interventions to effect long-lasting reduction in the prevalence of hookworm . For S . mansoni , among schools with moderate baseline prevalence , we also saw a nearly four-fold higher prevalence of infection among individuals last treated two years prior to the 2015 impact assessment as compared with those treated one year prior to the assessment , suggesting a rebound of infection over time , but the post-MDA prevalence of S . mansoni was low for both groups and the difference in prevalence was of borderline statistical significance . Rebound of hookworm and schistosomiasis infection has been described elsewhere [14–16] . We observed no significant rebound of infection for S . haematobium among schools treated every other year compared to schools receiving annual treatment . Hookworm accounted for 97% of STH infection . Ascaris and Trichuris infections were very rare , which may reflect the decades of ivermectin distribution for onchocerciasis in most villages in Togo . Hookworm disproportionately affected boys , and prevalence rose steadily with age . These age and sex trends have been reported previously [17 , 18] , but the increase in hookworm prevalence across only three years of age is striking , particularly among boys . Nine-year-old boys were approximately 1 . 6 times more likely to be infected with hookworm than were six-year-old boys and were two to three times as likely to be infected as were six-year-old girls . Hookworm infection is unique among the three STH examined because prevalence of infection increases with age and peaks in adults [17 , 18] . Togo’s policy of targeting school-age children for treatment and measuring impact in this same group is aligned with WHO recommendations , and also reflects the extent of treatment that the country could support with available funding . However , adults can serve as an unmeasured reservoir of infection that poses a challenge to elimination of hookworm as a public health problem . There was a significantly lower mean baseline prevalence of hookworm ( 15 . 7% ) in the eight districts that had received MDA through Togo’s Lymphatic Filariasis Elimination Program–high-coverage MDA with ivermectin and albendazole for all persons age five years and older for six to eight years prior to the baseline survey–compared with those districts that had never received any albendazole ( 36 . 7% ) . But there was much variation in the 2009 prevalence of hookworm in the eight LF districts , hookworm prevalence prior to LF treatment is not known , and many factors contribute to differences in hookworm prevalence at baseline and to the effect of MDA on hookworm prevalence . Indeed , at the 2009 baseline , two of the LF districts had hookworm prevalence of 29% and 30% in spite of prior intensive MDA for LF; after 5 years of MDA with albendazole for SAC , while the first district witnessed a reduction in prevalence to 13% , the latter district still had STH prevalence of 27% in 2015 , in spite of evidence of good program coverage . Conversely , two LF districts with STH prevalence of 5% and 7% in 2009 showed no change in prevalence in 2015 in spite of having received no albendazole in the interim period . Without baseline data on STH infection prior to LF MDA it is difficult to draw conclusions about the impact of treating adults on the prevalence as measured in children , but certainly additional interventions will be needed to eliminate hookworm as a public health problem [19 , 20] . We did not observe a correlation between treatment coverage rates in 2014 and disease prevalence in 2015 for either STH or schistosomiasis . The absence of a correlation may be due in part to the fact that coverage was overall very high and there were few villages with reported poor treatment coverage ( <80% coverage ) in which we also had school-based data on disease prevalence; 30 schools were located in villages with coverage <80% for albendazole and another 30 schools were in villages with coverage <80% for praziquantel , so the sample size was small . Unmeasured factors that could have contributed to the observed reduction in disease prevalence must be considered . There is certainly un-programmed deworming that occurs in Togo but could not be measured or accounted for in this analysis . At 45% of the sampled schools the schoolmaster reported school-based deworming had occurred in the previous twelve months . Such deworming is typically led by local or regional non-governmental organizations; there is no government-led or supported school-based deworming program in Togo . We found no association between reported school-based deworming and STH or schistosomiasis infection in our mixed effect models , but other unmeasured deworming activities such as self-treatment could have had an impact . Improvements in water , sanitation and hygiene ( WASH ) infrastructure or practices could also have had an impact on STH and schistosomiasis prevalence , but we could not track those . The ministry of health did not report any large-scale , government-supported WASH activities during the period from 2009–2015 . A related publication using our study data examined the relationship between STH infection and individual and school-based WASH characteristics and practices and found that the associations are complex , underscoring the difficulty of establishing the impact of WASH even when practices are documented [21] . The granular nature of the sampling frame used in both the baseline and 2015 surveys revealed the highly focal nature of both schistosomiasis and STH infections , as reflected in Figs 2–7 . Programs that develop MDA targets based on prevalence estimates obtained from a small sample of sentinel sites risk over- or under-treating large segments of the population [10] . Granular epidemiological information is of particular importance for national NTD programs aiming to eliminate these diseases as public health problems , as the need for detailed data about where to target interventions increases as disease prevalence drops . In Togo , the distribution of S . mansoni is particularly geographically circumscribed and high-intensity S . mansoni infections show pronounced clustering . Distribution patterns such as these may be driven by determinants such as temperature , elevation and distance to large water bodies; these patterns also highlight the need to identify risk factors that affect transmission on sub-district , village , or even smaller spatial scales [22 , 23] . Additional epidemiological work is needed to investigate reasons for remaining hot spots of infection , and these activities , along with granular sampling schemes , are costly [10] . This assessment was conducted in the context of monitoring and evaluation of the Integrated NTD Program in Togo , and certain aspects of the methodology pose limitations for the interpretation of the results . Convenience sampling was used to select the children at each school , and these results may not be representative of the prevalence of infection among all children in Togo . Due to the constraints of implementing such a large national field study , only one stool sample was collected per person . Additionally , the time from collection to final reading of stool samples was longer than desired for hookworm , whose eggs are prone to degradation over time [24] . This may have reduced the sensitivity of the Kato-Katz for detecting hookworm , but the same laboratory set-up and procedures were used in 2009 , so for the purposes of comparison across years the processing times are likely similar , although time from stool collection to processing was not recorded in 2009 . For the baseline survey in Binah district , schistosomiasis data were recorded only at the school level , so dual schistosomiasis infections could not be identified . We therefore took the higher of the S . haematobium and S . mansoni prevalence estimates to represent the overall prevalence of schistosomiasis at each school at baseline , effectively assuming that dual infections occurred whenever both infections were observed in a school; this may have resulted in an underestimation of the proportion of children who had infection with at least one of the two species at baseline , which would in turn result in an underestimate of disease reduction from 2009 to 2015 . Another limitation relates to Binah district’s 2007 pilot baseline survey , in which recruited children were 9 or 10 years old , rather than 6 to 9 years old . Given the significant increase in prevalence with age , the enrollment of older children at baseline than at follow-up may have exaggerated the impact of MDA in Binah district . This cross-sectional impact assessment demonstrates that Togo has made significant progress in the control of STH and schistosomiasis through MDA with albendazole and praziquantel . The findings from this impact assessment have been used to amend target populations and treatment frequency to consolidate gains and intensify efforts in those areas with persistent high prevalence of infection; the optimal treatment algorithm may differ for different settings . More frequent treatment for STH resulted in greater reduction of infection and/or less rebound of infection , but elimination of both STH and schistosomiasis as public health problems does not appear attainable with the current treatment algorithms and WASH interventions in Togo . Interventions such as expansion of STH treatment to adults , greater access to clean water , improved sanitation and hygiene , and snail control are likely necessary to eliminate these diseases as public health problems , and the country risks rebound of these infections if funding for these programs and MDA is withdrawn . The findings from this study may help refine treatment recommendations for sustaining control or attempting elimination of these diseases . | Mass drug administration ( MDA ) is a key component of programs aimed at controlling soil-transmitted helminths ( STH ) and schistosomiasis , diseases that disproportionately impact individuals in developing countries and adversely affect physical and cognitive development . The World Health Organization recommends evaluating the impact of mass drug administration on the prevalence of these infections after five years of MDA . We present here a study of the impact of four to five years of MDA on the prevalence and intensity of STH and schistosomiasis infections in school children in Togo . The prevalence and intensity of these infections in 2015 were significantly reduced compared to a baseline survey conducted in 2009 . Local baseline prevalence in 2009 was the strongest predictor of infection in 2015 . These infections are more prevalent in boys than in girls , and in older versus younger children . We found that in areas with high baseline prevalence of hookworm the risk of rebound of infection is high among children who do not receive bi-annual treatment . This information is important for programs weighing the decision to stop MDA in areas where prevalence has been reduced through treatment . This and other findings from this study may help refine treatment recommendations for these diseases . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"schistosoma",
"invertebrates",
"schistosoma",
"mansoni",
"medicine",
"and",
"health",
"sciences",
"body",
"fluids",
"education",
"helminths",
"sociology",
"tropical",
"diseases",
"hookworms",
"social",
"sciences",
"parasitic",
"diseases",
"animals",
"urine",
"neglected",... | 2018 | Impact of community-based integrated mass drug administration on schistosomiasis and soil-transmitted helminth prevalence in Togo |
An important paradigm in evolutionary genetics is that of a delicate balance between genetic variants that favorably boost host control of infection but which may unfavorably increase susceptibility to autoimmune disease . Here , we investigated whether patients with psoriasis , a common immune-mediated disease of the skin , are enriched for genetic variants that limit the ability of HIV-1 virus to replicate after infection . We analyzed the HLA class I and class II alleles of 1 , 727 Caucasian psoriasis cases and 3 , 581 controls and found that psoriasis patients are significantly more likely than controls to have gene variants that are protective against HIV-1 disease . This includes several HLA class I alleles associated with HIV-1 control; amino acid residues at HLA-B positions 67 , 70 , and 97 that mediate HIV-1 peptide binding; and the deletion polymorphism rs67384697 associated with high surface expression of HLA-C . We also found that the compound genotype KIR3DS1 plus HLA-B Bw4-80I , which respectively encode a natural killer cell activating receptor and its putative ligand , significantly increased psoriasis susceptibility . This compound genotype has also been associated with delay of progression to AIDS . Together , our results suggest that genetic variants that contribute to anti-viral immunity may predispose to the development of psoriasis .
Psoriasis is an immune-mediated , inflammatory skin disease that is associated with arthritis and other systemic co-morbidities [1] . Psoriasis is a highly heritable condition , with a monozygotic twin concordance rate of 70% [2] and an estimated sibling recurrence risk λs of 4–11 [3] . Linkage studies [4]–[6] and genome-wide association studies ( GWAS ) [7]–[11] have identified over 20 psoriasis susceptibilities alleles . However , the locus consistently displaying the strongest association signal , by many orders of magnitude , is the major histocompatibility complex ( MHC ) . We were intrigued by the observation that several of the most highly significant SNPs from psoriasis GWAS were identical to the top SNPs from GWAS of HIV-1 virologic control , a clinical phenotype whereby certain HIV-1 infected individuals , termed “HIV-1 controllers , ” are able to maintain low levels of plasma HIV-1 RNA in the absence of antiretroviral therapy and who generally do not develop clinical symptoms [12] . For example , rs2395029 within the MHC gene HCP5 and a proxy for HLA-B*57 , was identified in a psoriasis GWAS as the SNP with the largest odds ratio , OR = 4 . 1 , p = 2 . 13×10−26 [8] . This same SNP has been shown to be the first or second most significant SNP in three GWAS of HIV-1 control [13]–[15] . The two most significant SNPs identified in multiple psoriasis GWAS are rs10484554 and rs12191877 near HLA-C ( r2 = 1 with each other in Europeans ) [8]–[10] . rs10484554 and rs12191877 were found to be associated with HIV-1 control [Supplementary Materials in [13] , [15]] and both are in moderate linkage disequilibrium ( r2 = 0 . 33 ) with rs9264942 , another top SNP for HIV-1 control [14]–[16] . The relationship between psoriasis and HIV-1 is also interesting because of the clinical observation that HIV-1 infection can exacerbate existing psoriasis or trigger new-onset psoriasis [17] . As HIV-1 infection progresses and CD4+ T cell counts decrease , psoriasis can worsen [18] , [19] . This has puzzled dermatologists and infectious disease clinicians because it has been convincingly established that psoriasis is an immune disorder that is mediated through activation of T cells . Several explanations for this “psoriasis HIV-1 paradox” have been proposed , including HIV-1 induced destruction of regulatory CD4+ T cells [20] , an increase in number of memory CD8+ T cells late in disease [21] , HIV-1 proteins acting as superantigens [22] , or co-stimulation through traditional antigenic presentation [20] . Due to these genetic and clinical observations , we pursued a more in-depth analysis of the HLA region in psoriasis to determine whether patients with psoriasis are enriched for the major genetic determinants of HIV-1 control . The psoriasis data generated in this study were compared to the largest GWAS for HIV-1 control performed to date , involving 516 cases and 1 , 196 controls and for which detailed HLA allele information was available [15] .
We imputed to four-digit resolution the HLA class I alleles ( -A , -B , -C ) and HLA class II alleles ( -DQA1 , -DQB1 , -DRB1 ) of 1 , 727 psoriasis cases and 3 , 581 healthy controls which were obtained from 3 separate case-control cohorts of European ancestry ( Table S1 ) . Imputation was performed using the software HLA*IMP , which has been shown to have an accuracy of at least 96% for class I loci and 92% for Class II loci [23] . To further validate the accuracy of our imputation , we compared the imputed HLA alleles to empirically obtained HLA class I alleles for a subset of our samples ( n = 98 ) . The concordance was 566/581 alleles ( 97 . 4% ) , indicating that the imputation was of high accuracy . A sensitivity analysis examining the imputation accuracy of low frequency HLA alleles ( allele frequency between 1% and 5% ) demonstrated similar high accuracy ( 177/179 alleles = 98 . 9% ) . Only HLA alleles with a minor allele frequency greater than 1% in the control group were used for subsequent analyses . We tested all imputed HLA alleles for association with psoriasis using logistic regression , adjusting for gender , ancestry , and cohort . The top ten HLA associations for psoriasis are shown in Table 1 ( Full four-digit and two-digit results in Tables S2 and S3 , respectively ) . Overall , we observed a striking pattern in which the HLA alleles which are enriched in psoriasis patients are also enriched in HIV-1 controllers , and the HLA alleles which have decreased frequency in psoriasis patients are also decreased in HIV-1 controllers . We found that psoriasis patients are highly enriched for HLA-B*57:01 ( 12 . 5% in cases vs 3 . 9% in controls , p = 5 . 50×10−42 , OR = 3 . 61 ) , which in multiple studies has been shown to be the most significant predictor of both HIV-1 control and delayed progression time to AIDS [14] , [15] , [24]–[27] . Psoriasis patients also display a significant enrichment of the HIV-1 control allele B*13:02 , whereas they display a relative paucity of B*07:02 , B*40:01 , and C*04:01 which are associated with lack of virologic control [15] . The HLA allele B*35 , almost always seen with C*04:01 , and the most significant HLA allele associated with rapid progression to AIDS [28] , [29] , was significantly protective against psoriasis in our dataset ( p = 3 . 20×10−6 , OR = 0 . 65 [0 . 54–0 . 78] ) . HLA-B*35 alleles can be segregated into B*35-Px and B*35-PY alleles , where Px alleles bind peptides with hydrophobic , non-tyrosine residues at position 9 and PY alleles bind peptides with tyrosine at position 9 . It has been shown that the influence of HLA-B*35 in accelerating progression to AIDS is mostly attributable to HLA-B*35-Px alleles [30] . In our psoriasis dataset , the B*35-Px alleles B*35:02 and B*35:03 together demonstrated a stronger effect on psoriasis protection ( p = 2 . 9×10−4 , OR = 0 . 47 [0 . 31–0 . 71] ) than the B*35-PY allele B*35:01 ( p = 5 . 86×10−3 , OR = 0 . 74 [0 . 60–0 . 92] ) . To identify HLA alleles independently associated with psoriasis , we performed stepwise regression modeling , first conditioning the association results on the top allele HLA-C*06:02 , and then adding alleles to the model in a stepwise manner . We identified HLA-C*06:02 , B*38:01 , A*02:01 , B*39:01 , B*27:05 , B*08:01 , B*14:02 , B*55:01 , and B*57:01 as HLA class I alleles independently associated with psoriasis ( Table 2 ) . In the multivariate regression model including all of these alleles , the HIV-1 viral control alleles B*57:01 and B*27:05 both had significant effect on psoriasis susceptibility ( OR = 1 . 52 and 1 . 75 , respectively ) . The contribution of B*27:05 was more apparent in the regression model than when B*27:05 was analyzed as a single allele ( p = 0 . 016 , OR = 1 . 32 [1 . 05–1 . 66] ) . The HIV-1 progression allele B*35 remained independently associated with psoriasis after conditioning on the top allele C*06:02 ( p = 0 . 0064 , OR = 0 . 77 [0 . 63–0 . 93] ) , but further conditioning on B*38:01 and A*02:01 resulted in a residual association signal for B*35 of p = 0 . 0168 , OR = 0 . 78 [0 . 64–0 . 96] ) . We also performed stepwise regression modeling combining class I and class II HLA alleles . At 4 digit resolution , C*06:02 , B*38:01 , DQB1*05:02 , DQB1*06:04 , and A*02:01 were found to be independent risk factors; however , at 2 digit resolution , the class II alleles DQB1*05 and DQB1*06 were no longer significant ( data not shown ) . We performed haplotype analysis in psoriasis patients and HIV-1 controllers to help understand how combinations of HLA alleles contribute to the observed association signals . We estimated the frequency of HLA haplotypes in our psoriasis case-control cohort as well as in 214 Caucasian HIV-1 infected individuals ( 52 HIV-1 controllers , 162 non-controllers ) in the SCOPE cohort , whose HLA class I and II alleles had been previously genotyped . Our analysis revealed that both psoriasis patients and HIV-1 controllers are highly enriched for the B*57:01–C*06:02 haplotype as well as the extended haplotype B*57:01–C*06:02–DQA1*02:01–DQB1*03:03–DRB1*07:01 , thus explaining why these individual alleles are associated with both phenotypes ( Table 3 and Table 4 ) . We found that the association of DQA1*02:01 , DQB1*03:03 , and DRB1*07:01 with psoriasis was nearly completely due to the effects of C*06:02 or B*57:01 , since conditioning DQA1*0201 and DRB1*0701 on C*06:02 resulted in p = 0 . 017 , OR = 1 . 21 and p = 0 . 199 , OR = 1 . 11 , respectively; and conditioning DQB1*0303 on B*57:01 resulted in p = 0 . 038 , OR = 1 . 33 . Thus , the primary genetic determinants of both psoriasis and HIV-1 control reside within the class I alleles . Specific amino acid positions within the peptide binding groove of HLA class I molecules have been shown to serve as important mediators for the protective and risk effects of individual HLA alleles on HIV-1 control [15] . Namely , amino acid residues at positions 97 , 67 , and 70 within HLA-B were found to be more highly associated with HIV-1 control than HLA-B*57:01 and each of these amino acid positions was found to serve as a strong predictor of HIV-1 viral load levels in an independent cohort [15] . To determine whether psoriasis is associated with the groups of alleles that are marked by specific amino acids within HLA proteins , we used the official protein sequences [31] assigned to each four-digit HLA allele to perform association testing at each amino acid position within HLA-A , -B , -C , -DQA1 , -DQB1 , and -DRB1 . As before , the association testing was adjusted for gender , ancestry , and cohort . We found that the five most significant amino acid positions for psoriasis occurred at 3 positions within HLA-B ( residue 97 [p = 1 . 58×10−53] , residue 67 [p = 4 . 00×10−45] , and residue 70 [p = 1 . 35×10−40] ) and 2 positions with HLA-C ( residue 156 [p = 3 . 89×10−51] and residue 97 [p = 4 . 56×10−45] ) ( Table S4 ) . Each of these 5 positions is located within the peptide-binding groove of the HLA molecule and directly contacts the bound peptide [32] . At each of these positions , we investigated whether the direction of the association signal was consistent between psoriasis patients and HIV-1 controllers . We confirmed that for each position examined , the amino acid residues associated with psoriasis susceptibility were associated with HIV-1 virologic control , and the amino acids associated with a protective effect on psoriasis risk were associated with HIV-1 progression ( Figure 1 ) . For example , alleles marked by Asn97 , Thr97 , and Val97 in HLA-B were associated with psoriasis susceptibility and HIV-1 control while those marked by Arg97 and Ser97 in HLA-B were associated with psoriasis protection and HIV-1 progression . To rule out the possibility that the observed similarities between psoriasis and HIV-1 control were the result of systematic bias of the imputation process or general amino acid variability at those positions , we examined whether these 5 amino acids positions were associated with other autoimmune or inflammatory diseases . We examined GWAS data from 5 diseases studied by the Wellcome Trust Case Control Consortium [33]—rheumatoid arthritis , Crohn's disease , type 1 diabetes , type 2 diabetes , and coronary artery disease—and used the same imputation process as performed with psoriasis . We found that none of these diseases displayed the degree of similarity between psoriasis and HIV-1 control when considering the direction and magnitude of the association signal at these amino acid positions ( Figure 1 ) . Crohn's disease , which shares some pathophysiological features with psoriasis [34] and is also slightly enriched for HLA-C*06:02 ( p = 4 . 2×10−5 , OR 1 . 32 ) and HLA-B*57:01 ( p = 3 . 68×10−4 , OR 1 . 40 ) , showed some similarity to psoriasis and HIV-1 control at these positions , but the magnitude of the association was smaller and several important residues such as Asn70 and Asn97 in HLA-B and Gln156 in HLA-C were not concordant . Interestingly , type 1 diabetics showed the opposite effect at many of these residues ( i . e . patients with type 1 diabetes lack HIV-1 control alleles and have an excess of HIV-1 progression alleles ) , which could support the theory that type 1 diabetes is triggered by a viral infection . Another important amino acid within HLA-B that may be relevant for HIV-1 progression is position 116 , which not only interacts with the carboxy-terminal residues of peptides in the F pocket , but also strongly influences the interaction of HLA class I molecules with the peptide-loading complex [35] , [36] . Studies of HLA-B*44:05 and HLA-B*44:02 , which only differ at position 116 , have shown that B*44:05 ( containing tyrosine at position 116 , “116Y” ) utilizes a tapasin-independent pathway that leads to a less optimal peptide repertoire compared to the tapasin-dependent HLA-B*44:02 ( 116D ) [36] . 116Y is strongly associated with lack of HIV-1 control , p = 1 . 6×10−10 , OR = 0 . 57 [15] . Our data show that 116Y is strongly associated with decreased susceptibility to psoriasis , p = 7 . 96×10−17 , OR = 0 . 66 [0 . 60–0 . 73] ( Table S4 ) . In our dataset , all HLA-B alleles containing 116Y had an odds ratio less than 1 . 0 , including B*07:02 ( p = 3 . 4×10−6 , OR = 0 . 71 ) , B*08:01 ( p = 0 . 092 , OR = 0 . 88 ) , B*35:02 ( p = 0 . 038 , OR = 0 . 54 ) , B*40:01 ( p = 7 . 24×10−6 , OR = 0 . 60 ) , B*40:02 ( p = 0 . 013 , OR = 0 . 53 ) , and B*51:01 ( p = 0 . 084 , OR = 0 . 82 ) . Thus , protection against psoriasis may be associated with presentation of a less-optimized peptide repertoire . Together , our data indicate that the genetic similarity between psoriasis patients and HIV-1 controllers extends to specific amino acid residues within class I molecules that mediate peptide binding , influence peptide loading , and which mark viral control or progression . We additionally performed stepwise regression to identify amino acid residues that were independently associated with psoriasis and found Trp156 and Ala24 in HLA-C; Val97 , Leu145 , Cys67 , and Tyr99 in HLA-B; and Gly107 in HLA-A to be markers independently associated with psoriasis ( Table S5 ) . Similar to HIV-1 control [15] , we found that HLA-B positions 97 and 67 remained in the model , while position 70 dropped out due to linkage disequilibrium with positions 97 and 67 . However , we caution against an interpretation that the amino acids identified here as independent are necessarily the functional ones . Due to the complex LD patterns between the amino acids , the final output of the stepwise regression model is affected by the starting variable , and potentially functionally significant amino acids can be lost because they are tagged by other residues . For example , HLA-B Gly62 , part of the α1 helix located in the B-pocket of the peptide binding groove , shows strong independent association with HIV-1 control ( p = 4 . 6×10−27 , OR = 5 . 03 ) [15] . Gly62 is also strongly associated with susceptibility to psoriasis ( p = 2 . 03×10−39 , OR = 3 . 20 ) but is in high LD with HLA-B Val97 and thus drops out of the final psoriasis model . Expression levels of HLA-C are modulated by the G/- polymorphism rs67384697 located within the 3′ UTR of HLA-C , where the presence of the deletion inhibits the binding of the microRNA hsa-miR-148 to the 3′UTR and results in higher HLA-C surface expression [37] . In a multivariate model of HIV-1 control , the deletion allele of rs67384697 has a strong effect on viral control independent of the classical alleles HLA-B*57:01 and HLA-B*27:05 , although the high LD of rs67384697 with other HLA-B and HLA-C alleles makes it difficult to determine whether rs67384697 ( high HLA-C expression ) is directly mediating this effect , or whether the HLA alleles themselves are causal [37] . Nevertheless , rs67384697 has been proposed to be the functional variant that explains the previously identified protective effect of rs9264942 on HIV-1 control , where rs9264942 is located -35 kb upstream of HLA-C and is in moderately high LD with rs67384697 ( r2 = 0 . 74 ) . We investigated whether rs67384697 was associated with psoriasis by imputing the deletion genotype of all psoriasis cases and controls . This was made possible by the near perfect LD between HLA-C four digit classical alleles and presence or absence of the deletion [Supplementary Table 2 in [37]] . To confirm the validity of our imputation , we sequenced the region of the HLA-C 3′UTR containing rs67384697 in a subset of our samples ( n = 70 ) and found a concordance rate of 138/140 alleles ( 98 . 6% ) , indicating the imputation was robust . Using logistic regression and adjusting for sex , ancestry , and cohort , we found that deletion allele of rs67384697 was significantly associated with psoriasis ( p = 1 . 02×10−29 , OR = 1 . 72 ) ( Table 5 ) , again confirming the similarity between psoriasis patients and HIV-1 controllers . We found that the association of rs67384697 with psoriasis was largely driven by HLA-C*06:02 , since conditioning on HLA-C*06:02 resulted in only a marginally significant p-value for rs67384697 ( p = 0 . 044 , OR = 1 . 12 ) . We note , however , that among all HLA-C allotypes , HLA-C*06:02 shows the highest level of cell surface expression , which could explain , to some extent , its strong association with psoriasis . Natural killer ( NK ) cells , a major component of the innate immune system , respond in the early stages of viral infection by producing cytokines and killing infected cells . NK-cell responses are regulated in part by activating and inhibitory killer immunologlubulin-like receptors ( KIRs ) on NK cells which engage HLA class I molecules on target cells . The activating KIR allele KIR3DS1 on chromosome 19 , alone or in combination with its putative HLA-B ligand Bw4 , has been associated with delayed progression to AIDS and improved HIV-1 outcomes [38]–[43] . The HLA-B Bw4 epitope can be identified by the presence of isoleucine or threonine at amino acid position 80 , whereas the Bw6 epitope contains asparagine at position 80 . Our HLA data revealed that psoriasis is associated with HLA-B alleles carrying the Bw4 epitope ( p = 1 . 28×10−25 , OR = 1 . 66 , Table S4 ) . The association was stronger for Bw4-80I [isoleucine] ( p = 8 . 28×10−22 , OR = 1 . 80 ) than for Bw4-80T [threonine] ( p = 3 . 41×10−4 , OR = 1 . 22 ) , which is interesting because Bw4-80I is thought to have a higher binding affinity for its KIR receptor than Bw4-80T [44] . We therefore hypothesized that psoriasis susceptibility might be mediated through activation of NK cells through KIR3DS1 and its putative partner HLA-B Bw4-80I . We genotyped KIR3DS1 in a subset of our psoriasis samples ( n = 397 ) and compared the results to a healthy control cohort with available KIR3DS1 and HLA genotypes ( n = 282 ) . We found that the presence of the compound genotype KIR3DS1+Bw4-80I was a strong risk factor for psoriasis ( frequency 22 . 7% in cases vs 6 . 9% in controls , p = 1 . 54×10−7 , OR = 3 . 92 , Table 6 ) . Individuals positive for KIR3DS1 but lacking Bw4-80I had no increased risk for psoriasis ( p = 0 . 63 , OR = 0 . 91 ) , and individuals positive for Bw4-80I but lacking KIR3DS1 had only a borderline increased risk of psoriasis ( p = 0 . 058 , OR = 1 . 53 ) . To our knowledge , this is the first report that the compound genotype KIR3DS1+Bw4-80I is a strong risk factor for psoriasis susceptibility . This finding is again consistent with our observation that there is significant genetic similarity between psoriasis patients and HIV-1 controllers; however , replication of the KIR3DS1+Bw4-80I association in additional psoriasis cohorts is warranted .
In this study , we followed up on the observation that several of the top SNPs from genome-wide association studies of psoriasis were identical to the top SNPs from genome wide association studies of HIV-1 control . Using imputation of HLA alleles , we found that psoriasis patients are enriched for several of the most significant known genetic variants associated with HIV-1 control: HLA-B*57 and HLA-B*27 , which are associated with decreased viral load and increased time to AIDS [14] , [15] , [24]–[27]; specific amino acid residues at HLA-B positions 97 , 67 , and 70 that are strong markers of HIV-1 controller status and viral load [15]; the deletion SNP rs67384697 which is associated with decreased viral load independent of HLA-B*57 and HLA-B*27 [37]; and the activating KIR3DS1 allele in combination with HLA-B Bw4-80I [42] . Psoriasis patients also demonstrate a significant paucity of HLA alleles and variants associated with HIV-1 disease progression [15] , [28] , including HLA-B*35 ( especially B*35-Px ) , B*07:02 , B*40 , C*04:01 , C*07 , and tyrosine 116 in HLA-B associated with sub-optimal peptide loading ( Table S2 , Table S4 ) . These effects were consistent between the 3 psoriasis cohorts examined in this study , demonstrating that the effects observed were real ( Table S6 ) . An important question to address , however , is whether the structural similarity between HLA alleles in psoriasis and HIV-1 control reflects the same underlying causal variants , or merely a coincidental association due to linkage disequilibrium . Our data suggest that some , but not all , of the observed similarity can be attributed to linkage disequilibrium . Our haplotype analysis shows that both psoriasis patients and HIV-1 controllers are enriched for the same extended haplotype , B*57:01–C*06:02–DQA1*02:01–DQB1*03:03–DRB1*07:01 . In the HIV-1 controllers , this haplotype is likely primarily driven by selection for B*57:01 , since previous studies have shown that the association of C*06:02 with HIV-1 control is dependent on B*57:01 in Europeans [15] and B*5801 in Africans [45] , although one indirect benefit of C*06:02 for HIV-1 control is its high LD ( D′ = 1 ) with the rs67384697 deletion polymorphism . In psoriasis , the haplotype association appears to be driven more by C*06:02 than B*57:01 , since C*06:02 remains significant after conditioning on B*57:01 ( p = 6 . 86×10−39 , OR = 3 . 04 ) and a number of C*06:02 haplotypes that do not contain B*57:01 still remain associated with psoriasis ( Table 3 ) . In addition , the association of the deletion allele of rs67384697 with psoriasis appears to be largely driven by LD with C*06:02 . However , it should be noted that variants in high LD with C*0602 may be contributing to the observed association signal for C*0602 . Interestingly , one can take the association signal for C*06:02 in psoriasis and perform stepwise conditioning on all coding amino acids within C*06:02 to demonstrate that the coding residues do not account for the entire association signal ( Table S7 ) . Therefore , the association of C*06:02 with psoriasis reflects , in part , other variants in high LD with C*06:02 . Despite the effects of linkage disequilibrium , our data suggest that some HIV-1 control variants indeed contribute independently to psoriasis susceptibility . First , both HLA-B*57:01 and HLA-B*27:05 remain associated with psoriasis after conditioning on C*06:02 ( B*57:01 p = 1 . 43×10−3 , OR = 1 . 45; B*27:05 p = 4 . 83×10−4 , OR = 1 . 52 ) and both remain independently associated with psoriasis in our stepwise regression model ( Table 2 ) . A previous analysis of the HLA region in psoriasis by Feng et al . [46] also found that B*57:01 was independent of C*06:02 in Caucasians; moreover , in this study B*57:01 was also found to be independent of C*06:02 in a Chinese psoriasis cohort , which is notable because the LD between C*06:02 and B*57:01 is lower in Asians ( D′ = 0 . 41 ) compared to Europeans ( D′ = 0 . 90 ) [47] . Prior studies have also shown that B*27 is a strong risk factor for psoriasis in the subset of patients with psoriatic arthritis , especially those with axial disease [48]–[51] . Second , linkage disequilibrium with C*06:02 does not explain the lower frequency of the HIV-1 progression allele B*35 in psoriasis , nor can it account for the concordance of amino acid residues at HLA-B positions 67 , 70 , and 97 whose association with psoriasis was shown to be independent of HLA-B*57:01 and HLA-C*06:02 ( Cys67 , Ser67 , Lys70 , Asn97 , Arg97 , see Figure 1 . legend ) . Finally , the provisional association of KIR3DS1+HLA-B Bw4-80I with psoriasis cannot be due to linkage disequilibrium , because KIR3DS1 is located on chromosome 19 which segregates independently of chromosome 6 . Although B*57:01 , B*27:05 , and possibly B*35 may have independent effects in psoriasis , additional studies are needed to clarify the precise mechanism ( s ) by which these and other psoriasis-associated HLA alleles contribute to psoriasis susceptibility or protection . The observation that psoriasis patients and HIV-1 controllers display concordant amino acids within the peptide binding groove of HLA-B suggests the possibility that an unknown psoriasis antigen shares homology with HIV-1 epitopes . An alternative possibility is that B*57:01 , B*27:05 , and B*35 do not restrict antigen presentation in psoriasis , but primarily function through their ability or inability to activate NK cells . We have provisionally shown a strong effect of KIR3DS1+Bw4-80I on psoriasis susceptibility , and B*57:01 contains the Bw4-80I epitope . The second strongest HLA allele in our stepwise regression model , B*38:01 , also contains the Bw4-80I epitope . B*27:05 contains the Bw4-80T epitope , while protective alleles B*35 and B*40 contain the Bw6 epitope , which do not serve as ligands for KIR . Previous studies have shown that the activating KIR allele KIR2DS1 also contributes to psoriasis or psoriatic arthritis susceptibility [52]–[55] , supporting the notion that NK cells may play a role in the pathogenesis of psoriasis . Finally , we have discussed the potential role of peptide processing on susceptibility to psoriasis , with the presence of tyrosine at HLA-B position 116 associated with protection against psoriasis , where position 116 is located near the C terminus of the bound peptide . A role for peptide processing influencing psoriasis risk has been previously identified for the gene ERAP1 , an amino peptidase which regulates the quality of peptides bound to MHC class I molecules through trimming the peptide N terminus [10] . The genetic similarity between psoriasis patients and HIV-1 controllers has interesting implications . On a population level , the data would predict that Caucasian individuals with psoriasis are more likely than Caucasian individuals without psoriasis to be HIV-1 controllers , and HIV-1 controllers are more likely than non-controllers to develop psoriasis . This does not imply that every individual with psoriasis will be an HIV-1 controller , since only a fraction of psoriasis patients will harbor , for example , B*57 , and even the presence of B*57 does not guarantee HIV-1 control , as this allele is present in some HIV-1 progressors . Nevertheless , one would expect an enrichment of HIV-1 controllers in the psoriasis population relative to a non-psoriatic population . Our data also suggest a hypothesis that the existence of psoriasis may represent aberrant activation of evolutionarily-derived viral control alleles [56] . Barreiro et al . have shown that several of the top MHC SNPs associated with both psoriasis and HIV-1 control reside on haplotypes which show strong evidence of recent positive selection in the genome , as evidenced by long haplotypes indicative of rapid expansion of an advantageous allele in the population [57] . Psoriasis could subsequently result from the activation of viral control alleles due to the presence of a psoriasis antigen with sequence homology to HIV-1 , or due to other environmental triggers . Although this study has focused on HLA and KIR alleles , other non-MHC psoriasis genes are plausibly associated with host response to viral infection . ERAP1 is involved in class I peptide processing and demonstrates epistasis with C*06:02 [10] and IFIH1 encodes a cytoplasmic helicase that mediates induction of interferon response to viral RNA [10] , [58] . TNIP1 , TNFAIP3 , TRAF3IP2 , NFKBIA , and REL are associated with the TNF-α pathway and activation of NF-κB; while IL23R , IL12B , IL23A , and TYK2 are associated with activation of the Th17 pathway . Psoriasis is characterized by the upregulation of the cytokines IFN-α , IFN-γ , TNF-α , IL-17 , IL-22 , and IL-23 [59] , while TNF-α , IFN-γ , and Th17+ T cells have been shown to be important in HIV-1 controllers [60]–[62] . The enrichment of viral control alleles in psoriasis patients may also help explain the psoriasis HIV-1 paradox ( Figure 2 ) . Psoriasis patients are more likely to harbor alleles such as HLA-B*57 , HLA-B*27 , HLA-C*06 ( in high LD with the HLA-C 3′UTR deletion polymorphism ) , and KIR3DS1+Bw4-80I , theoretically resulting in vigorous cytotoxic T cell and NK cell responses upon infection with HIV-1 virus . The pro-inflammatory environment created by these anti-viral responses , resulting in the production of cytokines such as TNF-α and IFN-γ , would worsen the psoriasis . In addition , if the psoriasis antigen had sequence homology to HIV-1 , then antigen specific immune responses directed against HIV-1 might cross-react with the psoriasis antigen and also flare the psoriasis . In either case , reduction of viral load and removal of the antigenic trigger through treatment with anti-retroviral therapies would improve the psoriasis , which is indeed seen clinically [17] . This explanatory model is consistent with several observational studies that patients with severe psoriasis and HIV-1 infection tend to carry the HLA-C*06 and HLA-B*27 alleles [63] , [64] , because such alleles would trigger the vigorous immune response associated with exacerbation of psoriasis . The data presented here with psoriasis and HIV-1 control illustrate the delicate balance of the human immune response , in which processes that safeguard the body against pathogens may also engage deleterious inflammatory responses . A similar example occurs with a genetic variant in the SH2B3 gene which may be protective against bacterial infection but which increases susceptibility to celiac disease , an autoimmune disease of the gut resulting from gluten intolerance [65] . Another example can be seen with the identification of genetic variants in immune function genes that increase the risk of sepsis , a systemic inflammatory response to infection which can lead to death [66] , [67] . In summary , using a large dataset of psoriasis cases and controls , we have shown that psoriasis patients and HIV-1 controllers share a high degree of similarity at their HLA loci . While some of this similarity is attributable to linkage disequilibrium , we present evidence that much of the similarity may be attributable to shared biological mechanisms including activation of natural killer cells , specificity of antigen presentation , and use of optimal MHC class I peptide processing . The genetic similarity between psoriasis and HIV-1 control suggests the possibility that psoriasis represents aberrant activation of pathways associated with anti-viral immunity . If this hypothesis is true , then the study of the biological pathways active in psoriasis may provide new therapeutic insights for the treatment of HIV-1 .
The study population and source are shown in Table S1 . Two independent genome-wide association scan datasets were used as cohort 1 and cohort 2 in the present study . All cases and controls were of European descent . More details on subject characteristics and recruitment can be found in Liu et al . [8] and Nair et al . [9] . Only the subjects whose HLA alleles were successfully imputed ( see below ) were included in our analysis . In cohort 3 , 169 psoriasis cases recruited from Washington University , St . Louis were directly typed for the class I HLA alleles by combining locus-specific amplification with hybridization of sequence-specific oligonucleotide probes as described in [68]; 1 , 711 control samples of European ancestry were obtained from studies 66 and 67 of illumina iControlDB . There was no overlap between the subjects among the three cohorts . Informed consent was obtained from each participant . Data generated by this study were primarily compared against a published genome-wide association study of HIV-1 control involving 516 HIV controllers of European ancestry ( viral load <2 , 000 RNA copies/ml by three measurements over at least 12 months without antiviral therapy ) and 1 , 196 controls ( treatment-naïve chronically infected individuals with advanced disease , median viral load 61 , 698 copies/ml ) [15] . HLA haplotype analysis was performed on 214 Caucasian HIV-1 infected individuals ( 52 HIV-1 controllers , 162 non-controllers ) from the SCOPE cohort ( Study of the Consequences of the Protease Inhibitor Era ) , whose HLA class I and II alleles had been previously directly genotyped . SCOPE HIV-1 controllers were antiretroviral therapy-naïve subjects who had at least one year duration of documented plasma HIV RNA below 2 , 000 copies/ml , while SCOPE non-controllers were subjects who had at least one documented viral load above 10 , 000 copies/ml . The Wellcome Trust Case-Control Consortium data were obtained from the WTCCC official website ( http://www . wtccc . org . uk/ ) . In this study , we used Affymetrix 500 K genotyping data from approximately 2 , 000 samples from each of five diseases ( rheumatoid arthritis , Crohn's disease , type 1 diabetes , type 2 diabetes , and coronary artery disease ) and 3 , 000 shared control samples from the 1958 birth cohort ( 58C ) and the National Blood Service ( NBS ) . More details about these samples are described elsewhere [33] . KIR3DS1 typing was performed on 397 psoriasis subjects from cohorts 2 and 3 described above . Control HLA and KIR3DS1 data were obtained from 282 healthy Caucasian blood donors from the Carrington laboratory . The program HLA*IMP [23] was used to impute HLA loci -A , -B , -C , -DQA1 , -DQB1 and -DRB1 to 4-digit resolution in our genome-wide SNP datasets . Individuals or SNPs with a missing data frequency above 0 . 20 were excluded as recommended in the software manual . A call threshold of 0 . 7 on the modes of the posterior HLA type distributions was employed , which represents a good compromise between accuracy and call rate as suggested by the author . Imputation accuracy was assessed by comparing the imputed HLA alleles to directly typed HLA class I alleles in a subset of our samples ( n = 98 ) , comprising 42 samples imputed from Illumina 300 K SNP data and 56 samples imputed from Perlegen SNP data . We found that HLA*IMP produces highly accurate HLA type imputations at HLA class I loci at the 4-digit level . The concordance for the Illumina 300 K platform was 244/252 alleles ( 96 . 8% ) and the concordance for the Perlegen platform was 322/329 alleles ( 97 . 9% ) , for an overall concordance rate of 566/581 alleles ( 97 . 4% ) . To examine the imputation accuracy of infrequent/rare HLA alleles , we identified all HLA class I alleles with a population frequency of less than 5% in individuals of European descent , according to the online database: http://www . allelefrequencies . net . We then examined the accuracy of imputation at the 4 digit level for these infrequent/rare HLA alleles in our subjects for whom we had both directly genotyped HLA alleles and imputed HLA alleles . The concordance of HLA alleles with frequency <5% was 95/102 alleles ( 93 . 1% ) for the Illumina 300 K platform and 116/122 alleles ( 95 . 1% ) for the Perlegen platform , for an overall concordance rate of 211/224 alleles ( 94 . 2% ) . However , our manuscript excludes HLA alleles with frequency less than 1% . For HLA alleles with frequency greater than 1% but less than 5% , the concordance was 78/79 alleles ( 98 . 7% ) for Illumina 300 K and 99/100 alleles ( 99 . 0% ) for Perlegen , for an overall concordance rate of 177/179 alleles ( 98 . 9% ) . Additive logistic regression models in PLINK [69] were used for most of the association tests , except for the HLA haplotype association tests . To account for potential population stratification or admixture in these samples , principal component analyses ( PCA ) was performed using the EIGENSTRAT [70] . Seven PCs in cohort 1 and ten PCs in cohort 2 were used for ancestry adjustment , based on leveling off of the PCA scree plot . The principal component score for each individual was included as a covariate in all models along with cohort and gender in logistic regression models . Multivariate logistic regression was performed in R software package ( http://www . r-project . org/ ) . To examine the consistency of association signals seen in the 3 cohorts used , a heterogeneity index was calculated using the meta-analysis module in PLINK . Conditional and stepwise logistic regression was performed using the ‘condition’ function in PLINK to determine whether independent effects existed . The method begins with an empty model to which variables are added in an iterative process as described by Barcellos et al [71] . Briefly , starting with HLA-C*06:02 which exhibits the strongest association with psoriasis , we conditioned candidate HLA alleles on C*06:02 to determine the next most significant independent effect . For the model including both class I and class II alleles , the iterative process completes when no candidate allele demonstrates p<0 . 0006 , which corresponds to the Bonferroni correction for the 88 HLA candidate alleles with MAF>1% tested . For the model including only class I alleles , the iterative process completes when no candidate allele demonstrates p<0 . 00096 , which corresponds to the Bonferroni correction for the 52 class I candidate alleles with MAF>1% tested . The amino acid sequence of all HLA alleles is completely determined by the HLA type at four-digit resolution . We used the official amino acids sequences defined for known HLA alleles [31] and our imputed HLA allele data to determine the frequency of amino acid residues in cases and controls . HLA amino acids residues were tested for association using a logistic regression model that corrects for population substructure , gender and cohort using PLINK . For amino acid positions with >2 alleles , the omnibus test in the conditional haplotype analysis module in PLINK was used to determine a single p-value for all alleles at that position . We performed stepwise logistic regression in PLINK to determine the amino acid residues that were independently associated with psoriasis , using the same approach as done with the HLA alleles . For the amino acid analysis , the algorithm completes when no remaining candidate residue has p<0 . 0001 , which corresponds to the Bonferroni correction for the 480 HLA amino acid residues with MAF>1% tested . For the stepwise regression modeling of HLA-C 06:02 association signal ( Table S7 ) , the algorithm completes when no candidate residue has p<0 . 0006 ( 0 . 05/87 ) . Since there is strong linkage disequilibrium between specific HLA-C alleles and rs67384697 [Supplementary Table 2 in [37]] , we were able to determine the rs67384697 genotype of all subjects using their HLA-C four digit classical alleles . To ensure the accuracy of our imputation , we directly sequenced ( ABI 3730 DNA analyzer , Quintara Biosciences , Berkeley , CA ) the region of the HLA-C 3′UTR containing rs67384697 in a subset of our samples ( n = 70 ) and determined a high concordance rate of 138/140 alleles ( 98 . 6% ) . The following primers were used for sequencing of genomic DNA samples: forward 5′-gtgagattctggggagctga and reverse 5′-gaacagcaactaggcacagg as specified in [37] . Arlequin V3 . 5 , based on the EM algorithm , was used to estimate the frequency of HLA allele haplotypes in our psoriasis cohorts and the SCOPE HIV cohort . To ensure the accuracy of haplotype construction , we compared the haplotype frequencies generated by Arlequin to the frequencies obtained by direct counting of the phased HLA alleles output from HLA*IMP , and found the two methods to yield nearly identical results . Haplotype frequencies were tested for statistically significant differences between case and control groups using the Chi Square test or Fisher's exact test in the R software package . KIR3DS1 genotyping was performed by using multiplex PCR-SSP ( sequence-specific priming ) according to Kulkarni et al with other minor modifications [72] . Briefly , each reaction contained 15 ng of DNA , 200 µM dNTP , 1 . 5 mM MgCl2 , 0 . 5 µl 10× PCR buffer , 1 µM of each primer for KIR3DS1 and KIR3DL1 and 0 . 8 µM of each primer for HLA-DRB1 , and 0 . 025 µl of Platinum Taq polymerase ( Invitrogen , Carlsbad , CA ) in a 5 µL final volume . The polymerase chain reaction ( PCR ) conditions were: 3 min at 94°C; 5 cycles of 15 s at 94°C , 15 s at 65°C , 30 s at 72°C; 25 cycles of 15 s at 94°C , 15 s at 60°C , 30 s at 72°C; 4 cycles of 15 s at 94°C , 1 min at 55°C , 2 min at 72°C followed by a final 7 min extension step at 72°C . To confirm the accuracy of the results , samples were replicated using a second set of KIR3DS1 and KIR3DL1 primers from [73] . Phenotype frequencies for the presence of each gene were estimated by direct counting . Frequency differences between psoriasis and control groups were tested for significance by two-sided Fisher's exact test . | Individuals with autoimmune disease generally demonstrate excessive immune system activation , leading to inflammation and damage of specific target organs . However , in some cases the detrimental effects of an overactive immune system might be counterbalanced by a beneficial effect in protecting against certain infections . In this study , we investigated whether patients with psoriasis , a common autoimmune disease of the skin , harbor genetic variants that are associated with an enhanced ability to limit replication of the HIV-1 virus . We profiled the HLA ( human leukocyte antigen ) immune genes located on chromosome 6 in 1 , 727 Caucasian psoriasis cases and 3 , 581 healthy controls and found that psoriasis patients are significantly more likely than controls to have gene variants that are protective against HIV-1 disease . We found that this enrichment for HIV-1 protective variants was unique to psoriasis and largely absent in patients with other autoimmune or inflammatory diseases such as rheumatoid arthritis , Crohn's disease , type 1 diabetes , type 2 diabetes , and coronary artery disease . Our results suggest the possibility that the excessive skin inflammation in psoriasis may be associated with activation of anti-viral immune pathways that were important to human ancestors who encountered viruses similar to HIV-1 . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"medicine",
"dermatology",
"infectious",
"diseases",
"public",
"health",
"and",
"epidemiology",
"sexually",
"transmitted",
"diseases",
"epidemiology"
] | 2012 | Psoriasis Patients Are Enriched for Genetic Variants That Protect against HIV-1 Disease |
Understanding mosquito host choice is important for assessing vector competence or identifying disease reservoirs . Unfortunately , the availability of an unbiased method for comprehensively evaluating the composition of insect blood meals is very limited , as most current molecular assays only test for the presence of a few pre-selected species . These approaches also have limited ability to identify the presence of multiple mammalian hosts in a single blood meal . Here , we describe a novel high-throughput sequencing method that enables analysis of 96 mosquitoes simultaneously and provides a comprehensive and quantitative perspective on the composition of each blood meal . We validated in silico that universal primers targeting the mammalian mitochondrial 16S ribosomal RNA genes ( 16S rRNA ) should amplify more than 95% of the mammalian 16S rRNA sequences present in the NCBI nucleotide database . We applied this method to 442 female Anopheles punctulatus s . l . mosquitoes collected in Papua New Guinea ( PNG ) . While human ( 52 . 9% ) , dog ( 15 . 8% ) and pig ( 29 . 2% ) were the most common hosts identified in our study , we also detected DNA from mice , one marsupial species and two bat species . Our analyses also revealed that 16 . 3% of the mosquitoes fed on more than one host . Analysis of the human mitochondrial hypervariable region I in 102 human blood meals showed that 5 ( 4 . 9% ) of the mosquitoes unambiguously fed on more than one person . Overall , analysis of PNG mosquitoes illustrates the potential of this approach to identify unsuspected hosts and characterize mixed blood meals , and shows how this approach can be adapted to evaluate inter-individual variations among human blood meals . Furthermore , this approach can be applied to any disease-transmitting arthropod and can be easily customized to investigate non-mammalian host sources .
Many insects require a blood meal to complete their gonotrophic cycle . By feeding successively on different hosts , these insects can transmit blood borne pathogens that cause diseases responsible for significant burden on global health [1 , 2] . In particular , insects that seek human blood meals are vectors of devastating diseases such as malaria , dengue fever , sleeping sickness , filariasis , leishmaniasis , typhus and plague . Understanding the complex blood feeding patterns of the insects transmitting these human diseases is crucial for developing and prioritizing vector-based control program activities and identifying potential unrecognized disease reservoirs , and thus for reducing disease transmission and burden . The blood meals of arthropods have traditionally been analyzed using serological techniques such as ELISA or precipitin tests [3–5] . While these methods have provided valuable information , they have limited taxonomic resolution as they are generally only able to characterize hosts at the order or family levels [6] . In addition , since these approaches test for the presence of a protein from a specific organism , they only test for absence/presence of organisms that are a priori believed to be blood meal hosts . More recently , a number of PCR-based molecular techniques have been developed to characterize host blood meals ( [7] and references within ) and determine the blood feeding preference of mosquitoes [8–11] , ticks [12–14] , sandflies [15–17] and Tsetse flies [18 , 19] . While these PCR-based approaches enable rigorous identification of the host species , they typically focus on species-specific amplification of putative hosts and therefore are not designed to identify novel , unanticipated host blood sources . In addition , the detection of mixed blood meals ( i . e . , when an insect feeds on more than one host ) by these approaches is complicated as the dominant host signal can completely overwhelm signals from other minor hosts . These limitations may have biased our understanding of the transmission of many vector-borne diseases and have prevented identification of important disease reservoirs . Beyond the identification of the host species , it may also be important to understand which individuals of a given species are being fed upon: for example , knowing whether an insect preferentially bites specific individuals or , in contrast , feeds on multiple individuals per night , could influence our assessment of disease transmission . A number of studies have used microsatellites or other polymorphic genetic markers to generate individual DNA fingerprints from human blood meals of mosquitoes [20–25] and lice [26 , 27] . However , interpretation of these data can become complicated if DNA from more than one individual is present in a single blood meal . Anopheles punctulatus sensu latu ( s . l ) mosquitoes are the principal vectors of malaria and lymphatic filariasis in Papua New Guinea ( PNG ) and the South Pacific [28] . There are 13 sibling species in An . punctulatus s . l , five of which are major disease vectors: An . punctulatus s . s . , An . koliensis , An . farauti s . s . , An . hinesorum and An . farauti 4 . While these species have been little studied , they are generally characterized as unspecialized with regards to their feeding behaviors and ecological preferences [29] and shown to feed roughly indiscriminately on humans , dogs and pigs that are the most abundant species found in PNG villages [30 , 31] . Here , we describe a novel approach using next-generation sequencing technology to analyze the blood meal composition of individual mosquitoes in an unbiased manner . We first amplify DNA extracted from a single female mosquito using universal primers targeting the mammalian mitochondrial ( mt ) 16S rRNA genes . Following individual barcoding , PCR products from up to 96 mosquitoes are pooled and simultaneously sequenced using Illumina high-throughput sequencing methods . We also use the same approach to interrogate whether individual mosquitos fed on more than one person by sequencing a highly polymorphic region of the human mt hypervariable region I . We applied this approach to 442 Anopheles punctulatus sensu lato ( s . l ) mosquitoes captured in five villages of the Madang Province of Papua New Guinea and provide evidence that ( i ) Anopheles punctulatus s . l . mosquitoes feed on a variety of mammalian species , including several unanticipated hosts , and ( ii ) Anopheles punctulatus s . l . mosquitoes frequently feed on multiple mammalian hosts . We also show how this assay can be easily customized to examine the number of individual hosts within a specific species . Overall , our results illustrate the potential of this approach to comprehensively characterize host species for any blood feeding arthropods , to identify reservoirs of pathogens and to provide opportunities for developing better evidence-based strategies to decrease transmission of important infectious diseases .
This study was approved by the Papua New Guinea Institute of Medical Research Institutional Review Board ( 1203 ) and PNG Medical Research Advisory Board ( 12 . 05 ) . We collected mosquitoes from the villages of Dimer , Wasab , Kokofine , Mirap and Matukar in the Madang province of Papua New Guinea ( PNG ) in June and August 2012 . In each village , field technicians collected mosquitoes between 1800 and 0600 using barrier screens as described by Burkot et al [32] . These screens were manually searched every 20 minutes and resting mosquitoes were captured from the screens using an aspiration device . After collection , the sex and species of each mosquito were determined by morphology as previously described [33] . All male mosquitoes and non-Anopheles mosquitoes were discarded . We visually classified each female Anopheles mosquito as non-fed , partially-fed or fully-fed by examining the size and coloration of their abdomen . We individually stored each mosquito in an Eppendorf tube containing silica gel as desiccant . We extracted DNA from individual mosquitoes using a 96 well Qiagen DNeasy blood and tissue kit as previously described [34] . Mosquito species identification was determined using a PCR-based assay that evaluates species-specific polymorphisms in the ribosomal RNA internal transcribed spacer unit 2 ( ITS2 ) [35] . To test the range of mammals that should be amplified using mt 16S rRNA primers [36] , we conducted an in silico analysis using the primerTree package . We also conducted in silico analyses for two other previously published primers , cytochrome oxidase I ( COI ) and cytochrome b ( Cytb ) that have been previously used for mosquito blood meal identification [37] . Since the 16S rRNA locus appeared to be the most informative for our purposes ( S1 Fig ) , we restricted our further analyses to this locus . Briefly , we performed primer-BLAST searches using the mammalian mt 16S rRNA primer sequences against the NCBI nucleotide database using default parameters but allowing for up to three mismatches in the primer sequences . In our search , we set the maximum number of blast hits retrieved to 10 , 000 and retrieved the taxonomical information of each sequence retrieved . As this search can be biased by recent release of many DNA sequences from a specific taxon , we performed this search separately for each mammalian order . We then calculated how many different species were obtained from each order to calculate the total number of mammalian species likely to be amplified by this primer pair . Note that , when conducting the search without any taxonomic restrictions , these mammalian primers were also predicted to amplify amphibian and fish 16S rRNA genes . To estimate the total number of mammalian species for which the targeted locus has been sequenced and deposited in NCBI , we randomly selected one DNA sequence from each mammalian family and used BLAST searches to identify similar DNA sequences in the NCBI nucleotide database ( accessed on July 2015 ) . We filtered out any DNA sequence from the database that did not contain the primer sequences ( allowing for up to three mismatches ) . We then merged the results from the searches performed in each family and counted how many unique species were observed . These analyses provided us with the total number of mammalian species that should be amplified if the primers were truly universal . We also evaluated whether the 16S rRNA primers amplified sufficiently informative DNA sequences to support rigorous species identification ( i . e . , whether related species could be distinguished ) . First , we retrieved the mammalian DNA sequence alignment from the primerTree analysis and calculated the number of nucleotide differences ( including deletions ) between every pair of DNA sequences using the dist . dna program of the Ape package [38] . We then calculated the average proportion of nucleotide differences between species belonging to the same mammalian order and between species belonging to different orders . Second , we used the same approach to determine , for each mammalian order , how often two different species ( or genera ) have the exact same DNA sequence for the targeted region of the 16S rRNA gene . To interrogate the mammalian species composition of individual mosquito blood meals we amplified a 140 bp of the mammalian mt 16S rRNA gene using universal mammalian primers [36] modified to include a 5’-end tail complementary to the Illumina sequencing primers ( S1 Table ) . We also attempted to amplify a subset of 192 mosquitoes with universal avian primers ( [39] and S1 Fig ) using a pooled approach but failed to detect any bird DNA in these samples . To identify individual differences among human blood meals , we designed PCR primers to amplify 300 bp of the human mt hypervariable region I . We first aligned 795 whole mt genomes of individuals from Oceania [40] using MAFFT version 7 [41] to evaluate the extent of sequence variation across the mt genome hypervariable region I and then designed primers positioned in conserved flanking sequences with Primer3 [42] . As described above , we added a 5’ tail to each primer for sample barcoding and high-throughput sequencing ( S1 Table ) . For each sample and amplicon , we performed two rounds of PCR amplification to prepare products for Illumina sequencing ( Fig 1 ) . First , we performed a locus-specific amplification ( i . e . , targeting either the mammalian mt 16S rRNA or the human mt hypervariable region ) using the Promega GoTaq PCR kit protocol ( 50 μL reaction ) with 1μL of genomic DNA , 0 . 2mM of each dNTP , 1 . 25 units of GoTaq DNA polymerase , 4mM of magnesium and 0 . 2 μM primers . PCR amplification was carried out under the following conditions: 3 minutes at 94°C followed by 30 cycles at 94°C for 45 seconds , 50°C for 45 seconds , 72°C for 30 seconds and a final elongation step at 72°C for 3 minutes . We then purified these PCR products using the QIAquick 96 PCR purification kit protocol ( QIAGEN ) . Second , we incorporated Illumina adaptors , including unique 6-nucleotide sample identification barcode sequence through 10 additional PCR cycles , using barcoding primers complementary to the 5’-end tail incorporated during the first PCR amplification ( Fig 1 ) . For these reactions , we used the Promega GoTaq protocol as described above with 1uL of PCR product being added to each reaction . The same thermocycling conditions as described above were used but for an annealing temperature of 56°C . Predicted sizes for the mammalian mt 16S rRNA amplicons ranged from 265 to 343 bp; sizes for the human mt hypervariable region I amplicons ranged from 440 to 444 bp ( amplicon sizes include Illumina sequencing primers , unique barcode sequence and Illumina adaptors , Fig 1 ) . Finally , we pooled the barcoded amplification products from 96 individual mosquitoes and simultaneously sequenced them on an Illumina MiSeq instrument ( Sequences deposited in NCBI SRA: SRP062959 ) . We discarded from further analyses any read that did not carry the exact barcode and primer sequences . After recording the read origin using the barcode sequence , we removed the primer and barcode sequences to only keep the amplified DNA sequences . We discarded any resulting read smaller than 50 bp as these likely represent primer dimers . Since each amplified molecule was sequenced in both directions using paired-end reads , we merged each pair of sequencing reads using PANDAseq [43] ( Fig 1 ) keeping , at each position , the nucleotide with the highest sequencing quality . We then analyzed 16S RNA and human mtDNA sequences separately . Using all 43 , 743 , 363 16S rRNA sequences generated from the 442 mosquitoes , we identified all unique DNA sequences using Mothur [44] and recorded the number of reads carrying each unique DNA sequence . We removed any DNA sequence that was observed less than 10 times across all samples , as these likely resulted from sequencing errors . We compared the remaining unique DNA sequences against all DNA sequences present in the NCBI nucleotide database using blastn . For each DNA sequence , we recorded the best match ( es ) , only considering sequences with > 90% identity over the entire sequence length . We then retrieved the taxonomic information from each best-matched sequence using the ‘get_taxonomy’ function in PrimerTree . When an amplified sequence matched multiple species equally well , we recorded all species names associated with that sequence . We then summarized the blood meal of each mosquito by calculating the proportion of reads matching each species . As a small number of reads generated could reflect low level PCR contamination or an error in the sequence barcode identification , we only analyzed mosquito samples with at least 1 , 000 reads ( S2 Fig ) . For the same reason , we considered a mosquito as having fed on a single mammalian host if >90% of the sequencing reads aligned to the 16S rRNA of that species . Alternatively , if >10% of the sequencing reads aligned to a second species , we considered the mosquito to have fed on multiple mammalian hosts . For the human mt hypervariable region , we aligned the consensus reads to the human mitochondrial reference genome sequence ( NC_012920 . 1 ) using bowtie2 [45] and calculated , for each sample , the number of reads supporting each haplotype . Only haplotypes supported by more than 500 reads were considered to avoid incorporating sequencing or PCR errors ( i . e . , rare haplotypes that differed from an abundant haplotype by one nucleotide difference ) in the analyses . Finally , we reconstructed a phylogenetic tree with all identified human mt haplotypes using MEGA version 6 [46] .
We first conducted extensive in silico analyses to confirm that the primer pair selected [36] could amplify DNA sequences from a wide range of mammalian orders including Primates , Rodentia ( rodents ) , Artiodactyla ( even-toed ungulates ) , Carnivora ( carnivorans ) , Chiroptera ( bats ) , Cetacea ( cetaceans ) , Insectivora ( insectivores ) and Marsupials ( Table 1 ) . Overall , in silico analysis predicted that these primers should amplify 1 , 752 of the 1 , 779 mammalian species ( 98 . 5% ) sequenced at this locus and present in the NCBI nucleotide database ( Table 1 ) . Besides mammals , these primers were predicted to also amplify Actinopteri ( bony-fishes ) and Amphibia ( amphibians ) ( S3 Fig ) . In addition to amplifying a wide range of targets , our approach requires primers to amplify DNA sequences containing enough information to identify each species specifically . We tested this parameter by comparing the DNA sequences predicted to be amplified by this primer pair ( see Methods for details ) . Despite the short amplified DNA sequence ( ~140 bp ) , these primers enabled rigorous differentiation of most mammalian species as illustrated by the average proportion of nucleotide differences ( including deletions ) between sequences of species belonging to the same or different order ( S2 Table ) . For example , 27% of the nucleotide positions at this locus differ , on average , between one Carnivora and one Primate species and 17% of the nucleotides differ between the sequences of two Carnivora species . This high discriminating ability is also shown by the long branch lengths displayed by the phylogenetic tree reconstructed using these sequences ( S1C Fig ) . In fact , we found that one DNA sequence amplified by these primers typically matches a single genus and , in 86% of the cases , a single species ( Table 1 ) . We analyzed mosquitoes collected in five villages in the Madang Province in PNG: Dimer ( n = 45 ) , Wasab ( n = 81 ) , Kokofine ( n = 83 ) , Mirap ( n = 171 ) and Matukar ( n = 62 ) . These mosquitoes included several species of the Anopheles punctulatus group: An . punctulatus s . s . , An . koliensis , An . farauti 4 and An . farauti s . s . We characterized the blood meal composition of a total of 442 female Anopheles by amplifying the mammalian mt 16S rRNA genes from DNA extracted from these mosquitoes , pooling the PCR products of 96 samples after individual barcoding , and simultaneously sequencing the samples on an Illumina MiSeq instrument ( Fig 1 ) . We generated a total of 43 , 743 , 363 paired-end reads of 150 bp ( includes added primers ) . For 42 , 198 , 573 DNA sequences ( 96 . 5% ) , we were able to collapse the overlapping paired-ends ( Fig 1 ) and thus correct many sequencing errors . After combining the reads generated from all samples together , we identified 2 , 436 , 277 unique DNA sequences . We discarded from further analyses 2 , 404 , 684 unique DNA sequences that were carried by less than 10 reads across all samples as these likely represent DNA sequences with rare sequencing errors ( accounting for a total of 4 , 432 , 784 reads or 10 . 1% of the total number of reads generated ) . We then compared the remaining 31 , 593 unique DNA sequences , accounting for 39 , 310 , 579 reads ( 89 . 9% ) , to all DNA sequences deposited in the NCBI database . 28 , 999 of these DNA sequences ( representing 38 , 375 , 616 reads ) had > 90% nucleotide identity to at least one mammalian DNA sequence present in NCBI: 18 , 814 unique DNA sequences best matched a single mammalian species sequence while 10 , 185 unique DNA sequences matched equally well to multiple mammalian species sequences ( S3 Table ) . Overall , we generated an average of 82 , 528 reads per mosquito . The number of reads generated from each mosquito varied considerably ( S2 Fig ) as it depends on several factors including: the amount of starting template ( i . e . , quantity of mammalian DNA present in the mosquito ) , the amplification efficiency and uneven pooling or variations in sequencing output between MiSeq runs . For further analyses , we only considered mosquito samples with more than 1 , 000 reads . None of the 30 extraction controls ( i . e . , water samples that have been processed in parallel through DNA extraction , PCR and sequencing ) reached this cutoff illustrating the low level of cross-contamination or read mis-assignment due to errors in the barcode sequence ( if any ) . Overall , we analyzed mammalian DNA from 314 blood fed mosquitoes , including 258 out of the 337 mosquitoes characterized as fully-fed ( 76 . 6% ) and 56 out of the 86 mosquitoes visually-classified as partially-fed ( 65 . 1% ) . Only 5 out of the 19 mosquitoes visually classified as non-fed yielded mammalian 16S rRNA sequences: four yielded exclusively human 16S rRNA sequences , the last one a mix of human and pig sequences . These DNA sequences could indicate possible contamination either during field collection or in the laboratory , or detection of DNA from a previous , partially digested , blood meal . There was no statistical difference between fully-fed and partially-fed mosquitoes , however the number of sequencing reads generated for mosquitoes visually classified as fully-fed or partially-fed were significantly different from those classified as non-fed ( p<0 . 05 , Wilcoxon Rank-Sum test , S4 Fig ) . In total we successfully amplified and sequenced mammalian DNA from 319 Anopheles mosquitoes . We identified 201 Anopheles mosquitoes that carried human DNA , 111 carried pig DNA , 60 carried dog DNA and 5 carried mouse DNA ( Table 2; further details in S3 and S4 Tables ) . In addition to these expected hosts , we identified one mosquito that carried DNA from two different bat species: 7 . 2% of the reads matched perfectly Dobsonia moluccenis , a fruit bat commonly found in PNG , while 5 . 1% of the reads were most similar ( 94 . 4% identity ) to another megabat species , Dobsonia praedatrix , also endemic to PNG ( Table 2 ) . These bat DNA sequences were clearly distinct ( 8 nucleotide differences between them ) and unlikely to have been derived from sequencing errors , indicating that the mosquito fed on two different bats ( S5 Fig ) . Additionally , in another mosquito 13% of the total reads ( 7 , 599 reads ) were most similar to the common spotted cuscus ( Spilocuscus maculatus , 98% similarity ) , a marsupial found in the forests of PNG ( S6 Fig ) . Note that , consistent with our in silico analysis , we were not always able to identify the exact species that was fed upon . For example , we could not differentiate Canis lupus from Canis aureus ( S3 Table ) . Overall , these finding illustrate the unbiased nature of this sequencing approach to identify host species regardless of expectations for mosquito blood meal feeding ( as long as a closely related species has been sequenced ) . Out of 319 mosquitoes analyzed , 52 ( 16 . 3% ) showed clear evidence of having fed on more than one host species ( with >10% of the reads supporting the minor host ) : 44 mosquitoes carried DNA from two species and eight carried DNA from three species ( Fig 2 ) . Within each village , we identified three major mammalian hosts—humans , dogs and pigs—accounting for 37 to 100% of each mosquito blood meal . However , the proportion of mosquitoes that fed on each host varied within and between villages ( Fig 2 ) . For example , in Mirap , only 31 of the 127 Anopheles mosquitoes ( 24% ) fed on humans while 62 ( 49% ) fed on pigs , 11 fed on dogs ( 9% ) and 23 fed on two or three species ( 18% ) including one mosquito that fed on two bat species and one mosquito that fed on a common spotted cuscus ( Fig 2A ) . By contrast , in Kokofine , 52 out of the 62 mosquitoes fed on humans ( 84% ) while the remaining 10 mosquitoes fed on dogs ( n = 3 ) , pigs ( n = 3 ) or on multiple species ( n = 4 ) ( Fig 2B ) . The data for the three other villages are presented in Fig 2C–2E . Note that as host density information is not available for these villages , we were unable to test whether the observed differences in blood meal composition were caused by differences in mosquito feeding behavior among locations or species . Since we observed that 16 . 3% of the mosquitoes analyzed had fed on multiple mammalian hosts , we hypothesized that mosquitoes could also be feeding on multiple human individuals . We therefore investigated the number of different human DNA sequences present in 157 human-fed mosquitoes , using the same approach to sequence ~300 bp of the human mt hypervariable region . We generated an average of 26 , 721 sequencing reads of 250 bp for each sample and successfully amplified 102 of the 157 mosquitoes for the human mt hypervariable regions yielding a total of 20 different human mtDNA sequences ( S7 Fig ) . While a single DNA sequence was amplified from 78 . 5% ( n = 80 ) of the human-fed mosquitoes analyzed , 21 . 5% ( n = 22 ) mosquitoes carried two distinct DNA sequences ( S7 Fig ) . One sequence , identified in 14 of these potential mixed human blood meal , was always present at low frequency ( <8% of the reads ) and was actually more similar to a region of human chromosome 11 ( 98% similarity ) than to the mitochondrial genome sequence ( 91% ) . This DNA sequence likely resulted from the amplification of the nuclear insertion of the mitochondrial sequence ( numt , [47] ) and was excluded from further analyses . Nine mosquitoes , belonging to two species and collected in three locations , showed presence of two human mtDNA sequences ( S5 Table ) . For four of these mosquitoes , only one substitution ( out of the 300 bp amplified ) differentiated the two sequences and these could possibly be caused by a PCR error occurring at an early cycle . However , for the remaining five mosquitoes , 5–14 nucleotide substitutions differentiated the two sequences amplified and indicated that the mosquito successively fed on multiple individuals ( Fig 3 and S5 Table ) .
Previous studies have used microsatellites to compare the attractiveness of different individuals or group of individuals [22 , 24 , 25] , examine the blood feeding patterns of mosquitoes [20 , 53 , 54] or determine the effectiveness of insecticide treated bed nets [55–58] . DNA profiling with microsatellites allows for the identification of unique genetic profiles from human individuals fed on and can be a very powerful method to differentiate DNA from unrelated individuals . However , microsatellites can only detect the simultaneous presence of multiple individual DNAs ( typically two ) if their proportion in one sample is relatively similar . Otherwise , the signal from the less abundant DNA is typically obscured and not distinguishable from background noise . Rigorously identifying whether a disease vector feeds on a single or multiple individuals is however essential for disease control as vectors that feed on multiple individuals are more likely to rapidly spread the disease than those that only feed on a single individual . As an alternative to microsatellites , our approach relies on identifying unique human mitochondrial haplotypes carried by a mosquito by analyzing 300 bp of the mt hypervariable region I . We showed that at least five ( out of 102 mosquitoes analyzed ) carried human mitochondrial DNA sequences from more than one person . It is important to emphasize here that the number of mixed human blood meals is clearly underestimated as only maternal lineages can be detected by this approach: all offspring will carry the same DNA sequence as their mother and therefore it would not be possible to distinguish between siblings ( or cousins from mothers who are sisters ) . However , one could , at least partially , circumvent this limitation by including additional polymorphic nuclear loci in the assay and sequence them together with the mt hypervariable region locus ( and the 16S rRNA ) . Overall , our approach allows for a rapid evaluation of the number of maternal lineages a mosquito has fed on that can be added to the characterization of the blood meal at no additional costs , and could be used to determine if mosquitoes preferentially feed on some individuals and avoid other individuals . | Female mosquitoes require a blood meal to acquire the nutrients necessary for egg production . While feeding on host species , mosquitoes can transmit pathogens that cause several diseases including malaria , lymphatic filariasis and dengue . Understanding the mosquito host choice is important to better implement control strategies to reduce mosquito populations and therefore transmission of disease . Currently , the majority of methods for evaluating host species only test for the presence of pre-selected , expected hosts . Here , we describe an unbiased assay that combines amplification of any mammalian DNA with high-throughput sequencing to comprehensively characterize the composition of mosquito blood meals . We applied this approach to Anopheles mosquitoes collected in Papua New Guinea and observed that they fed on expected ( humans , dogs and pigs ) and unexpected hosts ( mice , bats , marsupials ) . In addition , we show that 16 . 3% of mosquitoes fed on multiple hosts , from the same or different species . Overall , this approach enables unbiased characterization of mosquito blood meals and can be easily applied to significantly improve our understanding of the feeding behavior of any disease-transmitting insect . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"sequencing",
"techniques",
"invertebrates",
"medicine",
"and",
"health",
"sciences",
"body",
"fluids",
"animals",
"molecular",
"biology",
"techniques",
"insect",
"vectors",
"cellular",
"structures",
"and",
"organelles",
"research",
"and",
"analysis",
"methods",
"sequen... | 2016 | Unbiased Characterization of Anopheles Mosquito Blood Meals by Targeted High-Throughput Sequencing |
The present pilot study investigating the minimum dose for short-course single and double-dose treatment of kala-azar with an apparently new liposomal formulation of amphotericin B , Fungisome , led to identification of immunological components for early detection of success and/or failure to cure . Patients were treated with 5 , 7 . 5 ( single-dose ) and 10 mg/kg body weight ( 5 mg/kg double-dose ) of Fungisome . Immunological investigations involving plasma cytokines and antigen-specific lymphoproliferation and cytokine responses from PBMCs were carried out before , 1 week after Fungisome treatment , at the time of relapse , and again after conventional amphotericin B treatment . At 1-month follow-up all the patients showed 100% initial cure . However , total doses of 5 , 7 . 5 and 10 mg/kg Fungisome showed 60% , 50% and 90% cure , respectively , at 6-months posttreatment . Patients successfully cured demonstrated downregulation of IL-12 and IL-10 in plasma , and two-fold or more elevation of IFN-γ , IL-12 and TNF , and significant down-regulation of IL-10 and TGF-β in culture supernatants 1-week posttreatment irrespective of drug-dose . A differential immune profile , involving insignificant decline in IL-10 and IL-12 in plasma and negligible elevation of IFN-γ , IL-12 and TNF , and persistence of IL-10 , despite decline in TGF-β in culture supernatants , in apparently cured individuals , corresponded with relapse within 6-months of treatment . Immunological investigations revealed significant curative and non-curative immunomodulation 1-week posttreatment , correlating with successful cure and relapse , respectively . Although immune-correlation was dose-independent , almost consistent curative response in patients treated with the highest dose 10 mg/kg reflected a definitive impact of the higher-dose on the immune response . Clinical Trials Registry - India ( CTRI ) CTRI/2009/091/000764
Visceral leishmaniasis ( VL ) or kala-azar , caused by members of Leishmania donovani complex , is the most severe form of leishmaniasis involving uncontrolled parasitization of liver , spleen and bone marrow . If left untreated VL is generally fatal [1] . Although the disease is endemic in 62 countries , 90% of the estimated 500 , 000 new cases occur in the rural areas of India , Nepal , Bangladesh , Sudan and Brazil [2] . Pentavalent antimonial is the standard therapy in the endemic regions . However , in north Bihar , India , 34–65% of the VL patients are now refractory to this treatment [3] , [4] . Amphotericin B ( AmB ) at 15–20 mg/kg body weight renders high cure rate of ∼100% in the antimonial unresponsive regions of Bihar [5] . However , serious nephrotoxicity , intravenous administration with constant monitoring , and prolonged hospitalization limit its use [6] . Furthermore , AmB stimulates production of proinflammatory cytokines including TNF-α [7] , [8] , [9] which although contributes to the antimicrobial activity of AmB also increases its infusion related toxicities ( fever , chill , and shivering ) [10] . With few satisfactory alternative therapies available , treatment of VL is a challenging task . Liposomal amphotericin B , Ambisome , has the highest therapeutic index of current antileishmanial drugs and reduces the toxicities observed with free AmB , ensuring administration of higher doses leading to short treatment courses with reduced side effects [11] . World Health Organization ( WHO ) recommends a total dose of 10–21 mg/kg of Ambisome for the treatment of VL [12] , [13] , [14] . However , even with the preferential pricing for developing countries [15] , Ambisome is almost 3-fold the price of conventional AmB . To bring down the cost , recent clinical studies to identify minimum effective total dose showed that a single infusion of 5 or 7 . 5 mg/kg would only leave a fraction ( 10–20% ) of patients needing further treatment [16] , [17] , [18] . Under these circumstances , an early prediction of lack of response in this fraction of patients who would require more drug for treatment would be desirable . However , there is no means to predict successful cure or failure of drug treatment in kala-azar . Active VL is characterized by immune suppression with weak T cell responses [19] . In vivo cytokine profile from the serum and at the mRNA level from bone marrow , lymphoid tissues , and splenic aspirate of patients showed mixed Th1/Th2 cytokine response during active disease [20] , [21] , [22] , [23] , [24] . In contrast , studies with leishmanial Ag-stimulated PBMC showed Th2-type bias at active VL with decreased or absent IL-2 and IFN-γ production and upregulation of both IL-10 and TGF-β [25] , [26] , [27] , [28] , [29] . With the onset of chemotherapy and cure with sodium antimony gluconate ( SAG ) and AmB , upregulation of IFN-γ , and IL-12 synchronized with a decline in IL-10 and TGF-β levels [25] , [26] , [27] , [28] , [29] . In addition to the direct antileishmanial activity , both AmB and SAG showed in vitro immunomodulatory activities on PBMCs resulting in decline of IL-10 , with AmB downregulating TGF-β also [29] . So a short-course therapy with liposomal formulation of AmB that maintains the efficacy of AmB and is non-toxic would be the ideal first line treatment for kala-azar . A liposomal AmB preparation , developed in India and commercially available as Fungisome , is safe and effective for treatment of VL [2] . As for Ambisome , Fungisome is also an intravenous infusion , and a total dose of 15–21 mg/kg showed an efficacy of 90 . 9–100% against responsive and unresponsive cases of VL [30] . To initiate short-course therapy , we investigated whether a relatively higher dose at a single- ( 5 , and 7 . 5 mg/kg ) or a double-dose ( 5 mg/kg ) ( total 10 mg/kg ) of Fungisome could be safe and effective for therapy of kala-azar . In addition , we were interested to investigate whether a similar modulation of cytokine profile as observed with AmB treatment could be obtained with Fungisome as early as 1 week after treatment to identify if possible immunological components for early detection of successful treatment and/or failure to cure .
This short-course pilot study was performed between 2006 and 2008 at the School of Tropical Medicine ( Kolkata , India ) . Patients were mainly from endemic regions of eastern India diagnosed with active VL . Based on the inclusion exclusion criteria ( Protocol S1 ) patients of all ages and both sexes were potentially eligible if they presented the clinical symptoms of prolonged fever , hepatosplenomegaly , and were confirmed to be VL by K39 strip test and detection of Leishmania parasites in the splenic or bone marrow aspirate . Three groups of VL patients were classified based on their mode of treatment . Group A ( n = 10 ) patients were categorized for treatment ( Lifecare Innovations , India ) with 5 mg/kg body weight Fungisome single-dose . Group B ( n = 10 ) patients were treated with Fungisome 7 . 5 mg/kg single-dose , and Group C ( n = 10 ) were treated with Fungisome 5 mg/kg×2 doses on subsequent days . 100 ml Fungisome was dissolved in 100 ml of normal saline and infused i . v . over 2 hours . Patients were monitored during infusion and further for 24 hours for any adverse effects , and were discharged 1-week after treatment . Some of the patients who suffered relapse within 6-months of treatment were further treated with conventional AmB ( Sarabhai Piramal Pharmaceuticals , India ) ( total dose 20 mg/kg body weight ) by i . v . drip in dextrose solution on alternate days . Longitudinal heparinized blood samples from the patients were collected at different time points . Biochemical and hematological characteristics are shown in Table S1 . The study was approved by the Institutional Review Board and Ethical Committee , Calcutta School of Tropical Medicine , Kolkata . Written informed consent was obtained from each patient enrolled in the study . A copy of the patient consent form was submitted to the Ethical Committee . Leishmanial antigen ( LAg ) was prepared from promastigotes of L . donovani ( MHOM/IN/1983/AG83 ) as described earlier [31] . Briefly , stationary phase promastigotes harvested after the third or fourth passages were washed four times in ice-cold 20 mM phosphate-buffered saline ( PBS , pH 7 . 2 ) and suspended in 5 mM cold Tris-HCl buffer ( pH 7 . 6 ) . The suspension was vortexed six times for 2 min each with a 10-min interval on ice and then centrifuged at 2310×g for 10 min . The crude ghost membrane pellet obtained was resuspended in the same buffer and sonicated on ice three times for 1 min in an ultrasonicator . The suspension was centrifuged at 5190×g for 30 min and the supernatant containing the LAg was harvested and stored at −70°C . The amount of protein obtained from 1-g cell pellet was ∼14 mg , as assayed by Lowry et al [32] . Heparinized blood samples for isolation of PBMCs from VL patients , and healthy volunteers ( n = 5 ) from the Indian Institute of Chemical Biology ( IICB ) , Kolkata , India ( non endemic for VL ) , were obtained after written informed consent and was approved by The Ethical Committee on Human Subjects , IICB . Briefly , PBMCs isolated from the blood samples by density sedimentation on Histopaque-1077 ( Sigma-Aldrich ) ( 400×g , 30 min at RT ) , were washed and resuspended in RPMI 1640 supplemented with 10% FCS ( Sigma , St Louis , MO ) , 2 mM L-glutamine , penicillin ( 100 U/ml ) , and streptomycin ( 100 µg/ml ) . PBMCs ( 1×106/ml ) of VL patients were cultured in triplicates in 96-well flat-bottom tissue culture plates ( Nunc ) and stimulated with and without LAg ( 12 . 5 µg/ml ) for 4 days at 37°C in 95% humidified air with 5% CO2 . 1 µCi of [3H] Thymidine ( sp . act . 5 Ci/mM; Amersham Biosciences ) was added to the wells and cultured for another 18–24 hours . Thymidine uptake was measured in a β scintillation counter ( Beckman Instruments , Fullerton , CA ) . Plasma from VL patients were collected at different time points and stored at −20°C for cytokine analysis . PBMCs of these patients were cultured in the presence of LAg and media as described above . After 96 hours supernatants were collected for cytokine analysis . To study the direct effect of AmB and Fungisome to modulate cytokine production , PBMCs ( 1×106/ml ) from healthy controls were incubated with various non-toxic concentrations of AmB ( 0 , 0 . 031 , 0 . 062 , 0 . 125 , 0 . 25 , 0 . 5 µg/ml ) and Fungisome ( 0 , 0 . 031 , 0 . 062 , 0 . 125 , 0 . 25 , 0 . 5 , 1 µg/ml ) with or without LPS ( 1 µg/ml ) for 48 hours at 37°C in 95% humidified air with 5% CO2 . Culture supernatants were collected and stored at −70°C until use . IFN-γ , IL-12 ( p40 ) , TNF and IL-10 ( BD OptEIA ELISA kit; BD Biosciences ) were measured according to the manufacturers' instructions . For total TGF-β measurement , culture supernatants were acidified to activate the latent TGF-β by adding 1 N HCl for 10 min and neutralized with 1 . 2 N NaOH in 0 . 5 M HEPES . Plasma samples were activated with 2 . 5 N acetic acid in 10 M urea for 10 min and neutralized with 2 . 7 N NaOH in 1 M HEPES . TGF-β was captured with monoclonal anti-TGF-β1 , MAB240 , and detected with biotinylated polyclonal anti-TGF-β1 , BAF240 ( R&D Systems ) . The standard curve was prepared using rTGF-β1 ( R&D Systems ) suspended in culture medium . The color reaction was performed using avidin-HRP and tetramethylbenzidine and read at OD 450 nm . Statistical analysis was performed using the nonparametric Wilcoxon matched pairs signed rank test for paired samples and the Mann-Whitney U test for unpaired samples . One-way ANOVA followed by Tukey's multiple comparison test was performed to assess the differences among various groups ( GraphPad Software , Inc . , San Diego , CA . ) . P<0 . 05 was considered to be significant .
Three groups of patients were categorized based on the mode of treatment with Fungisome; Group A ( 5 mg/kg body weight single dose , n = 10 ) , Group B ( 7 . 5 mg/kg single dose , n = 10 ) and Group C ( 5 mg/kg×2 doses on subsequent days , n = 10 ) . Figure 1 shows the flow of participants through the dose finding study . Within 3 months of study initiation , 2 out of 5 patients ( 40% ) in Group A returned with relapse . So this dose was considered suboptimal and the protocol was amended to continue treatment only in Group B and C . At 1-month follow up , initial response to treatment was noted and all the patients were considered clinically cured . Within 6-months follow-up , 5 out of 10 patients ( 50% ) in Group B and 1 out of 10 patients ( 10% ) in Group C ( Table 1 ) experienced relapse . Therefore cure rate 6-months posttreatment was 60% in Group A , 50% in Group B and 90% in Group C . The other 5 enrolled patients in Group A , who did not receive treatment with Fungisome , and the patients who suffered relapse in all the three groups , were treated with conventional AmB ( 20 mg/kg ) and were cured . All the three doses of Fungisome were remarkably well tolerated ( Table S2 ) . For immunological studies two groups of patients were categorized , responders who were successfully cured 6-months posttreatment and non-responders who suffered relapse within 6-months of treatment . The mean plasma cytokine levels of both Th1-type , IFN-γ , IL-12 , TNF , and Th2 type immunosuppressive cytokines IL-10 and TGF-β were elevated during active disease ( Table 2 ) . 1-week after treatment , there was significant downregulation of IL-12 and IL-10 ( P<0 . 005 ) in the responders compared to untreated condition , however , there was no significant change in the levels of IFN-γ , TNF and TGF-β ( Table 2 ) . The fall in IL-10 in the responders was also significant ( P<0 . 05 ) when compared with non-responders 1-week posttreatment . Difference in the cytokine levels of non-responders was significant neither at 1-week posttreatment nor at relapse compared to untreated patients . However , after cure with AmB there was significant downregulation again restricted to IL-12 and IL-10 ( P<0 . 05 ) compared to active diseased and relapsed patients ( Table 2 ) . Although there was a mixed Th1-Th2 type cytokine profile in the plasma of VL patients during active disease and after treatment with Fungisome and AmB , a significant fall in IL-10 and IL-12 in both corresponded with cure . Consistent with prior observations [19] , [29] , VL patients at active disease showed impaired lymphoproliferation in response to LAg-stimulation ( Table 3 , Figure S1 ) . Remarkably , 1-week after Fungisome treatment there was significant enhancement in lymphoproliferation in the responders of all the three groups of patients ( P<0 . 0001 ) ( Table 3 ) especially in Group B and Group C ( P<0 . 05 ) ( Figure S1 ) . However , non-responders showed no or low lymphoproliferation 1-week posttreatment and at the time of relapse . Subsequent treatment with AmB ( 20 mg/kg ) resulted in significantly ( P<0 . 05 ) increased lymphoproliferation . LAg-stimulated cytokine production during disease demonstrated low levels of IFN-γ , IL-12 and TNF ( Table 3 ) measured over background levels . 1 week after Fungisome treatment there were two or more fold increase in antigen-stimulated IFN-γ production in the responders of all the groups ( P<0 . 05 ) ( Table 3 ) , more consistent in Group C ( Figure S1 ) when compared with untreated patients and non-responders ( Table 3 ) . In contrast , there was no such increase in patients who suffered relapse 1-week posttreatment . Although there was significant upregulation of IL-12 and TNF in the responders of all the groups who were cured 6-months posttreatment compared to the untreated ( P≤0 . 0006 ) and non-responders ( P≤0 . 005 ) ( Table 3 ) there was low or insignificant increase in these cytokines in non-responders following treatment . Separate analysis indicated that immunological responses were dose-independent ( Figure S1 ) . Therefore , two or more fold upregulation of IFN-γ , IL-12 and TNF , 1 week after Fungisome treatment , could predict successful cure at 6-months posttreatment irrespective of the drug dose . At the time of relapse the patients had moderate levels of LAg-specific IFN-γ and IL-12 and low levels of TNF ( Table 3 ) , which increased significantly ( P<0 . 05 ) after completion of AmB treatment . Though IL-10 and TGF-β are not considered typical Th2 cytokines both have been implicated as immunosuppressive factors in human VL . Prior to treatment there were high levels of LAg-specific IL-10 and TGF-β , calculated over background levels , in most of the patients of all the three groups ( Table 3 , Figure S2 ) . There was significant ( P<0 . 005 ) downregulation of IL-10 in the responders of all the three groups 1-week after Fungisome treatment compared to before treatment ( Table 3 ) and non-responders ( P<0 . 05 ) . However , there was no significant downregulation of IL-10 in the non-responders who suffered relapse . Similarly , TGF-β levels were high prior to treatment in all the patients . 1 week following treatment there was decline in TGF-β in both responders and non responders , although the fall was significant ( P<0 . 005 ) only in the responders of all the groups . Persisting levels of IL-10 and TGF-β at the time of relapse showed significant decline following treatment with AmB . To compare the in vitro immunomodulatory effects of Fungisome with conventional AmB [9] , PBMCs from healthy individuals were cultured for 48 hours in the presence or absence of LPS and various nontoxic concentrations of these drugs ( Figure 2 ) . Although AmB significantly ( P<0 . 05 ) upregulated LPS-stimulated TNF production , Fungisome did not increase the TNF level ( Figure 2A ) . Unstimulated PBMCs led to the production of high levels of TGF-β that was enhanced further in the presence of LPS . Both AmB and Fungisome were instrumental in significantly ( P<0 . 01 ) downregulating inherent as well as LPS-induced TGF-β production at concentrations 0 . 25 µg/ml and above ( Figure 2B ) and LPS-stimulated IL-10 production ( P<0 . 05 ) in a dose-dependent manner ( Figure 2C ) . The effect was more significant ( P<0 . 05 ) for Fungisome as the highest dose of the drug could be extended to 1 µg/ml .
The present pilot study evaluated the efficacy for short course single- ( 5 , and 7 . 5 mg/kg ) and double-dose ( 5×2 mg/kg ) Fungisome treatment of kala-azar patients . Interestingly , a differential immune profile of the patients with modulation of in vivo and antigen-stimulated cytokine responses was observed before and 1 week after drug treatment which correlated with successful cure or relapse 6-months posttreatment . Commercially available lipid formulations of AmB have largely remedied the drawbacks of conventional AmB enabling safe and short-duration administration of higher doses for cost-effective treatment . A cure rate of 90 . 9–100% was reported for Fungisome at 3 mg/kg for 5–7 days ( total dose 14–21 mg/kg ) [30] . To establish the lower limit of efficacy and reduce treatment duration we initiated Fungisome treatment at 5 and 7 . 5 mg/kg single-dose , and 5 mg/kg double-doses ( total 10 mg/kg ) on consecutive days . All the three doses of Fungisome showed an initial clinical cure of 100% at 1-month follow-up . However , total doses of 5 , 7 . 5 and 10 mg/kg of Fungisome , showed 60% , 50% and 90% cure rate at 6-months posttreatment . There are few reports on comparative short-course dose-finding studies with liposomal AmB and these show varying results [33] , [34] , [35] , [36] , warranting further testing of single and double-dose regimens of liposomal AmB . It is well established that infection with L . donovani and its control are modulated by a range of T cell responses and cytokine network . We therefore investigated the immune environment of the patients at 1-week posttherapy . The results interestingly throw light on the possible mechanisms of successful cure and relapse , despite an apparent cure . Investigation of plasma cytokines revealed a mixed Th1–Th2 type profile in the VL patients during active disease as observed earlier [23] , [24] , [37] , [38] . Interestingly , 1-week following treatment resulted in a significant fall in the levels of IL-10 and IL-12 corresponding with cure . Similar fall was also observed in the cured individuals treated with conventional AmB . High serum levels of IFN-γ during active VL do not reconcile with large parasite burdens observed in this disease . Lack of IFN-γ activity may be related to the simultaneous presence of elevated levels of IL-10 and TGF-β , the macrophage deactivating cytokines in human leishmaniasis [39] . The observed fall in IL-10 levels 1-week following treatment is coincident to the control of parasite growth lending support to an important role of this cytokine in human VL . Investigations on the antigen-specific cytokine production by PBMCs showed an enhanced Th1-type response with upregulated IFN-γ , IL-12 and TNF and reduced IL-10 and TGF-β production 1 week after treatment in VL patients who were cured at 6-months posttreatment irrespective of the drug-dose . IL-12 helps in the expansion of IFN-γ that synergistically acts with TNF-α to activate macrophages to kill Leishmania parasites through release of NO [40] , [41] . IL-10 and TGF-β neutralize the effects of IFN-γ [37] , [39] , [42] and favor survival of Leishmania by inhibiting NO production by the macrophages [43] thus helping in disease progression [39] , [44] , [45] , [46] , [47] , [48] . Patients who suffered relapse , showed neither significant elevation of Th1 cytokines nor downregulation of IL-10 , at this time point , although TGF-β level was reduced . This differential immune response 1 week posttreatment between the responder and non-responder patients further substantiates the significance of upregulation of Th1 cytokines and simultaneous downregulation of IL-10 for the success of therapy . Further , our studies show that this polarization occurs as early as 1 week of treatment and thus has the potential to serve as markers to predict therapeutic success . Elevation of antigen-specific IL-10 and TGF-β along with the Th1 cytokines at the time of relapse substantiates the significance of immunosuppressive activities of IL-10 and TGF-β . Conventional AmB treatment ( 20 mg/kg ) of relapsed patients generated a dominant Th1-type response as observed earlier [29] . Since cure is a combinatorial effect of drug and immune status of the host in VL , the rationale approach towards antileishmanial chemotherapy would be to potentiate the immune functioning of the host [49] in addition to parasite killing . Targeting of Th1 cell mechanism increased the efficacy of AmB and permitted lower doses to be used with comparable activities [50] , [51] . In addition , our previous study revealed that AmB can downregulate IL-10 and TGF-β in the LPS-stimulated PBMCs of healthy individuals [29] . Interestingly , in the present study similar results were maintained by Fungisome . Simultaneous downregulation of IL-10 and TGF-β is important as their combined in vivo blockade led to an apparent sterile immunity in mice infected with Leishmania parasites [48] . TNF , responsible for infusion-related toxicities of AmB ( 10 ) was increased only in the presence of AmB . No such increase was observed with Fungisome corroborating earlier reports [52] of reduced TNF production by lipid carriers of AmB . The mechanism of the observed immunoenhancement in Fungisome treated patients was due to the combined effect of the drug's ( AmB ) parasite killing effect together with unexpectedly rapid regulation of cytokine responses . A significant downregulation of IL-12 and IL-10 in the plasma , elevation of IFN-γ , IL-12 and TNF , and downregulation of IL-10 and TGF-β in PBMCs , in patients' 1-week after treatment corresponded with successful cure irrespective of drug dose . Negligible decline in plasma IL-12 and IL-10 levels , negligible elevation of IFN-γ , IL-12 and TNF and persistence of IL-10 , despite decline in TGF-β in culture supernatants of apparently cured individuals , predicted disease relapse within 6-months posttreatment . Although the immune modulation was dose-independent , low cure rates in Group A and B corresponded with fall in TGF-β levels without much effect on IL-10 expression in some of the patients who eventually relapsed . In contrast , treatment with the highest dose in Group C simultaneously downregulated IL-10 and TGF-β , and upregulated IFN-γ , IL-12 and TNF . This dual effect might be responsible for the successful cure , with only one relapse , in this Group . Fungisome maintained increased immunomodulatory activity of AmB at doses that are toxic with the conventional AmB . To extend this clinically appealing , novel approach , studies should be initiated with higher total doses of 10 mg/kg ( 5 mg/kg double-dose ) and 15 mg/kg ( 7 . 5 mg/kg double-dose ) of Fungisome with larger number of patients to better define the role of this new liposomal AmB . | Visceral leishmaniasis ( VL ) is a potentially fatal disease without treatment , characterized by prolonged fever , enlargement of spleen and liver , anaemia and weight loss . Treatment for VL is difficult , as it requires prolonged and painful application of toxic drugs with adverse side effects . It is therefore important to develop alternative satisfactory therapies for VL . Herein , we report the efficacy of a new liposomal formulation of amphotericin-B , Fungisome , and the immunological changes that take place 1-week after treatment . Patients treated with 5 and 7 . 5 mg/kg ( single-dose ) and 10 mg/kg ( 5 mg/kg double-dose ) of Fungisome showed 60% , 50% and 90% successful cure at 6-month posttreatment , respectively . Successfully cured patients showed reduced IL-12 and IL-10 levels in the plasma and two-fold or more increase in Th1 type-cytokines IFN-γ , IL-12 and TNF , and down-regulation of immunosuppressive factors IL-10 and TGF-β in the culture supernatants , 1-week after treatment independent of drug-dose . Insignificant decrease of plasma IL-12 and IL-10 , negligible increase of Th1-cytokines , and persistence of IL-10 , despite decrease in TGF-β in culture supernatants , correlated with relapse within 6-months of treatment . These interesting results pave the way for further testing of this drug as a new alternative in the chemotherapy of leishmaniasis . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"infectious",
"diseases",
"microbiology/immunity",
"to",
"infections",
"immunology/immunomodulation",
"infectious",
"diseases/neglected",
"tropical",
"diseases",
"immunology/immune",
"response",
"microbiology/innate",
"immunity",
"immunology/innate",
"immunity",
"immunology",
"infe... | 2010 | A Curative Immune Profile One Week after Treatment of Indian Kala-Azar Patients Predicts Success with a Short-Course Liposomal Amphotericin B Therapy |
During millions of years of coevolution with their hosts , cytomegaloviruses ( CMVs ) have succeeded in adapting to overcome host-specific immune defenses , including the protein kinase R ( PKR ) pathway . Consequently , these adaptations may also contribute to the inability of CMVs to cross species barriers . Here , we provide evidence that the evolutionary arms race between the antiviral factor PKR and its CMV antagonist TRS1 has led to extensive differences in the species-specificity of primate CMV TRS1 proteins . Moreover , we identify a single residue in human PKR that when mutated to the amino acid present in African green monkey ( Agm ) PKR ( F489S ) is sufficient to confer resistance to HCMVTRS1 . Notably , this precise molecular determinant of PKR resistance has evolved under strong positive selection among primate PKR alleles and is positioned within the αG helix , which mediates the direct interaction of PKR with its substrate eIF2α . Remarkably , this same residue also impacts sensitivity to K3L , a poxvirus-encoded pseudosubstrate that structurally mimics eIF2α . Unlike K3L , TRS1 has no homology to eIF2α , suggesting that unrelated viral genes have convergently evolved to target this critical region of PKR . Despite its functional importance , the αG helix exhibits extraordinary plasticity , enabling adaptations that allow PKR to evade diverse viral antagonists while still maintaining its critical interaction with eIF2α .
HCMV is a ubiquitous virus that persists for the lifespan of the infected host , highlighting its ability to evade host defenses [1] . While most infections are asymptomatic , HCMV causes life-threatening diseases in immunocompromised patients and is the most frequent congenital viral infection in developed countries , leading to permanent neurological deficits in thousands of newborns each year [2] . Despite its success in spreading throughout the human population , HCMV is unable to cross species barriers . Genomic analyses have demonstrated that CMVs have been co-speciating with their hosts for ~80 million years [3 , 4] . Through this process , each CMV has specifically adapted to its cognate host and in doing so , diverged from closely related CMV species . Among the many factors that may contribute to cross-species barriers to infection , cell-intrinsic immune factors likely play a central role because the selective pressure imposed by viral antagonists has driven their rapid evolution . Support for this arms race paradigm comes from computational and functional studies that demonstrate ongoing reciprocal innovation by host and viral factors at host:virus interfaces [5 , 6] . The millions of years of shared evolutionary history between CMVs and their hosts provide an invaluable model for investigating the consequences of host-virus arms races . Protein Kinase R ( PKR ) is a broadly acting restriction factor that phosphorylates the translation initiation factor eIF2α in response to cytoplasmic double-stranded RNA ( dsRNA ) , resulting in a block to translation initiation and viral replication [7] . The importance of PKR in the host cell’s anti-viral arsenal is highlighted by the presence of PKR antagonists in many virus families [8 , 9] . Furthermore , deletion of PKR antagonists renders many viruses replication deficient [10–15] , demonstrating that PKR poses a strong molecular barrier to viral replication . To overcome the onslaught of diverse viral antagonists , PKR has had to continually adapt while still being constrained by the need to maintain its critical interactions with dsRNA and eIF2α . Consistent with this perspective , evolutionary analyses have identified dramatic episodes of positive selection in PKR during primate evolution [16 , 17] . Thus , we first leveraged the long co-evolutionary history of CMVs and their hosts to investigate how the rapid evolution of PKR has impacted the evolution of the CMV-encoded PKR antagonist TRS1 .
To determine whether the evolutionary divergence of PKR in primates has affected the ability of CMVs to antagonize PKR , we used a recombinant VacV system to readily test TRS1 alleles from several primate CMV species . The VacVs used in these studies were engineered to express lacZ , allowing us to measure β-gal activity as a proxy for viral replication as there is strong correlation between β-gal activity and viral titers ( S1 Fig ) . This recombinant VacV system takes advantage of the fact that wild type VacV ( WT VacV ) replicates well in a broad range of primate cells , including human ( HeLa ) and Agm ( BSC40 ) cells , while deletion of the PKR antagonist E3L ( VacVΔE3L ) markedly reduces replication [14] ( Fig 1A ) . PKR activity is responsible for the replication blockade in HeLa cells , as knocking out PKR by CRISPR/Cas9 gene editing ( HeLa PKR KO ) completely rescued VacVΔE3L replication . To evaluate whether the CMV TRS1 proteins can antagonize human or Agm PKR , we recombined four CMV TRS1 genes from species that infect hominoids , Old World monkeys , and New World monkeys ( HCMV , African green monkey CMV ( AgmCMV ) , Rhesus CMV ( RhCMV ) ) , and Squirrel monkey CMV ( SmCMV ) ) into VacVΔE3L and evaluated their ability to rescue replication in human and Agm cells . All viruses replicated well in the HeLa-PKR KO cells ( Fig 1A ) and expressed comparable levels of the TRS1 proteins ( although in this particular experiment SmCMVTRS1 expression was slightly higher ) ( Fig 1B ) . However , we observed species-specific differences in viral replication in the human and Agm cell lines . Consistent with previous findings , VacVΔE3L+HCMVTRS1 replicated well in human but not Agm cells [18] . Conversely , AgmCMVTRS1 and RhCMVTRS1 rescued VacVΔE3L in Agm but not human cells . Surprisingly , SmCMVTRS1 rescued replication in both human and Agm cells , indicating that it is more broadly acting . These results suggest that evolutionary pressures have had a substantial impact on the CMV TRS1 genes , resulting in species-specific differences in their ability to block host defenses necessary to rescue VacVΔE3L replication . We hypothesized that the differences in the ability of TRS1 genes to rescue VacVΔE3L in human and Agm cells are due to differing abilities to antagonize human vs . Agm PKR . To test this , we utilized an assay in which expression of PKR inhibits translation of a co-transfected secreted embryonic alkaline phosphatase ( SEAP ) reporter gene [17 , 19] ( Fig 2A ) . Co-transfection of a functional PKR antagonist reverses , at least in part , the inhibitory effect of PKR , leading to an increase in reporter gene expression . We co-transfected HeLa PKR KO cells with plasmids expressing SEAP along with HuPKR or with a control plasmid and a panel of TRS1 antagonists or a vector control . In the absence of any antagonist , transfection of the HuPKR expression plasmid reduced SEAP expression ( Fig 2B , bars 1 vs 6 ) . Co-transfection of HCMVTRS1 or SmCMVTRS1 each lessened the inhibitory effect of HuPKR ( bars 7 and 10 ) , while co-transfection of AgmCMVTRS1 or RhCMVTRS1 had little effect ( bars 8 and 9 ) . In contrast , in cells transfected with Agm PKR , AgmCMVTRS1 rescued reporter activity relative to the vector control while HCMVTRS1 did not ( Fig 2C , S2A and S2B Fig ) , consistent with what was observed with VacVΔE3L rescue in Agm cells ( Fig 1A ) . These results substantiate the co-evolutionary history of CMVs with their hosts , during which specific adaptations to their cognate PKR limited hominoid and Old World monkey CMVs ability to restrict more distant PKR alleles . These functional differences are likely explained by species-specific changes in the binding interface of PKR and the CMV TRS1 proteins . However , the location of this interface on PKR is unknown . As an alternative to blind approaches like alanine scanning and random mutagenesis , we leveraged the power of the species-specific differences in TRS1 activity to precisely map the PKR:TRS1 interface . We generated chimeras between the susceptible HuPKR allele and the resistant AgmPKR allele . The 98 amino acids that differ between HuPKR and AgmPKR are scattered throughout the protein , as are the sites that were previously found to be evolving under positive selection in primates [16] , making it difficult to predict which region ( s ) is likely to be responsible for their differential sensitivity to HCMVTRS1 . Therefore , we first made chimeras with a break point within the linker region near the middle of PKR and found that the species-specificity of HCMVTRS1 and AgmCMVTRS1 mapped to the C-terminal half of PKR ( S2A Fig ) . Further subdivision identified a small region ( codons 475–520 , designated region D2 ) within Agm PKR which when introduced into HuPKR was sufficient to confer resistance to HCMVTRS1 ( S2B Fig ) . Six amino acids within this region differ between HuPKR and AgmPKR . To determine whether any of these differences alone is necessary for sensitivity to HCMVTRS1 , we generated point mutants at each of these sites by introducing the amino acid present in AgmPKR into HuPKR . Strikingly , mutating position 489 of HuPKR from phenylalanine to serine ( F489S ) was sufficient to confer resistance to HCMVTRS1 , while the other five point mutants had no effect ( Fig 2C ) . Thus , this system allowed us to rapidly identify position 489 as a critical species-specific determinant of sensitivity to HCMVTRS1 . We next wished to test the impact of this mutation in the context of a complete viral infection by challenging viruses with heterologous PKRs . Unfortunately , this approach has been difficult to establish because PKR overexpression inhibits cell growth . To circumvent this problem , we stably transduced PKR knockout cells with heterologous PKR genes ( HuPKR , AgmPKR , HuPKR F489S , or the empty vector ) under the control of a doxycycline-inducible promoter . These cells were then used to assess the replication profiles of our recombinant VacVs . All of the viruses replicated well in the control empty vector cell line regardless of doxycycline induction ( Fig 3A ) , and TRS1 proteins were expressed to similar levels in these cells ( Fig 3B ) . Upon induction , the HuPKR and HuPKR F489S cell lines expressed PKR to levels comparable to that observed in wild-type HeLa cells ( Fig 3C ) . Although we could not assess expression of AgmPKR as it does not cross-react with the PKR antibody , the fact that the AgmPKR cell line restricted VacVΔE3L replication after induction of PKR by addition of doxycycline strongly suggests that AgmPKR was expressed in these cells . As expected , WT VacV was able to overcome PKR antiviral activity in each of these cell lines , while VacVΔE3L did not replicate well in any ( Fig 3A ) . In the HuPKR- and AgmPKR-expressing cell lines , the replication profiles of the panel of TRS1-expressing viruses mirrored those observed in human ( HeLa ) and Agm ( BSC40 ) cell lines ( Figs 1A and 3A ) . Importantly , the cell line expressing HuPKR F489S restricted VacVΔE3L+HCMVTRS1 , demonstrating that this mutation renders HuPKR resistant to HCMVTRS1 activity in the context of viral infection . RhCMVTRS1 and AgmCMVTRS1 were also unable to antagonize HuPKR F489S efficiently and rescue replication , suggesting that mutating position 489 of HuPKR to the AgmPKR variant is not sufficient for conferring sensitivity to these Old World monkey CMV TRS1 proteins ( Fig 3A ) . Interestingly , SmCMVTRS1 rescued VacVΔE3L replication even in cells expressing HuPKR F489S , consistent with its broad PKR inhibitory activity ( Fig 3A ) . The inability of VacVΔE3L+HCMVTRS1 to replicate in HuPKR F489S cells strongly suggests that HCMVTRS1 is unable to prevent activation of HuPKR F489S . To test this , we evaluated the levels of phosphorylated PKR and eIF2α following infection . As expected , PKR phosphorylation was observed in response to VVΔE3L infection in both HuPKR and HuPKR F489S cells , demonstrating that activation of the PKR pathway had been initiated ( Fig 4 ) . While we did not observe increased eIF2α-P levels in response to VVΔE3L infection in the HuPKR cell line in this experiment , we did observe increased eIF2α-P in the HuPKR F489S cells . Consistent with previous findings , VacVΔE3L+HCMVTRS1 was able to block phosphorylation of PKR and eIF2α in HuPKR cells [18] . However , robust phosphorylation of both PKR and eIF2α was observed in HuPKR F489S cells in response to VacVΔE3L+HCMVTRS1 infection ( Fig 4 , lanes 4 vs 8 ) . Thus , consistent with the replication data ( Fig 3A ) , HCMVTRS1 prevents activation of the PKR pathway in HuPKR cells but not in HuPKR F489S cells . We next investigated how this single amino acid change made HuPKR resistant to HCMVTRS1 . Since prior studies suggested that HCMVTRS1 must bind directly to PKR to effectively antagonize the PKR pathway [20–22] , we tested the hypothesis that altering residue 489 of HuPKR from phenylalanine to serine interfered with HCMVTRS1 binding to PKR . We transfected HeLa PKR KO cells with either WT HuPKR or HuPKR F489S along with a panel of His-tagged TRS1 constructs or His-tagged GFP as a negative control . Following cell lysis , we affinity purified the His-tagged proteins along with any bound PKR . We detected TRS1 and PKR in the lysate and bound fractions using anti-His and anti-PKR antibodies , respectively ( Fig 5 ) . As expected , HCMVTRS1 bound to WT HuPKR; however , this interaction was severely disrupted by the F489S mutation . In contrast , SmCMVTRS1 bound to both WT HuPKR and the F489S mutant equally well , consistent with its ability to antagonize both forms of HuPKR ( Fig 3A ) , while AgmCMVTRS1 did not bind to either PKR variant . These results indicate that position 489 is a critical residue mediating the interaction between HCMVTRS1 and HuPKR and that disrupting this interaction impairs HCMVTRS1 activity . Amino acid 489 falls within the αG-helix of PKR , which mediates the interaction between PKR and its downstream substrate , eIF2α [23 , 24] . In fact , structural data suggest that F489 directly interacts with eIF2α by projecting into a hydrophobic pocket [23] ( Fig 6A ) . Despite this seemingly critical interaction , position 489 is rapidly evolving in primates , consistent with it being engaged in an ongoing arms race with viral antagonists [16] ( Fig 6B ) . To determine how tolerant HuPKR function is to changes at position 489 , we generated HuPKR variants by introducing all possible amino acid substitutions at position 489 . These variants were then evaluated for their ability to restrict SEAP reporter gene expression , which reflects the ability of PKR to bind to and phosphorylate eIF2α and thereby inhibit translation [7] . Surprisingly , with the exception of the weak activity of proline , all of the amino acid variants maintained the ability to restrict SEAP expression to levels comparable to that of wild-type HuPKR ( Fig 6C , gray bars ) . In contrast , a point mutant of PKR that lacks kinase activity [25] ( KD HuPKR ) and cannot phosphorylate eIF2α had little effect on SEAP activity . This demonstrates that , position 489 is highly tolerant of amino acid changes and suggests it does not play an essential role in the interaction between HuPKR and eIF2α . Of note , the other three cellular eIF2α kinases , which do not appear to be engaged in arms races with viral factors , have completely conserved the site corresponding to position 489 of PKR [16] . Thus , these results demonstrate the robustness of PKR’s interaction with eIF2α despite the pressure for continual innovation within the αG-helix in order to evade viral antagonists . We next wanted to know whether HCMVTRS1 exhibited similar flexibility it its ability to recognize different amino acids at position 489 . We found that in addition to phenylalanine , HCMVTRS1 could antagonize HuPKR expressing tryptophan , methionine , or tyrosine at position 489 , and had moderate activity against leucine ( Fig 6C ) . These amino acids each have hydrophobic side chains , suggesting that they may interact with a hydrophobic pocket in HCMVTRS1 . In contrast , HCMVTRS1 had little activity against the other amino acid substitutions . Thus , our results clearly indicate that while HuPKR can tolerate a diverse range of amino acids at position 489 without losing its functional interaction with eIF2α , HCMVTRS1 is active in counteracting only a small subset of these variants . The surprising plasticity at this site may contribute to PKR’s ability to remain competitive in the arms race with rapidly evolving viral genes . In addition to position 489 , two other codons ( 492 and 496 , Fig 6B ) within the αG-helix are rapidly evolving among primates [16] , suggesting that this helix may be a hot spot for targeting by viral antagonists . Consistent with this hypothesis , poxviruses encode a PKR pseudosubstrate that structurally mimics eIF2α and competitively inhibits eIF2α docking by binding to the αG helix . In fact , previous work on the VacV eIF2α mimic K3L identified position 492 as a major species-specific determinant of PKR activity within hominoids [16] . Another neighboring codon in human PKR , A488 , has been implicated in the differential sensitivity of human vs . mouse PKR to VacV K3L [17] . The proximity of the amino acids that affect K3L and HCMVTRS1 activities indicate that interactions at the αG helix are essential for both antagonists . However , unlike K3L , HCMVTRS1 shows no obvious primary sequence homology to eIF2α , suggesting that two unrelated viruses convergently evolved to target this vulnerable region of PKR . Given the shared interaction at the αG-helix , we investigated whether altering position 489 of HuPKR would also impact the activity of K3L . WT K3L has very little activity against HuPKR , while an experimentally evolved form of K3L containing a single amino acid change ( H47R ) confers moderate activity against HuPKR [26 , 27] . Thus , we evaluated the replication profiles of viruses expressing WT K3L and K3L H47R in our HuPKR and HuPKR F489S cells . All of the viruses replicated well in the empty vector cell line and expressed similar levels of K3L ( Fig 7 ) . Similar to previous findings [26] , K3L H47R rescued VACVΔE3L replication relative to WT K3L in cells expressing HuPKR ( Fig 7A ) . Notably , the replication advantage conferred by K3L H47R was abrogated in cells expressing HuPKR F489S , demonstrating that this single amino acid change in PKR confers resistance to two unrelated viral antagonists . This result highlights the complexity of the host-virus arms race in broadly acting factors like PKR where multiple viral antagonists may target a shared interface . In these scenarios , mutations driven by one virus have the potential to alter host sensitivity or resistance to unrelated viral antagonists .
Our results provide evidence for a co-evolutionary history between PKR and the CMV TRS1 proteins , as reflected by specific adaptation of HCMVTRS1 and AgmCMVTRS1 to their cognate PKRs . These adaptations have led to divergence of the PKR:TRS1 interface , resulting in species-specific differences in TRS1 activity . For example , in contrast to the critical role of amino acid 489 of PKR for HCMVTRS1 activity , this site does not appear to mediate sensitivity to the orthologous proteins from three other primate CMVs . SmCMVTRS1 retains the ability to bind to and antagonize HuPKR F489S ( Figs 3A and 5 ) , and in fact it appears to be active against a broad range of primate PKRs ( Fig 1A ) . Thus , compared to HCMVTRS1 , SmCMVTRS1 may have evolved to recognize and bind to a different site that may be more conserved among primate PKRs . On the other hand , the RhCMV and AgmCMV TRS1 alleles are unable to inhibit HuPKR even when codon 489 is the variant found in their Old World monkey hosts ( 489S ) . Although we have not precisely mapped species-specific determinants in these cases , we observed that the Agm PKR kinase domain was sufficient to confer sensitivity to inhibition by AgmCMVTRS1 ( S2A Fig ) . Unlike HCMVTRS1 , which inhibits the PKR pathway prior to autophosphorylation , RhCMVTRS1 allows PKR autophosphorylation but blocks eIF2α phosphorylation [18] . Thus , the evolution of PKR in primates appears to have led to quite divergent adaptions in the CMV TRS1 antagonists that changed their precise binding interactions and resulted in alternative species-specific mechanisms of PKR inhibition . Even while PKR has been evolving to evade viral antagonists , it is constrained by the need to maintain recognition of dsRNA and eIF2α . Thus , residues that are critical for these interactions are more likely to be immutable , while those that are detected by the viral antagonists are more likely to change . In fact , codons evolving under positive selection in primates were reported to be widely dispersed across the PKR coding sequences , though none were found within the dsRNA-binding domains [16] . On the other hand , at least three codons within the αG helix are highly variable among primates [16] , even though the PKR kinase domain-eIF2α co-crystal shows that this region , and residue 489 in particular , makes direct contact with eIF2α [23] and thus might be expected to be particularly intolerant of substitutions . Although variation at this site in other primate PKR genes might depend on epistatic mutations elsewhere in PKR that help maintain the interaction with eIF2α , we found that HuPKR retains its inhibitory activity when any of 19 different amino acid substitutions are introduced at residue 489 in human PKR . The only exception is proline , which may not be tolerated due to its helix-breaking properties . These findings support structural data showing that the helical insert between βstrands 3 and 4 of eIF2α is relatively flexible ( high B-factor ) [28 , 29] and thus able to adapt to maintain interactions with a rapidly evolving PKR interface . In contrast , VacV K3L , the structural mimic of eIF2α , has been proposed to be more rigid and thus less tolerant of mutations in the αG helix [30] . Prior studies have identified other mutations in the αG helix that preserve ( codons 490 and 499 ) or eliminate ( codons 487 and 495 ) eIF2α recognition by PKR [24] , but none of these are variable among primate PKR alleles . Thus , HuPKR appears to retain its eIF2α kinase function despite mutations at a subset of positions within this otherwise critical structure . This mutational tolerance at codon 489 and nearby residues may facilitate acquisition of adaptive changes in PKR during its arms race with viral antagonists that target this site . Because PKR senses dsRNA , which accumulates during replication of diverse viruses [31] it has broad anti-viral activity and has also been the target of antagonists encoded by many different virus families [8 , 9] . Even within the CMV subfamily , adaptations to diverging PKR alleles during co-speciation of CMVs with their hosts have altered the PKR:TRS1 binding interfaces . In contrast to this divergent evolution of closely related PKR antagonists , here we also identify convergent evolution by the unrelated antagonists HCMVTRS1 and poxvirus K3L . Prior studies mapped species-specific determinants of VacV K3L sensitivity to PKR codons 488 [17] and 492 [16] , both of which are within the αG helix and proximal to the site we found to be critical for sensitivity to HCMVTRS1 . Furthermore , we demonstrate that codon 489 also impacts K3L activity , as introducing a serine at position 489 of human PKR is sufficient to confer resistance to both HCMVTRS1 and VacV K3L ( H47R ) . Thus , the conflict between PKR and its viral antagonists should be viewed as a multilateral arms race , in which mutations driven by one virus can have collateral effects on other antagonists . This is true for the αG helix of PKR , which might be targeted by unrelated viral antagonists precisely because it is critical for PKR function . In response , during the evolution of host restriction factors like PKR , preservation of robustness may depend on residues , such as 489 , embedded within critical functional domains that can serve as mutable decoys to enable rapid adaptation to multiple viruses without compromising the core functions of the protein .
All primers used to construct plasmids are listed in S1 Table . The SEAP assay was carried out as described previously [19] , except that 0 . 05 μg of SEAP reporter was transfected along with a given PKR and TRS1 construct at a ratio of 1:2 , respectively . The SEAP expression vector ( pEQ886 ) has been described previously [32] . A plasmid expressing EGFP with a 6x-His tag ( pEQ1100 ) was used as a the control and vector control in SEAP assays and has been described previously [33] . HCMVTRS1 was expressed from pEQ1180 ( formerly 981 ) [34] . AgmCMVTRS1 was PCR amplified from AgmCMV DNA ( ATTC VR-706 ) using primers 931 and 932 and TOPO cloned into pcDNA3 . 1V5-6xHis ( ThermoFischer Scientific ) to yield pEQ1377 . RhCMVTRS1 was digested from pEQ1215 [18] with HindIII and NotI and ligated into pEQ1180 that had been cut with the same enzymes to yield pEQ1261 . Because we were unsuccessful in PCR amplifying SmCMVTRS1 from viral DNA , possibly due to its very high GC content , we synthesized a mammalian codon-optimized form of SmCMVTRS1 flanked by the MCS from pcDNA3 . 1v5-His to facilitate later cloning ( GenScript Inc . ; GenBank accession number KX518569 ) . This synthesized construct was inserted into pUC57 using EcoRI and HindIII to produce pEQ1494 . SmCMVTRS1 was removed from pEQ1494 using Asp718 and NotI and ligated into pEQ1377 cut with the same enzymes , resulting in pEQ1495 , which was used in SEAP experiments . A knockdown resistant form of HuPKR ( SR#329 ) generously provided by Stefan Rothenburg ( Kansas State University ) [17] was used for HuPKR expression in Figs 2B and 6C . To generate a construct expressing active HuPKR with a 6xHis tag , pSB819+HuPKR [16] was digested with EcoRI to yield a PKR fragment that was then ligated into the same sites in pEQ1198 [18] , resulting in pEQ1356 . However , PKR expression from pEQ1356 was insufficient to repress SEAP expression in the reporter assay , so HuPKR with a 6xHis tag was then PCR amplified from pEQ1356 using primers 2058 and 2059 , and cloned into SR#329 cut with KpnI and HindIII using Gibson assembly , resulting in pEQ1563 . Similarly , AgmPKR was isolated from pSB819+AgmPKR , which has been described previously [16] , and introduced into a pcDNA3 . 1v5-His backbone that also contained a biotinylation signal between XhoI and XbaI , resulting in pEQ1357 . AgmPKR was PCR amplified from pEQ1357 using primers 2058 and 2059 and introduced into SR#329 that had been digested with KpnI and HindIII using Gibson assembly ( New England Biolabs ) to produce pEQ1564 . pEQ1563 and pEQ1564 were used in S2A Fig . To generate chimeras between HuPKR and AgmPKR , the N-terminal half of HuPKR or AgmPKR was amplified from pEQ1356 and pEQ1357 , respectively , using primers 2033 and 2036 , while the C-terminal half of each PKR was amplified using primers 2034 and 2035 . The N-terminal and C-terminal fragments were then cloned into pEQ1357 digested with BamHI and EcoRV using Gibson assembly to yield a Hu-Agm chimera ( pEQ1555 ) and an Agm-Hu chimera ( pEQ1556 ) in the pcDNA3 . 1v5-His backbone . The two his-tagged chimeras were then PCR-amplified from these constructs using primers 2058 and 2059 and inserted into SR#329 digested with KpnI and HindIII using Gibson assembly , resulting in pEQ1571 ( Hu-Agm PKR-6xHis ) and pEQ1572 ( Agm-Hu PKR-6xHis ) , which were used in S2A Fig . Constructs without tags were generated as follows: HuPKR and AgmPKR were PCR amplified from pEQ1356 and pEQ1357 using primers 2102 and 2104 or 2102 and 2103 , respectively . The PCR products were then introduced into SR#329 cut with KpnI and HindIII by Gibson assembly to yield pEQ1602 ( HuPKR ) and pEQ1598 ( AgmPKR ) , which were used in Fig 2C and S2B Fig . To generate the Hu-AgmD chimera , the N-terminal portion of HuPKR was PCR amplified from pEQ1356 using primers 2102 and 2085 , while the AgmD fragment was amplified from pEQ1357 using primers 2084 and 2103 . These fragments were cloned into SR#329 using Gibson assembly , resulting in pEQ1600 . The Hu-AgmD1 , D2 and D3 chimeras were first constructed with 6xHis tags by PCR amplifying the various regions of HuPKR and AgmPKR as follows . Hu-AgmD1: pEQ1356 was PCR-amplified using primer pairs 2058 , 2085 and 2098 , 2059; pEQ1357 was PCR amplified using primers 2084 , 2099 . Hu-AgmD2: pEQ1356 was PCR-amplified using primer pairs 2058 , 2099 and 2100 , 2059; pEQ1357 was PCR amplified using primers 2098 , 2101 . Hu-AgmD3: pEQ1356 was PCR amplified using primers 2058 , 2101; pEQ1357 was PCR amplified using primers 2100 , 2059 . The fragments were then cloned into SR#329 using Gibson assembly , resulting in pEQ1595 ( Hu-AgmD1-His ) , pEQ1596 ( Hu-AgmD2-His ) , and pEQ1597 ( Hu-AgmD3-His ) . To remove the 6xHis tags , the chimeras were PCR-amplified using the following primers: 2102 , 2104 ( pEQ1595 and pEQ1596 ) or 2102 , 2103 ( pEQ1597 ) . The PCR products were cloned into SR#329 to generate pEQ1603 ( Hu-AgmD1 ) , pEQ1604 ( Hu-AgmD2 ) and pEQ1599 ( Hu-AgmD3 ) , which were used in S2B Fig . Point mutations in HuPKR were constructed by using complementary forward and reverse primers harboring the desired mutation paired with primers at the beginning or end of PKR , as shown in S2 Table . pEQ1356 was used as a template for PCR amplification and the resulting N and C terminal fragments containing the desired mutations were cloned into SR#329 cut with KpnI and HindIII of using Gibson assembly . WT VacV ( VC2+lacZ ) , VacVΔE3L , VacVΔE3L+HCMVTRS1 ( VVeq1148 ) , and VacVΔE3L+RhCMVTRS1 ( VVeq1233 ) have been described previously [18] . To generate VacVΔE3L+AgmCMVTRS1 , AgmCMVTRS1 was first removed from pEQ1377 ( see SEAP plasmids , above ) using Asp718 and NotI and ligated into pEQ1233 [18] , that had been digested with the same enzymes , resulting in pEQ1453 . pEQ1453 was then used to introduce AgmCMVTRS1 into the thymidine kinase locus in VacVΔE3L through homologous recombination to generate VacVΔE3L+AgmCMVTRS1 ( VVeq1453 ) . Similarly , SmCMVTRS1 was cut from pEQ1495 ( see SEAP plasmids ) with Asp718 and NotI and inserted into the same sites of pEQ1453 to produce pEQ1497 , which was recombined into VacVΔE3L to produce VacVΔE3L+SmCMVTRS1 ( VVeq1497 ) . Recombinant VacVs were propagated and titered in BHK cells . In all viral replication experiments , cells were infected at an MOI of 0 . 1 for 1 hour , after which the medium was replaced . Viral replication was evaluated at 48 hpi by measuring β-Galactosidase ( β-Gal ) activity via a fluorometric substrate cleavage assay , as described previously [33] , or through plaque assays in HeLa PKR KO cells ( S1B Fig ) . With the stable cell lines expressing PKR in an inducible manner ( Figs 3 and 7 ) , cells were treated with medium containing 1 μg/ml of doxycycline 24 hours prior to infection , infected in the absence of doxycycline , and refed with medium containing 1ug/ml of doxycycline after the hour-long infection period . To evaluate the status of PKR phosphorylation during infection , cells were treated with medium containing 1 μg/ml of doxycycline for 24 hours and then infected at an MOI of 3 for 1 hour , after which the cells were re-fed with medium containing 1 μg/ml of doxycycline . Cell lysates were harvested in 2% SDS at 6 hpi . HeLa , BSC40 , and HeLa PKR KO cell lines were maintained at 37°C with 5% CO2 in Dulbecco’s modified Eagle’s medium ( DMEM ) supplemented with 10% NuSerum ( BD Biosciences ) . HeLa cells were provided by Bertram Jacobs ( Arizona State University ) and BSC40 cells were provided by Stanley Riddell ( Fred Hutchinson Cancer Research Center , Seattle , WA ) . HeLa PKR KO cell lines were generated by transfecting HeLa cells with plasmid vectors that express Cas9 ( HCas9 , a gift from George Church , Addgene #41815 [35] , two guide RNAs ( pEQ 1451 and pEQ1452 ) that target genomic sequences upstream ( 5’-TCTCTTCCATTGTAGGATA-3’ ) and downstream ( 5’-CTTTTCTTCCACACAGTCA-3’ ) of PKR , and a homologous recombination vector containing mCherry and puromycin ( pEQ1489 ) . After puromycin selection , single-cell clones were evaluated by sequencing and immunoblotting to identify clean knockouts; Clone #6 was used in these studies . To generate stable cell lines expressing different PKR constructs in the HeLa PKR KO cell line , we first cloned each PKR into the pSLIK-Hygro lentiviral vector ( a gift from Iain Fraser , Addgene plasmid # 25737 [36] ) . HuPKR was PCR amplified from pEQ1356 using primers 2105 and 2106 and inserted into pEN_TmiRc3 Entry Vector ( a gift from Iain Fraser , Addgene plasmid # 25748 [36] ) at sites SpeI and XbaI using Gibson assembly to produce pEQ1606 . HuPKR was then moved from pEQ1606 into the pSLIK-Hygro destination vector using Gateway Cloning ( Thermo Fisher Scientific ) , resulting in pEQ1607 . To move AgmPKR into this vector , AgmPKR was PCR amplified from pEQ1357 using primers 2175 and 2176 , and inserted into pEQ1607 that had been digested with BstEII using Gibson assembly , yielding pEQ1641 . Similarly , HuPKR F489S was PCR-amplified from pEQ1624 using primers 2173 and 2174 and introduced into pEQ1607 that had been digested with BstEII using Gibson assembly , resulting in pEQ1642 . HeLa PKR KO cells were transduced with lentiviral vectors encoding pEQ1607 , pEQ1641 , pEQ1642 , and pSLIK-Hygro ( empty vector ) . After hygromycin selection , single cell clones were evaluated for their ability to restrict VacVΔE3L and for PKR expression levels , with the following clones used in experiments: HeLa PKR KO+1607#9 , 1641#3 and 1642#2 . Cell lysates were harvested in 2% sodium dodecyl sulfate ( SDS ) and equal amounts of lysates were separated on a 10% SDS-polyacrylamide gel ( except for Fig 7 , where a 15% SDS-polyacrylamide gel was used ) and transferred to polyvinylide difluoride ( PVDF ) membranes . Proteins were detected using the Western-Star chemiluminescent detection system ( Applied Biosystems ) with the following primary antibodies: PKR D7F7 ( #12297 , Cell Signaling Technology ) , P-PKR E120 ( ab32036 , abcam ) , Penta-His ( 34660 , Qiagen ) , K3L ( a gift from J . Tartaglia [37] ) , P-eIF2α ( #3597 , Cell Signaling ) , total eIF2α ( #2103 , Cell Signaling ) , and Actin ( A2066 , Sigma ) . HeLa PKR KO cells were transfected using Lipofectamine 2000 ( Invitrogen ) with 0 . 15 μg of HuPKR ( pEQ1602 ) or HuPKR F489S ( pEQ1624 ) and 2 . 35μg of HCMVTRS1 ( pEQ1180 ) , AgmCMVTRS1 ( pEQ1377 ) , SmCMVTRS1 ( pEQ1495 ) or EGFP ( pEQ1100 ) . At 48 hours post-transfection , cells were washed with PBS and harvested in 250μl of cold NiNTA Lysis Buffer ( 50mM NaH2PO4 , 300mM NaCl , 10mM imidazole , 0 . 75% Tween 20 , 1μM benzamidine and 100 μM PMSF ) . Lysates were incubated on ice for 20 minutes with occasional vortexing and centrifuged at 16 , 000xg for 10 min at 4°C to remove cell debris . 20 μl of each lysate was reserved , while 200 μl of the remaining lysate was added to 30 μl of PerfectPro NiNTA Superflow agarose ( 5prime ) and incubated at 4°C on a rotating mixer for 2 hours . After binding , the beads were pelleted and washed 3x with 500μl of NiNTA Wash buffer ( 50mM NaH2PO4 , 300mM NaCl , 20mM imidazole , 0 . 75% Tween 20 ) . The lysate and bound samples were denatured in SDS-PAGE sample buffer at 95°C ( 5 minutes ) and separated on a 10% SDS-polyacrylamide gel , transferred to a PVDF membrane and probed with PKR D7F7 ( Cell Signaling Technology ) and Penta-His ( Qiagen ) antibodies . The structure of PKR in complex with eIF2α was visualized using data from the protein databank ( http://www . pdb . org; ID 2A1A ) and MacPyMol . For sequence comparisons of the αG helix among primate PKRs , the following sequences were obtained from NCBI and aligned using Clustal Omega: Homo sapiens ( human; BC093676 ) , Gorilla gorilla ( Gorilla; EU733258 ) , Pongo pygmaeus pygmaeus ( Bornean orangutan; EU733259 ) , Nomascus leucogenys ( northern white-cheeked gibbon; EU733257 ) , Hylobates agilis albibarbis ( Agile gibbon; EU733270 ) , Colobus guereza ( Colobus monkey; EU733267 ) , Macaca mulatta ( Rhesus monkey; EU733261 ) , Cercopithecus aethiops ( African green monkey; EU733254 ) , Miopithecus talapoin talapoin ( Talapoin monkey; EU733269 ) , Ateles geoffroyi ( Black-handed spider monkey; EU733263 ) , Callicebus moloch ( Dusky titi; EU733265 ) , Saimiri boliviensis boliviensis ( Squirrel monkey , XM_003926814 ) . The sequence of the mammalian codon-optimized SmCMVTRS1 has been deposited in GenBank under the following accession number: KX518569 . | Cytomegaloviruses ( CMVs ) are highly species-specific , but the host factors that prevent replication in heterologous species are largely unknown . Based on data indicating that the broadly-acting host antiviral factor protein kinase R ( PKR ) has diversified rapidly during evolution , we hypothesized that PKR may contribute to cross-species barriers to CMV replication . In support of this hypothesis , we find that primate CMVs differ in their ability to antagonize PKRs from different primates . By leveraging these differences , we identified a single amino acid at codon 489 in human PKR that dictates PKR susceptibility to the human CMV PKR antagonist , HCMVTRS1 . This amino acid is positioned within a helix that mediates the critical interaction between PKR and its downstream substrate eIF2α . Despite this seemingly important structural role , human PKR is highly tolerant of amino acid substitutions at position 489 , allowing it the flexibility to adapt in order to evade viral antagonists without disrupting its antiviral activity . Remarkably , position 489 also dictates PKR sensitivity to the entirely unrelated poxvirus-encoded PKR antagonist , K3L . Thus , mutations driven by one virus can impact the host’s sensitivity to unrelated viral antagonists , illustrating the multilateral nature of the host-viral “arms-races” between viruses and broadly acting antiviral host defenses . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"organismal",
"evolution",
"medicine",
"and",
"health",
"sciences",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"microbiology",
"vertebrates",
"cytomegalovirus",
"infection",
"animals",
"mammals",
"plasmid",
"construction",
"primates",
"viruses",
"dna",
"vi... | 2016 | A Single Amino Acid Dictates Protein Kinase R Susceptibility to Unrelated Viral Antagonists |
Due to the large number of putative microRNA gene targets predicted by sequence-alignment databases and the relative low accuracy of such predictions which are conducted independently of biological context by design , systematic experimental identification and validation of every functional microRNA target is currently challenging . Consequently , biological studies have yet to identify , on a genome scale , key regulatory networks perturbed by altered microRNA functions in the context of cancer . In this report , we demonstrate for the first time how phenotypic knowledge of inheritable cancer traits and of risk factor loci can be utilized jointly with gene expression analysis to efficiently prioritize deregulated microRNAs for biological characterization . Using this approach we characterize miR-204 as a tumor suppressor microRNA and uncover previously unknown connections between microRNA regulation , network topology , and expression dynamics . Specifically , we validate 18 gene targets of miR-204 that show elevated mRNA expression and are enriched in biological processes associated with tumor progression in squamous cell carcinoma of the head and neck ( HNSCC ) . We further demonstrate the enrichment of bottleneckness , a key molecular network topology , among miR-204 gene targets . Restoration of miR-204 function in HNSCC cell lines inhibits the expression of its functionally related gene targets , leads to the reduced adhesion , migration and invasion in vitro and attenuates experimental lung metastasis in vivo . As importantly , our investigation also provides experimental evidence linking the function of microRNAs that are located in the cancer-associated genomic regions ( CAGRs ) to the observed predisposition to human cancers . Specifically , we show miR-204 may serve as a tumor suppressor gene at the 9q21 . 1–22 . 3 CAGR locus , a well established risk factor locus in head and neck cancers for which tumor suppressor genes have not been identified . This new strategy that integrates expression profiling , genetics and novel computational biology approaches provides for improved efficiency in characterization and modeling of microRNA functions in cancer as compared to the state of art and is applicable to the investigation of microRNA functions in other biological processes and diseases .
Since the discovery of microRNAs as important regulators of broad biological processes [1]–[5] , characterization of their functions in cancer has been hindered by lack of microRNA profiling information in tumors such as squamous cell carcinoma of the head and neck ( HNSCC ) . Previous reports show that only one or a few gene targets , identified among predicted or differentially expressed genes , were directly targeted by the microRNA under investigation [6]–[8] . While sequence-based computational algorithms have been applied for predicting all potential microRNA gene targets; false positive rates remains relatively high [9] , [10] . Further , sequence-based predictions are unable , by design , to account for biological contexts ( e . g . cell and tissue types , normal or disease conditions ) and thus are not optimized for predicting the biological function of genes targeted by cancer microRNAs . Moreover , genome-scale and biological studies have yet to identify key regulatory networks perturbed by altered microRNA functions in cancer . To investigate microRNA function in HNSCC , we sought to develop an effective computational approach that is complementary to microRNA profiling and , in addition , is capable of simultaneously predicting tumor suppressor microRNAs as well as their functional targets from gene expression . In this report we illustrate how phenotypic knowledge of genetic disorders ( OMIM database ) can be utilized jointly with gene expression analyses to achieve this goal . Using this approach , we selected miR-204 among ten prioritized microRNAs for biological characterization , as miR-204 is located at the cancer-associated genomic region ( CAGR ) 9q21 . 1–q22 . 3 locus exhibiting high frequency of Loss of heterozygosity ( LOH ) in human HNSCC [11]–[15] , and a CAGR for which candidate tumor suppressor gene targets have not been identified . Additionally , we report the first computationally predicted and biologically validated microRNA-regulated network that is dependent on the epidermal growth factor receptor ( EGFR ) whose overexpression occurs in over 80% of head and neck cancer . We further demonstrate that gene targets of miR-204 exhibit enriched bottleneck and hub network topology properties in a predicted protein-protein interaction network ( PPIN ) . Moreover , we confirm the validity of our computational predictions of a microRNA function , as well as its gene targets and system's properties through conducting extensive and thorough biological characterization using a clinically relevant in vivo metastatic model of head and neck cancer . In summary , we show how such a high throughput system's strategy can accelerate the investigation of microRNA function in cancer by illustrating altered complex biological processes and regulatory pathways associated with microRNA dysfunction in cancer , by identifying among all putative microRNA gene targets only those that are dysregulated , and by elucidating molecular interactions underpinning microRNA regulation of malignant transformation and progression . The ability to characterize tumor suppressor microRNAs through a network analysis of mRNA expression datasets would be a major advance with potentially wide application . Further , we provide experimental evidence linking microRNA function to the genetic risk of HNSCC . We show at the LOH 9q21 . 1–22 . 3 locus , miR-204 could serve as a tumor suppressor of HNSCC oncogenesis and progression .
At the time of initiating this study , comprehensive analysis of microRNA expression profile in head and neck cancer ( HNSCC ) was not available and would have required time-consuming accruement of tumor tissues for conducting such analysis , a situation that is not limited to HNSCC research . We hypothesized that the development of a computational capability to simultaneously predict tumor suppressor microRNAs as well as their functional targets from more widely available genome-wide gene expression datasets could be an efficient reverse engineering approach for identifying deregulated microRNAs and their functional gene targets . We first developed IMRE , a statistical method to predict altered expression of microRNAs from genome-wide mRNA expression and putative microRNA targets databases ( Supporting Figure 2 in Text S1 , Materials and Methods ) . This strategy is based , in part , on the observations that at genome scale the expression of microRNAs and their direct mRNA targets are , in general , inversely correlated [16] , [17] . To conduct this analysis , we integrated five complementary microRNA target databases to generate “miRNOME” that contains 534 human microRNAs and 17 , 343 microRNA gene targets ( Materials and Methods , and Table 1 in Text S2 ) . We validated this method using two independent cancer expression profiling experiments in GEO comprised of paired mRNA and microRNA expressions for tumors and normal tissue ( GSE2564 [18]: multiple epithelial cancer; GSE8126 [19]: prostate cancer ) . IMRE-predicted downregulated microRNAs that are exclusively inferred from mRNA expression and microRNA targets datasets ( Materials and Methods ) were enriched in the expression analysis of the corresponding microRNA array dataset ( GSE2564: P = 0 . 014; GSE8126: P = 0 . 0002 respectively , cumulative hypergeometric test , data not shown ) . A recent study also demonstrated the increased prediction specificity of microRNA and its gene target relationship via intersecting the results of multiple prediction algorithms [20] . Subsequently , we applied the IMRE method to analyze two independent HNSCC mRNA microarray datasets for predicting deregulated microRNAs from genome-wide mRNA expression ( Supporting Figure 2 in Text S1 , Materials and Methods ) : first , the GSE6631 set that provides differential mRNA gene expression between 22 HNSCC non-microdissected patient tumor samples and their paired normal squamous tissues [21] , and second , the GSE2379 [22] set that contains 34 micro-dissected node-positive HNSCC tumors of the hypopharynx . We noted that vast majority of the known microRNAs had at least one putative target in the top 500 deregulated genes of the HNSCC expression arrays ( GSE6631 ) , with a median of 19 targets . Therefore , it is unfeasible to manually select microRNA candidates from their deregulated targets for biological validation . Applying IMRE method to each dataset , we predicted a set of down-regulated microRNAs in HNSCC ( 113 and 43 , respectively; FDR ≤0 . 05 , Materials and Methods ) , of which 34 were consistently found in both prediction sets ( P-value = 2 . 0×10−16 , Fisher's exact test , Figure 1A and Table 2 in Text S2 , FDR <0 . 05 , Materials and Methods ) . Predictions of up-regulated microRNAs did not reach reproducible statistical significance ( not shown ) . To further reduce the number of microRNAs to the most promising candidates for HNSCC , we conducted a statistical enrichment analysis of putative microRNA targets among inheritable cancer genes in the OMIM human disease gene database [Online Mendelian Inheritance in Man , http://www . ncbi . nlm . nih . gov/omim/ ( downloaded Dec . 1 , 2006 ) ] . OMIM contains 610 biologically validated cancer genes among which 586 ( 96% ) are predicted targets of 527 microRNAs in miRNOME . We observed that each of the 527 microRNAs could , on average , target 30 OMIM cancer genes ( not shown ) . Thus , it is also unfeasible to manually select microRNA candidates from OMIM cancer genes for biological validation . Our analyses identified 46 microRNAs significantly enriched in the inheritable cancer gene subset of OMIM in the miRNOME ( Figure 1A; Table 3 in Text S2 , Materials and Methods , Protocol S1/Section A , and Dataset S1 ) . Since microRNAs can be deregulated across cancers of different tissue origin [23] , we performed a review of literature and confirmed the validity of these 46 predictions ( OMIM; Supporting Figure 3 in Text S1 , P = 0 . 039; cumulative hypergeometric test , Table 4 in Text S2 , Materials and Methods ) . Thereafter , we reduced the list of candidates in HNSCC to ten microRNAs ( Figure 1A ) that were predicted in the HNSCC gene expression ( 34 microRNAs; Table 2 in Text S2 ) as well as in inheritable cancer genes ( 46 microRNAs; Table 3 in Text S2 ) . Among the ten prioritized microRNAs , four belong to the let-7 tumor suppressor microRNA family ( Figure 1A ) . We chose miR-204 among the ten prioritized microRNAs for thorough biological characterization based on the following considerations . First , miR-204 is located within the sixth intron of the host gene transient receptor potential melastatin 3 cation channel ( TRPM3 , NM_020952 ) and is transcribed in the same direction as TRPM3 [24] . TRPM3 is located on human chromosome 9q21 . 11 that is within the 9q21 . 1–q22 . 3 locus exhibiting high frequency of Loss of heterozygosity ( LOH ) in human HNSCC [11]–[15] . LOH at 9q21 . 1–q22 . 3 occurs in 37% of premalignant head and neck lesions , and increases to 67% in HNSCC [14] . Second , in addition to the genomic imbalance at 9q21 . 1–q22 . 3 locus , chromosomal aberrations occur most frequently at 3p , 5q , 9p , 11q and 17p in HNSCC [11] , [12] , [25] . With the exception of let-7g that is located at the 3p21 locus ( note that let7g is also included in the class of microRNAs with related mature sequence “Let7/98” ) , the other 7 prioritized microRNAs are not in the cancer associated genomic regions ( CAGRs ) . Third , while potential tumor suppressor gene candidates have been identified for other CAGRs in HNSCC , gene candidates possessing tumor suppressor activity associated with the 9q21 locus have not been identified . Thus the mechanisms by which changes at this locus affecting HNSCC oncogenesis remain uncharacterized . Fourth , the role of miR-204 in human cancer has not been established . We first examined miR-204 host gene TRPM3 expression by quantitative PCR ( qPCR ) and observed near complete TRPM3 suppression in four micro-dissected HNSCC tumors ( Figure 1B ) and in a panel of 10 low passage HNSCC cell lines generated from tumors of diverse head and neck locations ( Figure 1C and Table 5 in Text S2 ) [26] . We subsequently measured miR-204 expression in HNSCC tumors and cell lines . Consistent with the observed near complete loss of TRPM3 ( Figures 1B–C ) , miR-204 expression was inhibited in all four tumors by 85% to 99% ( Figure 1D ) , and by more than 90% in all ten HNSCC cell lines ( Figure 1E ) compared to samples of pooled normal buccal mucosa . The frequent allelic loss at 9q21 . 1–q22 . 3 in HNSCC [11]–[14] provides genetic evidence that loss of miR-204 microRNA function may occur as a result of genomic imbalance at this site and that miR-204 may be a potential candidate associated with the tumor suppressor activity of 9q21 . 1–q22 . 3 . Since miR-204 was also predicted in the OMIM analysis to be associated with lymphoma ( Table 3 in Text S2 ) , we quantified miR-204 expression in immortalized “normal B cell 11365” and three Burkitt B-cell lymphoma cell lines and found its expression significantly reduced ( Figure 1F ) . Further , paired comparison of miR-204 expression between 6 types of adenocarcinomas and their respective normal tissues was conducted using the microRNA array dataset GSE2564 [18] . miR-204 was significantly down-regulated in breast ( P = 0 . 014 ) , kidney ( P = 0 . 004 ) and prostate ( P = 0 . 0001 ) tumors ( Figure 1G ) . Additionally , significant miR-204 down-regulation was recently reported in a subtype of acute myeloid leukemia bearing cytoplasmic mutated nucleophosmin [27] . Here , we demonstrate for the first time , the accuracy and efficiency of joint analyses of mRNA expression , inheritable disease genes , and microRNA target databases to prioritize deregulated microRNAs for biological characterization . Collectively , these biological findings support the validity of our computational predictions of miR-204 downregulation in HNSCC and suggest that it may possess tumor suppressor activity . Among the 1 , 088 putative miR-204 targets predicted in the miRNOME , 34 mRNA transcripts that were significantly upregulated in HNSCC ( GSE6631 ) led to the enrichment of miR-204 ( Figure 2A and Table 6 in Text S2 ) . We first conducted statistical functional enrichment analyses using Gene Ontology ( GO ) [28] and found a number of biological processes ( BP ) and molecular functions ( MF ) of GO were significantly enriched among 32 of the 34 miR-204 gene targets ( referred to as “functionally prioritized miR-204 targets” ) ( Table 7 in Text S2 , Materials and Methods , and Protocol S1/Section C ) . We next examined mRNA expression status of 21 representative “functionally prioritized miR-204 targets” in four laser capture microdissected HNSCC tumor samples and observed increased expression of 18 of these genes compared with their respective expression in five pooled normal buccal mucosa ( Figure 2B ) . Additionally analysis of thirteen “functionally prioritized miR-204 targets” , those enriched with the listed GO functions ( the table in Figure 2B , Materials and Methods ) , showed overexpression of nine targets in ten HNSCC cell lines ( Figure 2C ) . These results indicate that predicted miR-204 targets , upregulated in HNSCC , share similar functions and may participate in similar biological processes . To provide evidence that miR-204 can directly suppress the expression of its predicted targets in HNSCC , examination of the 3′UTR confirmed that all 34 predicted target genes contain at least one miR-204 binding site as expected by our predictions using sequence homology databases of the miRNOME ( Table 8 in Text S2 , and Materials and Methods ) . Thereafter , we selected 21 “functionally prioritized miR-204 targets” overexpressed in HNSCC ( Figure 2B ) for biological validation . We conducted in vitro miR-204 gain-of-function analyses by transiently transfecting JSQ3 and SQ38 HNSCC cells with mature miR-204 mimics ( Dharmacon ) to enhance miR-204 function in these two cell lines . Restoration of miR-204 function achieved significant inhibition ( between 30% to 75% ) of endogenous mRNA expression in 18 out of 21 predicted targets examined for both cell lines , while non-specific control mimics had no significant effect ( Figure 2D and Supporting Figures 4–5 in Text S1 ) . The specificity of miR-204 mimics was further confirmed by unaltered expression of four endogenous housekeeping genes ( GUSB , HPRT1 , HUPO and PPIA ) that lack target homology to miR-204 ( Figure 2D and Supporting Figure 4 in Text S1 ) . Comparing with sequence-based microRNA gene target prediction algorithms that have true positive rates of about 40% [9] , [10] , the accuracy of our prediction methods is higher ( ∼90% ) . Collectively , these observations indicate that down-regulation of functionally related miR-204 targets upon miR-204 mimics treatment was sequence specific and was not due to artifacts of transfection or the “off target” effect of miR-204 mimics . Following functional enrichment analysis of upregulated miR-204 targets in HNSCC , we next examined the role of miR-204 targets in modulating the function of a protein-protein interaction network ( PPIN ) . To identify genome-wide changes in PPINs associated with altered microRNA functions in HNSCC , we first integrated seven protein-protein interaction databases ( Materials and Methods ) and generated a “genome-scale PPIN” that contains 44 , 695 protein-protein interactions and 7 , 321 predicted human genes targets for the 532 microRNAs in the miRNOME . We subsequently could map 260 out of 382 ( 68% ) up-regulated genes in GSE6631 to the PPIN ( refer to as “HNSCC PPIN” ) , of which 24 were miR-204 targets predicted in miRNOME . We next computed the empirical probability of interactions among these 260 genes in the network using permutation resampling ( Materials and Methods ) . To identify the most important interactions in the HNSCC PPIN , we retained proteins for which the number of observed interactions was significantly increased in single protein network modeling as compared to those found in the empirical distribution ( Materials and Methods ) . As a result , we identified a protein regulatory network in HNSCC consisting of 56 prioritized upregulated genes in GSE6631 at a low false discovery rate of 7% ( Figure 3 and , Materials and Methods ) ( referred to as “prioritized HNSCC PPIN” ) . Among the 24 miR-204 targets mapped to the genome-scale PPIN , seven were present in the “prioritized HNSCC PPIN” ( Figure 3 , shown in red ) . Further , six of the seven-miR-204 targets remained prioritized when computed using different network modeling conditions demonstrating the robustness of our analyses ( not shown , and Materials and Methods ) . We next analyzed two topological features of the PPIN: the “hub” and “bottleneck” properties . “Hubs” , the highly connected node proteins , and “bottlenecks” , the key connector proteins , are central to controlling the connectivity of biological sub-networks to one another [29] . Further , our prior studies showed proteins possessing both properties ( hub-bottleneck ) as essential and efficient network components to alter the functional output of a PPIN upon their dynamic changes in gene expression [30] . Here , we observed significant enrichment of hubs , bottlenecks , and hub-bottleneck proteins in the 56-gene “prioritized HNSCC PPIN” as compared to either the “genome-scale PPIN” or to the “HNSCC PPIN” ( hub: P = 8 . 7×10−8; bottleneck: P = 7 . 31×10−7; hub-bottleneck: P = 1 . 61×10−8; Materials and Methods ) . Additionally , the proportion of hub-bottleneck genes was further enriched among the seven miR-204 targets present in the “prioritized HNSCC PPIN” ( P = 0 . 002; Fisher's exact test , MMP9 , SHC1 , CDC25B and AURKB in Figures 3A–B ) . Moreover , in a genome-scale analysis , we observed a statistically significant association between the proteins that exhibit PPIN network topology , such as hub and bottleneck properties , and the number of predicted microRNA targets . Indeed , bottleneck proteins and hub-bottleneck protein of the “genome-scale PPIN” were both targeted on average by more microRNAs than those that are neither bottleneck nor hub-bottleneck ( bottleneck: P = 0 . 0009; hub-bottleneck: P = 0 . 022 , Materials and Methods ) . These results indicate that the enrichment of bottleneck and hub-bottleneck properties among miR-204 gene targets in the “prioritized HNSCC PPIN” is a system's property of microRNAs . They also suggest that the efficiency and specificity of microRNAs in regulating biological functions is further strengthened through alteration of the translation of these bottleneck proteins . In a protein-protein interaction network , proteins that are tightly linked are likely to function in the same biological process or pathways [31] , [32] . To characterize functional relationships among the 56 interacting proteins in the “prioritized HNSCC PPIN” , we conducted statistical enrichment analysis using Gene Ontology ( Materials and Methods , Protocol S1/Section C ) . The biological processes ( BP ) and molecular functions ( MF ) enriched in this network ( Figure 3C , Materials and Methods ) overlapped with our findings of functional enrichment among 34 predicted miR-204 targets ( Figures 2B–C ) . Two EGFR-dependent regulatory sub-networks were identified: cell cycle regulation and extracellular matrix ( ECM ) remodeling/Cell-matrix adhesion ( Figure 3C ) . Based on the importance of hub-bottleneck genes in regulating the function of a PPIN [33] , the enrichment of four hub-bottlenecks miR-204 targets in the EGFR-dependent “prioritized HNSCC PPIN” predicts that their up-regulation upon miR-204 suppression in HNSCC could significantly augment cell cycle and extracellular matrix remodeling . Among miR-204 gene targets that are potential regulators of cell-matrix interaction and proteolysis , overexpression of APRC1B [34] , CTSC [35] , FAP [36] , MMPs [37] , BMP1 [38] , CDH11 [39] and ITGB4 [40] is associated with cancer metastasis and/or poor prognosis . Therefore , we evaluated the role of miR-204 in HNSCC tumor progression . For these studies , we selected JSQ3 and SQ38 HNSCC cell lines for in vitro and SQ38 for in vivo characterization . The two cell lines were derived from nasal cavity and sinus HNSCC tumors , respectively ( Table 5 in Text S2 ) [26] . In vitro , ectopic restoration of miR-204 function by miR-204 mimics had no effect on the viability and proliferation of the two cell lines ( Supporting Figure 6 in Text S1 ) . In contrast , increased miR-204 function led to a significant inhibition ( P<0 . 05 ) of the ability of JSQ3 and SQ38 cells to adhere to laminin-rich basement membrane ( Figure 4A ) , to migrate through porous Transwell ( Figure 4B ) , and to invade through Matrigel-coated basement membrane ( Figure 4C ) . These results demonstrate that increased miR-204 function via its synthetic mimics is sufficient to suppress cell-matrix interaction , motility and invasiveness in vitro . To assess whether miR-204 could inhibit HNSCC tumor metastasis in vivo , we increased miR-204 function in SQ38 with miR-204 mimics treatments for three days prior to tumor transplantation . We employed an experimental model of lung metastasis by tail vein injection of tumor cells allowing characterization of tumor cell extravasation and colonization in the lung . For conducting in vivo fluorescent imaging analysis , we generated SQ38 cells stably expressing high levels of GFP fluorescent protein ( Materials and Methods ) . To initiate the study , one million of GFP-SQ38 cells transfected with either control mimics or miR-204 mimics were transplanted into athymic mice via tail-vein injection . GFP-SQ38 micrometastatic foci developed in the lung over a period of three weeks were scored lobe by lobe for each freshly isolated lung under fluorescent stereoscope ( Materials and Methods ) . Control mimics-treated SQ38 cells efficiently extravasated , established micro-metastases in 100% of animals and produced a mean number of lung metastatic foci of 75 on the whole lung surface . In drastic contrast , 50% of animals ( 7 out of 14 ) receiving miR-204 mimics treated SQ38 cells failed to develop any lung metastasis ( Figure 4D–E and not shown ) , while the other 50% of animals developed significantly less GFP-SQ38 lung foci at this early three-week time point ( P = 0 . 011 Figure 4E ) . Moreover , consistent with the predicted role of miR-204 targets AURKB and CDC25B as hub/bottleneck regulators of the cell cycle sub-network ( Figure 3 ) , restoration of miR-204 function in vivo significantly decreased the number of Ki-67 positive proliferating single SQ38 cells ( indicated by * ) and micro-foci ( indicated by arrows ) in the paraffin embedded lung sections ( P = 0 . 001 , Figure 4F ) . Moreover , Ki-67 positive SQ38 cells that received miR-204 mimics treatment were mostly single-cell foci and were in striking contrast to the multi-cell foci observed in the lungs of control mimics treatment group ( Figure 4F ) . Taken together , these observations indicate that miR-204 can significantly suppress experimental lung metastasis of SQ38 HNSCC tumors , thereby acting as a potent suppressor of metastasis . The novelty of our illustration of metastatic suppressor functions of miR-204 in head and neck cancer and its relevance to metastasis stems from our demonstration of miR-204 function at multiple scales of biology that collectively show its potential as a key regulator microRNA . Definitive demonstration of the role of miR-204 in head and neck progression requires future studies using cohorts of head and neck tumors . To explore the clinical relevance of miR-204 down regulation in HNSCC , we conducted an unbiased hierarchical clustering analysis of 60 HNSCC tumors harvested from representative anatomical sites of HNSCC in GSE686 [41] based on the mRNA expression pattern of 34 miR-204 targets identified in GSE6631 [21] ( Materials and Methods ) . The original study reported a 582-gene signature set in GSE686 that classified this set of tumors into four distinct groups: ( 1 ) an EGFR-pathway signature subtype , ( 2 ) a mesenchymal-enriched subtype , ( 3 ) a normal epithelial-like subtype , and ( 4 ) a subtype with a high level of antioxidant enzymes [41] . Hierarchical clustering using 19-upregulated genes , a subset of miR-204 targets that could be mapped to this dataset , identified two clusters ( Figure 5 ) . Tumors in Cluster A were enriched with the EGFR signature and correspond to Group 1 of the classification of Chung et al . ( P<0 . 0001 ) . In comparison , tumors in Cluster B were enriched with the Group 3 “normal epithelium-like subtype” tumors ( P<0 . 011 ) [41] . This is consistent with our observation that miR-204 targets were hub-bottleneck regulators of an EGFR-dependent regulatory network in HNSCC ( Figure 3 ) . Further , consistent with the prognostic capability of a 582-gene signature set reported by Chung et al . [41] , Cluster A showed overall earlier relapse than Cluster B ( Figure S7 ) . The fact that very comparable prognostic predictions can be derived using only 19 miR-204 gene targets suggest a potentially important role of miR-204 in HNSCC prognosis and merits further investigation and validation using a larger cohort of HNSCC tumor samples with well-characterized clinical outcomes .
Here , we developed an efficient combined computational and biological approach to predict and to prioritize cancer microRNAs for biological investigation . We demonstrated this strategy as an effective economical alternative to comprehensive microRNA analysis in cancers such as HNSCC for which prior genomic array datasets ( mRNA or microRNA ) are less abundant . This approach also allowed the identification of functional gene targets of the deregulated microRNAs that would otherwise require paired profiling of mRNA and microRNA expression for which the feasibility is often limited by the additional costs , or by the lack of access to the tissue . Employing this method that integrates the analysis of microRNA target predictions , differential HNSCC gene expression and the cancer genes in the OMIM genetic dataset , we identified and characterized miR-204 , located within its host gene TRPM3 at the 9q21 . 1–q22 . 3 region frequently incurring allelic loss [11]–[15] , as a potential tumor suppressor microRNA of HNSCC and possibly of other epithelial cancers . The high propensity of LOH at 9q21 . 1–q22 . 3 that occurs in 37% HNSCC pre-malignant conditions , further increases to 67% in cancer state [14] suggesting the presence of tumor suppressor gene candidates . While tumor suppressor genes at other frequent allelic loss loci in HNSCC have been identified , gene candidates responsible for the tumor suppressor activity associated with the 9q21 locus remain elusive . Here , we provided a plausible mechanism that loss of tumor suppressor function of miR-204 as a result of allelic imbalance at 9q21 . 1–q22 . 3 may significantly increases the genetic susceptibility to HNSCC oncogenesis and progression . LOH at this locus is also seen in the squamous cell carcinoma ( SCC ) of the esophagus [42] and SCC of the lung [43] suggesting a common somatic genetic lesion underlies the development of SCC of diverse tissue origin . The highly coordinated and nearly complete suppression of miR-204 and its host gene TRPM3 ( Figure 1B–E ) raises the possibility that TRPM3 mRNA expression may serve as a marker to indicate miR-204 expression status in HNSCC or other tumors , and also potentially LOH at 9q21 . 1–q22 . 3 . Since a small variation in the expression of a specific microRNA is expected to affect the expression of tens or hundreds of target mRNAs , genetic variations in a microRNA expression at the chromosomal break point , as we observed with miR-204 at the 9q21 . 1–q22 . 3 locus , could represent an effective mechanism of cancer predisposition , a hypothesis that is supported by emerging experimental evidences [44] , [45] . A few recent studies have reported genome-wide microRNA expression changes using HNSCC cancer cell lines [46]–[49] or tumor tissues [49]–[51] . While similar miR-204 downregualtion was reported in head and neck cancer cell lines based on microarray analysis [46] , [48] , its expression status was not further confirmed by PCR or other methods and its biological functions were not explored . Additionally , since its identification [24] biological characterization of miR-204 functions in normal development remain limited . Thus far , miR-204 was implicated in affecting global mRNA expression levels in the retina [52]; and was shown to regulate mesenchymal progenitor cell differentiation [53] . Through enrichment analysis and network modeling using mRNA gene expression profile , we identified a set of functionally related miR-204 targets that showed increased mRNA expression in HNSCC upon miR-204 suppression ( Figures 2A–C ) . The presence of miR-204 binding sites ( Table 8 in Text S2 ) , the coordinated up-regulation and the ability of increased miR-204 function to specifically inhibit the expression of 18 out of 21 gene targets ( 86% ) ( Figure 2D , Supporting Figures 4–5 in Text S1 ) suggest that these predicted genes are very likely selective and direct miR-204 targets in HNSCC . This finding is consistent with the genome-wide association between microRNA binding sites and the ability of corresponding targeting microRNAs to alter their gene expression [54] . This is the first report of a large set of functionally related cancer microRNA targets that was identified via high throughput computational approaches and confirmed biologically . In addition , the joint analyses of sequence-base information and mRNA expression arrays yielded an accuracy rate of 86% of miR-204 target predictions which surpasses the published accuracy ( about 40% ) of each sequence-based method when used alone [10] , [55] , [56] . More broadly , we demonstrated a computational framework for predicting altered regulatory networks and biological functions associated with differentially expressed microRNA targets . Indeed , our combined systems biology approach uncovered previously unknown connections between microRNA regulation , network topology , and expression dynamics for which we obtained thorough biological validations . While genome-scale analyses of interactions among microRNA gene targets in the context of a cellular or protein-protein interaction networks have been conducted computationally [57]–[59] , such methods and observations await biological confirmation . Here we significantly extended the observations of two recent reports on network modeling [31] , [32] and demonstrated the feasibility and validity of deploying statistical and bioinformatics approaches to derive regulatory networks corresponding to altered expression of proteins targeted by microRNAs ( Figure 3 ) . Further , combining functional enrichment analysis with network modeling leads to the unbiased prioritization of an EGFR-dependent protein regulatory network connected via up-regulated gene targets of microRNAs in human HNSCC ( Figure 3C ) . Topological analyses of hub and bottleneck properties further identified key regulatory proteins within the EGFR network ( Figures 3A–B ) . miR-204 appeared critical to regulate the function of this “prioritized HNSCC PPIN” as its gene targets exhibited significant enrichment of hub and bottleneck properties ( Figures 3A–B ) . Since the EGFR network was derived from overexpressed genes in HNSCC , the functional enrichment of its 56 proteins suggests their positive regulation of cell cycle , cell/matrix adhesion and extracellular matrix modeling . Using this approach , the biological effect of altering the function of a specific microRNA , such as miR-204 , can be accurately predicted via its gene targets that are key regulators of a protein network . Accordingly , enhancement of miR-204 function inhibited the expression of its functionally related gene targets ( Figure 2D , Supporting Figures 4–5 in Text S1 ) in the “prioritized HNSCC PPIN” and lead to the reduced adhesion , migration and invasion in vitro ( Figures 4A–C ) and experimental lung metastasis in vivo ( Figures 4D–F ) . Further , the strong association of overexpression of functional miR-204 gene targets with an earlier relapse in a sub-type of HNSCC tumors expressing an EGFR-pathway signature ( Figure 5 ) suggests that miR-204 expression and its deregulated gene targets could be potentially used for mechanism-based prognostic stratification of HNSCC patients to complement the conventional clinical-pathological tumor diagnosis . In fact , the feasibility of employing microRNA as sensitive and informative biomarkers for molecular diagnosis has recently been demonstrated [60] . Collectively , these findings show that single protein network modeling and statistical functional enrichment of a PPIN can illuminate altered complex biological processes and regulatory pathways associated with microRNA dysfunction in cancer with high precision . Complementary approaches have been developed to analyze gene expression changes in the molecular and biological context for candidate gene prioritization and for deriving mechanistic understandings that are most relevant to cancer biology [61]–[64] . The system's properties and microRNA-regulated molecular networks we discovered could be exploited for the design of “network mechanism”-based therapies to specifically restore tumor suppressor microRNA functions as an alternative to the single-gene target paradigm and merits further investigation .
All animal works have been conducted according to IACUC guidelines and were approved at the IACUC committee at the University of Chicago . All research involving human participants have been approved by the authors' institutional review board . Informed consent has been obtained . Microarray datasets were downloaded from NCBI GEO database . The . cel file of HNSCC mRNA transcription array sets GSE6631 [21] and GSE2379 [22] were processed using the Bioconductor Package [65] implementation of GCRMA in R Software [66] . To identify differentially expressed genes , SAM analysis [67] was performed using paired T-test between the HNSCC tumor and its corresponding paired normal tissue obtained from the same patient . The criteria for gene selection were fold change ≥2 and False Discovery Rate ( FDR ) ≤0 . 0006 ( Figure1A and Supporting Figure 3 in Text S1 ) . The association of miR-204 targets with clinical parameters was analyzed using HNSCC mRNA array set GSE686 [41] . The intensity ratios of red to green channel of the predicted miR204 targets were retrieved from GSE686 dataset . Missing values were assigned a constant value of 0 . Redundant probes representing an identical gene were reduced to a single one using the mean expression value . The miR-204 targets predicted in Figure 2A and filtered by coefficient of variation >0 . 3 were used for hierarchical clustering . In Figure 5 , the two-way hierarchical clustering was conducted with the dChip software using its default parameters ( distance metric: 1-Pearson correlation; centroid linkage clustering ) [68] , while the significance of the association between the hierarchical clusters and molecular groups of HNSCC samples [41] was determined by two-tailed Fisher's exact test adjusted with Bonferroni correction . The sample information file was obtained from the Table S1 of Chung et al [41] . The time to recurrence ( termed relapse time ) , shown in Supporting Figure 7 in Text S1 , was analyzed with the Kaplan-Myer method using the Logrank test of GraphPad Prism software ( version 4 ) [69] , and right censoring was conducted for subjects alive at the end of the study ( subjects identified by “*” in Figure 5 ) . To determine the miR-204 expression status in epithelial tumors ( Figure 1G ) , the expression values of miR-204 were extracted from microRNA array set GSE2564 [18] . Only six solid tumor types , colon , kidney , prostate , uterus , lung and breast that contained more than one samples in both tumor and the respective norm tissue were included in the analysis . Comparisons between tumors and their respective normal tissues were performed by unpaired two-tail t-test with unequal variances . MicroRNAs most likely to regulate a large number of specific inheritable cancer genes were predicted using an enrichment statistics . The Online Mendelian Inheritance in Man ( OMIM ) [76] is a semistructure database in which we computationally coded cancer genes ( OncoMIM , Protocol S1/Section A ) to a clinical nomenclature and mined with statistical enrichment to predict microRNAs that could deregulate a large number of genes , each associated with a certain type of cancer . OncoMIM contains 610 biologically validated or clinically demonstrated inheritable cancer genes among which 586 ( 96% ) are predicted targets of 527 microRNAs in the miRNOME , from which we can calculate significantly enriched microRNAs . The cumulative hypergeometric distribution ( Equation 4 ) was applied to identify significantly enriched microRNAs . We calculated the P-values based on the Equation 4 with the following variables: N is the number of OMIM genes also found in the miRNOME ( 3232 for anatomy , 2181 for disease ) , M is the number of genes associated to a specific cancer term in OncoMIM and also targeted by any microRNAs in the miRNOME , n represents the number of genes targeted by a specific microRNA in the miRNOME and also found in OncoMIM associated to any cancer term , m is the number of genes associated to both a specific cancer term in OncoMIM and to a specific microRNA in the miRNOME ( m = M∩n ) . ( 4 ) To control p in Equation 4 for multiple comparisons , we applied the Bonferroni-type adjustment method known as Šidák single-step adjusted P-value for multiple comparisons ( Equation 5 ) [77] . Significant correlations are first refined to remove false positive signals inherited in the hierarchies of the clinical nomenclature ( Protocol S1/Section B ) and then adjusted P-values ( p′ ) are less than 0 . 05 ( n = number of comparisons , p taken from Equation 4 ) ( 5 ) We developed an algorithm to identify and filter out false positive P-values derived from enrichment studies in ontologies ( hierarchical classifications ) due to the inheritance of genes in ancestry classes of a significantly enriched class [78] , [79] ( Supporting Figure 8 in Text S1 , Protocol S1/ection B ) . A gold standard of microRNAs deregulated in cancers was derived from the literature and was used to evaluate microRNA predictions in from OMIM cancer genes ( Supporting Figure 3 in Text S1 ) . To provide insights into biological functions and processes potentially regulated by miR-204 in HNSCC , we conducted standard statistical enrichment analyses based on the functional assignments of gene in Gene Ontology ( GO ) [28] to infer significantly deregulated functions associated with altered miR 204 target expression in the HNSCC according to their presence in the miRNOME and/or the PPINs ( details in Protocol S1/Section C ) . We previously established 10 low passage human head and neck squamous cell carcinoma lines ( HNSCC ) ( SCC25 , SCC35 , SCC58 , SCC61 , SCC135 , SCC151 , SQ20B , SQ38 , and JSQ3 ) , from head and neck tumor specimens of different head and neck primary sites [26] . This panel of cell lines was established from head and neck tumor specimens of different primary sites and most of the patients quickly developed local failure and eventually died of the disease [26] . Nu61 was derived from SCC61 tumors that developed radioresistance after serial passage and radiation treatment in vivo [85] . All cell lines were cultured and maintained in 1∶1 DME/F12 supplemented with high glucose and 10% fetal bovine serum . GFP-SQ38 cells were established via retroviral-mediated gene transfer using pLEGFP-N1 retroviral vector ( Clontech ) . Total RNA from normal and tumor tissues of esophagus , lung and cervix were obtained from the Ambion FirstChoice collection of RNA that is compatible with both mRNA and microRNA analysis . Total RNA from HNSCC cell lines , tumors and normal tissues was extracted and purified using TRIzol ( Gibco/BRL ) according to manufacturer's instructions . Tissues from primary HNSCC tumors were obtained from surgical procedures performed at our institution . Samples were snap frozen immediately in liquid nitrogen and stored at −80 °C . Laser micro-dissection was performed on frozen sections and approximately 10 , 000 cells were captured for RNA extraction . Normal buccal mucosa was obtained from healthy volunteers with no history of smoking and drinking according to an approved open IRB protocol . miR-204 expression was measured using TaqMan MicroRNA quantitative PCR ( qPCR ) assay ( Applied Biosystems ) according to manufacturer's instructions . . Real-time PCR was carried out using the Applied Biosystems 7900 Sequence Detector System ( Applied Biosystems ) . All qPCR reactions were run in triplicate . Human TATA-binding protein ( TBP ) ( Applied Biosystems ) was used as an endogenous control for miR-204 expression normalization . The fold changes of miR-204 expression between normal and tumor tissues or cell lines were calculated using the ΔΔCt method of relative comparison . For mRNA expression quantification , First-strand cDNA synthesis was carried out as above described except that random primers were used for reverse transcription ( High Capacity cDNA Reverse Transcription Kit , Cat#4368814 ) . Amplification of predicted miR-204 targeted genes was performed by Sybr Green qPCR assays using custom designed primers . Specific primers for each gene were designed using Invitrogen D-LUX Designer ( https://orf . invitrogen . com/lux/ ) and sequences provided in Table 13 in Text S2 . The mean Ct ( cycle threshold ) was calculated from the triplicates and used for the calculation of RQ values . qPCR condition for each gene was optimized that so that the standard error among the triplicates was <0 . 15 Ct . TBP was also used as endogenous control for data normalization . The fold changes of target gene were calculated using the ΔΔCt method of relative comparison . In addition , as negative controls for the off target effect of miR-204 mimics treatment , real time qPCR was performed to include three additional endogenous controls: PPIA ( AB , Cat#4333763 ) , GUSB ( AB , Cat# 4333767 ) and , HPRT1 ( Cat#4333768 ) using commercially designed Taqman gene expression assays ( Applied Biosystems ) . Quantitative mRNA expression data were acquired and analyzed in either 96- or 384-well-plate format using an Applied Biosystems 7900 Sequence Detector System ( Applied Biosystems ) . 40% confluent JSQ3 and SQ38 cells were transfected with 50–200 nM Control [Cat#110CN-001000-01] or miR-204 miRIDIAN mimics [Cat#110C-300069-02] ( Dharmacon ) using Oligofectamine ( Invitrogen ) . Transfection efficiency was optimized and estimated to be >90% . Proliferation assay , cell adhesion assay . Migration assay and Matrigel invasion assay were conducted at 72 hours after transfection . In vivo tail-vein injection of mimics treated GFP-SQ38 cells was performed at 48h after transfection . Cell adhesion was measured using the InnoCyte ECM cell adhesion assay kit ( Calbiochem , Cat#CBA025 ) according to manufacturer's instructions . Control or miR-204 miRIDIAN mimics treated JSQ3 and SQ38 cells were trypsinized and re-suspended in fully supplemented medium . 20 , 000 cells and 15 , 000 cells were added to each well for JSQ3 and SQ38 cell lines , respectively . Cells were incubated for 2h at 37 °C . The plates were then washed with PBS to remove non-adherent cells . 100 µl Calcein-AM was added to each well , incubated with cells for 1h at 37 °C , and read with a fluorescent plate reader at an excitation wavelength of ∼485 nm and an emission wavelength ∼520 nm . Results were expressed as percent of cell adhesion compared to that of control mimics treated controls±standard error ( SE ) of 3 replicates . Control or miR-204 miRIDIAN mimics treated JSQ3 and SQ38 cells were trypsinized and re-suspended in fully supplemented medium . Cells were then seeded at 10 , 000 cells per well for migration assay or at 20 , 000 cells per well for invasion assay into trans-well inserts ( 8 µm pore size , BD Falcon ) . For invasion assay , the trans-well inserts were coated with 60 µg/45 µl/well of Matrigel ( BD Falcon ) . Complete culture medium was used as chemo-attractant in the lower chamber . The assays were taken down with three PBS washes followed by fixation with 10% formalin and staining with 1% crystal violet after 6h for migration assay and 18h for invasion assay . The cells migrated to the basal side of the porous membrane was visualized with a Zeiss Axiovert microscope at ×20 magnification . 10 random fields from three replicate wells were counted and the number of cells that had migrated or invaded was presented as number of cells counted per field of the porous membrane . Cell proliferation assays were conducted in 96-well format by the MTT assay . Specifically , HNSCC cell lines were seeded at 5×103 cells/well in 96-well plates and let incubated for 24 hours prior to treatment with control or miR-204 miRIDIAN mimics . After drug or siRNA exposure , 10 µl of MTT reagent was added to each well and incubated for 4h . The precipitates were dissolved in 100 µl of stop solution overnight and proliferation rate was determined by absorbance at 570 nm wavelengths with 690 nm as the reference wavelength using a spectrophotometer . Animal work was conducted in accordance with an approved protocol . Age and weight-matched ( 4–6 weeks old weigh 18–20 g ) NCI athymic female mice were used for induction of experimental lung metastasis via the tail-vein injection of tumor cells . GFP-SQ38 cells were treated with miR-204 miRIDIAN mimics or non-specific control mimics for 2 days prior to tumor cell inoculation . 1×106 viable cells were re-suspended in 100 µl of PBS and injected into the lateral tail vein . Metastatic colonization of lung by GFP-SQ-38 cells was determined at 3 weeks post tumor injection . 28 Mice were sacrificed on day 21 after tumor cell inoculation . Lungs were perfused through tracheal with 2–3 ml of PBS , excised and then fixed in 10% formalin for 12 hours . Prior to fixation with formalin , lungs were examined under 4× magnification using fluorescent stereoscope ( Leica ) and scored lobe by lobe for GFP-SQ38 lung foci on the whole lung surface . Thereafter , the University of Chicago Immunohistochemistry Core Facility performed paraffin embedding , sectioning and H and E staining . 5-micron sections were stained with Ki-67 and Ki-67 positive SQ-38 cells or micro-foci were scored under 40× magnifications in 10 randomly selected fields for each section . A total of 6 lungs from each treatment group were examined . | MicroRNAs regulate the expression of genes in cells and are important in cancer development and progression . Designing new microRNA-based treatments requires the understanding of their mechanisms of action . Previous biological studies lack in depth since only a few genes are confirmed as microRNA targets . Additionally , key biological systems perturbed by altered microRNA functions in the context of cancer remain to be identified . Here , we demonstrate for the first time how genetic knowledge about the inheritance of cancer can be utilized jointly with data about the expression of genes in cancer samples to model deregulated microRNAs and their functions at multiple scales of biology . Our approach further uncovers previously unknown connections between microRNAs , their regulated genes , and their dynamics . Using head and neck cancer as a model , we predict the presence , functions , and gene targets of a new tumor suppressor microRNA in a cancer-associated chromosomal region where a candidate gene has not been identified . We then confirm their validity with extensive and thorough biological characterization and show attenuation of lung metastasis in mice . The discovery of molecular networks regulated by microRNAs could be exploited for the design of new treatments as an alternative to the single-gene target paradigm . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"genetics",
"and",
"genomics/genomics",
"genetics",
"and",
"genomics/gene",
"expression",
"genetics",
"and",
"genomics/disease",
"models",
"molecular",
"biology/translational",
"regulation",
"oncology/head",
"and",
"neck",
"cancers",
"genetics",
"and",
"genomics/bioinformatic... | 2010 | Network Modeling Identifies Molecular Functions Targeted by miR-204 to Suppress Head and Neck Tumor Metastasis |
Autophagy is a cellular process required for the removal of aged organelles and cytosolic components through lysosomal degradation . All types of eukaryotic cells from yeasts to mammalian cells have the machinery to activate autophagy as a result of many physiological and pathological situations . The most frequent stimulus of autophagy is starvation and the result , in this case , is the fast generation of utilizable food ( e . g . amino acids and basic nutrients ) to maintain the vital biological processes . In some organisms , starvation also triggers other associated processes such as differentiation . The protozoan parasite Trypanosoma cruzi undergoes a series of differentiation processes throughout its complex life cycle . Although not all autophagic genes have been identified in the T . cruzi genome , previous works have demonstrated the presence of essential autophagic-related proteins . Under starvation conditions , TcAtg8 , which is the parasite homolog of Atg8/LC3 in other organisms , is located in autophagosome-like vesicles . In this work , we have characterized the autophagic pathway during T . cruzi differentiation from the epimastigote to metacyclic trypomastigote form , a process called metacyclogenesis . We demonstrated that autophagy is stimulated during metacyclogenesis and that the induction of autophagy promotes this process . Moreover , with exception of bafilomycin , other classical autophagy modulators have similar effects on T . cruzi autophagy . We also showed that spermidine and related polyamines can positively regulate parasite autophagy and differentiation . We concluded that both polyamine metabolism and autophagy are key processes during T . cruzi metacyclogenesis that could be exploited as drug targets to avoid the parasite cycle progression .
Autophagy is a major intracellular degradation/recycling system ubiquitous in eukaryotic cells . It contributes to the turnover of cellular components by delivering portions of the cytoplasm and organelles to lysosomes , where they are digested [1] . Depending on the mechanisms used for the delivery of cargo to lysosomes , three different types of autophagy have been described in mammalian cells: macroautophagy , microautophagy , and chaperone-mediated autophagy ( CMA ) [2] . Macroautophagy , referred to as autophagy in the rest of this work , involves a first step of autophagosome formation followed by autophagosome maturation . Initially , the cytoplasmic materials are sequestered by the phagophore , a curved membrane that elongates around the cargo to form a double membrane vesicle called autophagosome . Autophagosomes next interact with endocytic compartments and finally fuse with lysosomes to form autolysosomes where the enclosed materials are hydrolyzed [1] . Several genes required for autophagy have been described . Their products , the so-called Autophagy ( Atg ) -related proteins , comprise the core molecular machinery responsible for the sequential activation of this pathway [3] . The Atg8 protein ( or LC3 in mammalian cells ) , is the best marker of autophagy . Atg8 is present in the membrane of all compartments of this pathway , from the phagophore to the autolysosome [4] . The formation of autophagosomes and execution of autophagy critically depend on proteolytic processing of Atg8 by the cysteine protease Atg4 , and its subsequent conjugation to the phosphatidylethanolamine in the expanding phagophore membrane [5] . It is known that two major kinases differentially regulate mammalian autophagy: the mammalian target of rapamycin ( mTOR ) and the class III PI3K Vps34 . mTOR is an evolutionary conserved kinase that senses the nutrient and energy status of cells by forming two distinct complexes . One of them , mTORC1 enhances glycolysis and biosynthetic processes and inhibits autophagy [6] . Therefore , inhibition of mTORC1 by treatment with rapamycin ( Rap ) , an immunosuppressive drug , results in a potent induction of autophagy . In contrast , the activity of Vps34 is essential for autophagy . In mammalian cells Vps34 forms a complex with beclin-1 ( the mammalian ortholog of yeast Atg6 ) and other proteins to promote the production of phosphatidylinositol 3-phosphate , thereby facilitating lipid membrane changes required for autophagosome formation and maturation [7] . The PI3K inhibitor wortmannin ( Wort ) has been widely used to inhibit yeast and mammalian autophagy for its inhibitory action on the beclin-1/Vps34 complex [8] . The polyamine spermidine ( Spd ) has been recently described as a new modulator of autophagy since Spd inhibits the activity of histone acetyl transferase , leading to the upregulation of several ATG genes including ATG7 , ATG11 and ATG15 [9] . When added to culture media , Spd is also able to directly induce autophagy in a transcription-independent manner . The mechanism has not been fully elucidated yet; however , this phenomenon could be due to the enhanced deacetylation of essential autophagy-related proteins such as ATG5 and ATG7 [10] . Furthermore , the same concentrations of Spd that exert proautophagic effects also have a marked life span-extending action on yeast , nematodes and flies . Conversely , the genetic inhibition of essential ATG genes abrogates the life span extension induced by Spd , indicating that this polyamine can prolong the life span by the induction of autophagy [9] . The parasitic protozoan Trypanosoma cruzi , which is the causative agent of Chagas’ disease , presents four well differentiated stages in its complex life cycle , which alternates between insect vectors and mammalian hosts . The bloodsucking triatomine bugs acquire the parasites by ingestion during a blood meal of an infected mammalian host . A few hours after the meal , in the anterior region of the midgut , bloodstream trypomastigotes transform into proliferative , non-infective epimastigotes . After several rounds of replication , epimastigotes transform into the non-proliferative , infective metacyclic trypomastigotes ( MT ) , a process called metacyclogenesis . MT are released along with the feces and urine of the insect and may infect a new mammalian host . Firstly , MT infect macrophages and epithelial cells in the site of entry and then cardiac and smooth muscle fibers that are the major targets of T . cruzi . The infection of these cells is responsible for the main clinical manifestations of the disease . Inside the cell , the parasite undergoes another dramatic transformation into proliferative intracellular amastigotes . After intense multiplication in the host cell cytoplasm , amastigotes transform into bloodstream trypomastigotes that can infect other neighboring cells or reach the circulatory system , thus completing the cycle [11] . Autophagy also occurs in trypanosomatid parasites . Half of the known yeast and mammalian ATG proteins have also been found in vertebrate pathogenic trypanosomatids ( Trypanosoma brucei , Trypanosoma cruzi and Leishmania spp . ) , although with low sequence conservation [12] . More than one ATG8 gene was identified in the Trypanosomatidae family: three in T . brucei , two in T . cruzi and , unexpectedly , four families comprising together 25 genes in Leishmania major . T . cruzi has a ‘true’ TcATG8 . 1 and a TcATG8 . 2 , which does not seem to participate in autophagy [13] . T . cruzi also contains two ATG4 isoforms , TcATG4 . 1 and TcATG4 . 2 whose products , which are called autophagins , are in charge of ATG8 . 1 processing [14] . The cellular remodeling during differentiation is essential for the progression of the life cycle of many unicellular eukaryotic pathogens such as Leishmania spp . [15] , and T . cruzi [16]; however , the mechanisms involved in these processes have not been fully characterized . The first morphological indications that autophagy occurs during differentiation of trypanosomatids were provided by electron microscopy images of T . brucei taken by Vickerman and colleagues in the 1970s [17] . Long after that , molecular studies corroborated the presence of ATG genes in trypanosomatids and their participation in the differentiation processes [12 , 13 , 18] . Although ultrastructural and molecular approaches have demonstrated the existence of autophagosomes in T . cruzi [13] , a functional characterization of this pathway is still lacking . Metacyclogenesis of T . cruzi takes place in the insect's rectum due to many factors such as nutrient scarcity produced by the fast replication of epimastigotes , specific components of intestinal wall and lumen of the vector , etc . The in vitro stimulation of this process has been achieved in aged cultures of epimastigotes [19] or by thermic and nutrient stress [20 , 21] as explained below . Regardless of the method , nutrient deprivation , the classical inducer of autophagy , is the common most frequent stimulus required to trigger epimastigote differentiation . In this report , we made a functional analysis of T . cruzi autophagy under different conditions . We have observed that autophagy is induced during T . cruzi metacyclogenesis and that different drugs known to regulate mammalian autophagy exert either a positive or a negative effect on parasite autophagy and differentiation . We have also shown that similarly to mammalian cells , spermidine and related polyamines are important inducers of parasite autophagy and that mutant parasites that produce their own polyamines display higher autophagic activity and higher metacyclogenesis efficiency , as compared to wild type parasites . Taken together , these data demonstrate the key role of autophagy for T . cruzi differentiation and highlight a new target that could be used to interrupt the T . cruzi cycle progression .
We are cognizant of the Argentinean ( ANMAT 5330/97 ) and international ( Declaration of Helsinki ) principles and bioethical codes , and guarantee that all procedures carried out in conducting the research reported here were in compliance with both . Human subjects were involved in this project for the purpose of sera donation . The subject population consisted of healthy male donors , 25 year of age or over , which signed a written Informed Consent form at the time of their enrollment . The Research Committee of the Central Hospital of Mendoza and the Bioethical Committee of the Diego Paroissien Hospital of Mendoza ( Comité de Investigación del Hospital Central de Mendoza , President: Dr Carlos Zanessi y Comité de Bioética del Hospital Diego Paroissien de Mendoza; President: Dr Jorge Sotile ) approved our protocol for the collection and manipulation of human serum samples . All laboratory procedures followed the safety regulations of the Hospitals and Medical School . TAU medium was prepared with 190 mM NaCl ( Biopack ) , 17 mM KCl ( Biopack ) , 2 mM MgCl2 ( Biopack ) , 2 mM CaCl2 ( Biopack ) and 8 mM sodium phosphate buffer ( pH 6 to 6 . 8 ) . Modified TAU medium ( TAU-AAG ) was prepared with TAU medium supplemented with 50 mM sodium glutamate ( Sigma ) , 10 mM L-proline ( Tetrahedron ) , 2 mM sodium aspartate ( Sigma ) , and 10 mM glucose ( Biopack ) . Diamond medium contains 6 . 25 g/l tryptose ( Sigma ) , 6 . 25 g/l tryptone ( Sigma ) , 6 . 25 g/l yeast extract ( Sigma ) , 7 . 16 g/l KH2PO4 ( Biopack ) ( pH 7 . 2 ) and 6 . 66 mM hemin ( Calbiochem ) prepared in 3 ml 1N NaOH ( Tetrahedron ) and 20 ml 1M Tris HCl ( Tetrahedron ) ( pH 6 . 8 ) . BHT medium was prepared with 33 g/l Brain heart infusion broth ( Britania ) , 3 g/l tryptose , 0 . 4 g/l KCl , 0 . 3 g/l glucose and 3 . 2 g/l Na2HPO4 ( Biopack ) . SDM79 medium , which contains only traces of polyamines , was prepared with 8 . 4 g/l 199 TC 45 medium ( Sigma ) , 8 ml/l MEM amino acids 50x ( Gibco ) , s/c L-glutamine ( Carbiochem ) , 6 ml/l MEM Non-essential amino acids 100x ( Gibco ) , 1 g/l glucose , 8 g/l HEPES ( Carbiochem ) , 5 g/l MOPS ( Carbiochem ) , 2 g/l NaHCO3 ( Biopack ) , 100 mg/l sodium pyruvate ( Sigma ) , 200 mg/l L-alanine ( Tetrahedron ) , 100 mg/l L-arginine ( Sigma ) , 300 mg/l L-glutamine ( Sigma ) , 70 mg/l L-methionine ( Sigma ) , 80 mg/l L-phenylalanine ( Sigma ) , 600 mg/l L-proline ( Sigma ) , 60 mg/l L-serine ( Tetrahedrum ) , 160 mg/l L-taurine ( Sigma ) , 350 mg/l L-threonine ( Sigma ) , 100 mg/l L-tyrosine ( Sigma ) , 10 mg/l adenosine ( Sigma ) , 10 mg/l guanosine ( Sigma ) , 50 mg/l glucosamine-HCl ( Sigma ) , 4 mg/l folic acid ( Sigma ) , ( pH 7 , 3 ) . Epimastigotes of Y or Y-GFP strain were cultured in Diamond medium with 10% fetal bovine serum ( Natocor ) at 28°C . Y-GFP-ODC [22] and Y-GFP-PAT12 [23] mutants co-expressing GFP and the ornithine decarboxylase gene ( ODC ) ( AN Y08233 . 1 ) or the PA transporter PAT12 ( AN AY526253 , also annotated as FJ204167 ) respectively were maintained in the semisynthetic medium SDM79 , to select auxotrophy at 28°C . All cultures contain 20 mg/l hemin ( Calbiochem ) , 10% inactivated fetal bovine serum , 250 μg/ml geneticin ( Gibco ) for GFP selection , 100 mg/ml streptomycin ( Gibco ) and 100 U/ml penicillin ( Gibco ) . To induce T . cruzi metacyclogenesis we performed a previously published in vitro protocol schematized in S1 Fig [20 , 21] . Briefly , epimastigotes of T . cruzi Y or Y-GFP strain ( or the mutants Y-GFP-ODC or Y-PAT12 ) grown to stationary phase ( 5 x 107 cells/ml ) were collected by centrifugation at 2000 g for 15 min , and resuspended at 5 x 108 cells/ml in TAU medium . After 2 h at 37°C ( 1st stage of metacyclogenesis ) , parasite samples were processed for microscopy or molecular studies . Similar procedures were conducted in control parasites maintained in Diamond , BHT or SDM79 medium at 28°C . In other cases , to complete the differentiation process , parasites were diluted 100 times in TAU-AAG or control media and maintained at 28°C for 48 h ( 2nd stage of metacyclogenesis ) . After this period , differentiated parasites ( MT ) were then directly quantified using human fresh serum or used for infection assays ( see below ) . In some experiments , TAU medium was supplemented with 100 nM wortmannin ( Wort , Sigma-Aldrich ) , 100 nM bafilomycin ( Baf , Sigma-Aldrich ) or 1 mM difluoromethylornithine ( DFMO , Sigma ) as autophagy inhibitors . For autophagy induction , 50 ng/μl rapamycin ( Rap , LC Laboratories ) , 100 μM spermidine ( Spd , Sigma ) or 100 μM spermine ( Spm , Sigma ) was added to control media . After the first period of metacyclogenesis parasites were stained with the Trypan blue vital dye to study parasite viability . Control and TAU samples were deposited on coverslips and stained parasites were counted by conventional microscopy . An aliquot of parasites were exposed to UV and used as a positive control of mortality . To study autophagic activity parasites were subjected to the first period of metacyclogenesis and processed to detect autophagosomes by indirect immunofluorescence with a specific antibody against the TcAtg8 . 1 protein ( AN ABH07412 ) generously given by Dr . Vanina Alvarez ( IIB-INTECH UNSAM-CONICET ) . Briefly , parasites were fixed with 4% paraformaldehyde ( Sigma-Aldrich ) solution in PBS for 15 min at room temperature , washed with PBS , and quenched with 50 mM NH4Cl ( Merck ) for 15 min at room temperature . Subsequently , cells were permeabilized with 1% saponin ( Sigma-Aldrich ) in PBS containing 1% bovine serum albumin ( BSA-Sigma ) , and then incubated with the primary antibody against Atg8 . 1 ( 1:500 ) followed by incubation with Cy-3 ( excitation wave: 550 nm and emission wave: 570 nm , ThermoFisher ) or Alexa 488 ( excitation wave: 490 nm and emission wave: 525 nm , ThermoFisher ) conjugated anti-rabbit ( 1:500 ) secondary antibodies . After that parasites were mounted on coverslips with mowiol 4–88 reagent ( Calbiochem ) and examined by confocal microscopy . For colocalization studies , after detection of Atg8 . 1 , parasites were incubated with 10 μg/ml of dequenched BSA ( red DQ-BSA , excitation wave: 590 nm and emission wave: 620 nm , Invitrogen ) for 2 h , washed three times with PBS and then mounted on coverslips with mowiol before examination . Epimastigotes were exposed to the first period of metacyclogenesis in Diamond ( control ) or TAU ( starved ) medium , during 2 h at 37°C and then fixed and processed by Electron Microscopy . Briefly , parasites were fixed with 2% glutaraldehyde ( Ted Pella ) in PBS for 2 h at 4°C , washed three times with PBS pH 7 . 2 and subsequently treated with 1% osmium tetroxide ( Ted Pella ) for 2 h at 4°C . In a next step , parasites were washed again with PBS and sequentially dehydrated in solutions with increasing concentrations of acetone . Finally , samples were included in the epoxy resin ( Spurr ) and ultrathin sections in an ultramicrotome Leica Ultracut R were performed . Sections were contrasted with uranyl acetate / acetone for 3 min , washed with distilled water and colored with lead citrate for 2 min before observation with the Zeiss 900 electron microscope . T . cruzi epimastigotes were subjected to the first period of metacyclogenesis in Diamond ( control ) or TAU ( starved ) medium during 2 h at 37°C and processed for molecular studies . Parasites were collected and washed and RNA was obtained by TRIzol reagent ( ThermoFisher ) . RT step was performed with Oligo ( dT ) 15 Primer and M-MLV Reverse Transcriptase according to the manufacturer instructions ( Promega ) . Levels of expression of TcAtg8 . 1 were determined by PCR assay in not saturating conditions using cDNA from both populations of metacyclogenesis induced epimastigotes . Primers 5’-CTTTGGAGCACCGCATCG-3’ ( forward ) and 5’-CAAAAGTTGCCTCACCCGAG-3’ ( reverse ) were used to amplify a fragment from the TcAtg8 . 1 transcript ( 318 pb ) ; and primers 5’- ATATTTAAACCCATCCAAAATCGAGTAAC-3’ ( forward ) and 5’- GTCAATTTCTTTAAGTTTCACTCTTGC-3’ ( reverse ) for 18S rRNA transcript ( 1029 pb ) , used as the housekeeping gene . PCR products were separated in 1% agarose gel , stained with SYBR Safe ( Invitrogen ) and quantified using ImageJ software ( http://imagej . nih . gov/ ) . The results were expressed as arbitrary units ( AU ) , normalized to rRNA levels . Data shown represents the mean from 3 independent experiments . Statistical analysis was performed using Student t-test . Epimastigotes from Y-GFP strain were subjected to the first period of metacyclogenesis in TAU ( starvation ) or BHT ( Control ) medium for 2 h at 37°C . Thirty min before the end of incubation , 0 . 15 mg/ml monodansylcadaverine ( MDC , excitation wave: 365 nm and emission wave: 525 nm , Sigma-Aldrich ) was added to samples . After that , parasites were centrifuged and washed three times with PBS . Subsequently they were deposited on coverslips previously coated with poly-L-lysine ( Merck ) and then observed in a confocal microscope Olympus FV 1000 in a thermostatized chamber . Another aliquot of parasites were processed to measure fluorescence intensity in a Multiplate reader . Data were represented using the mean values of percentage of MDC fluorescent parasites and standard errors ( SE ) of at least three independent experiments . Statistical calculations ( Tukey test , * p <0 . 05 , ** p <0 . 01 , *** p <0 . 001 ) and graphics were prepared using the software KyPlot . The method was similar to that of the previous section , with the addition of 10 μg/ml dequenched BSA instead of MDC . This compound emitted red fluorescence after BSA hydrolysis into small peptides in lysosomes , thus identifying degradative compartments . Data were represented using the mean values of percentage of DQ-BSA positive parasites and the error bars indicate SE of at least three independent experiments . Statistical calculations ( Tukey test , * p <0 . 05 , ** p <0 . 01 , *** p <0 . 001 ) and graphics were performed using the software KyPlot . Epimastigotes from Y strain ( 15 x 106 cells ) were subjected to the first period of metacyclogenesis in BHT medium ( control ) in the absence or presence of 100 μM spermidine for 2 h at 28°C or in TAU medium ( starvation ) for 2 h at 37°C . Cells were collected by centrifugation at 2000 g for 15 min , resuspended in sample buffer and incubated for 10 min at 95°C . Protein extracts were run on 18% SDS-PAGE and transferred to Hybond-ECL ( Amersham ) nitrocellulose membranes . The membranes were blocked in Blotto for 1 h at 4°C ( 10% non-fat milk , 0 . 05% Tween 80 in PBS ) , washed twice with 0 . 05% Tween 80 in PBS and incubated with a primary antibody anti-LC3 ( 1:800 dilution , Sigma-Aldrich ) followed by a peroxidase-conjugated anti-rabbit secondary antibody ( 1:10 , 000 dilution ) . Anti-Tubulin ( 1:300 dilution , Developmental Studies Hybridoma Bank ) was used to detect Tubulin ( AN ESS55047 ) as a loading control . Detection was accomplished with a chemiluminescence system from Millipore ( WBKLS , Biopore , Buenos Aires , Argentina ) on a Luminescent Image Analyzer LAS-4000 ( Fujifilm , Tokyo , Japan ) . To quantify the efficiency of metacyclogenesis at different conditions , parasites were subjected to the first ( F ) or the complete period ( T ) of metacyclogenesis in control or TAU medium in the presence of wortmannin ( in TAU medium ) or rapamycin ( in control medium ) . Subsequently , samples of mixed parasitic forms ( epimastigotes / metacyclic trypomastigotes ) were centrifuged to remove inhibitors or inducers of the autophagic pathway . Pellets were then resuspended in fresh human serum recently obtained from a healthy male donor , which produces the complement dependent lysis of epimastigotes and facilitates the direct quantification of the serum-resistant metacyclic trypomastigotes in a Neubauer chamber . Epimastigotes/ trypomastigotes mixed samples generated as in the previous section were placed on Vero cell ( ABAC-Asociación Banco Argentino de Células ) monolayers for 24 h at 37°C . After three washes with PBS , to remove non-internalized parasites , cells were fixed with 3% paraformaldehyde for 15 min at room temperature , and quenched with 50 mM NH4Cl in PBS . To facilitate visualization , cellular actin ( AN XP_008017958 ) were stained with rhodamine-conjugated phalloidin ( excitation wave: 540 nm and emission wave: 570 nm , Invitrogen ) for 1h at 37°C in a humid chamber . Parasite nuclei were visualized in green due to the stable expression of TcH2b histone fused to GFP [24] . Cells were also treated with Hoechst for DNA staining , mounted onto glass slides with Mowiol and analyzed with an Olympus Confocal Microscope FV1000-EVA ( Olympus ) , with the FV10-ASW ( version 01 . 07 . 00 . 16 ) software .
As mentioned in the Introduction , previous works have demonstrated the presence of ATG genes in T . cruzi as well as increased levels of TcAtg8 in parasites undergoing spontaneous differentiation [13] . To better characterize the participation of autophagy during this process , we followed a standardized protocol of differentiation initially published by Contreras et al . [20] and then modified by Ferrari et al [25] . Briefly , epimastigotes were subjected to a short period of nutritional and thermic stress in triatomine artificial urine ( TAU ) medium at 37°C during 2 h , followed by a chase in TAU medium supplemented with three amino acids ( indicated in the Methods section ) and glucose ( TAU-AAG ) during 48 h at 28°C ( S1A Fig ) . The efficiency of this protocol was analyzed by both direct quantification of differentiated MT and by infection assays ( see details in the Methods section ) . Data showed that a significant number of infective forms were obtained from the parasites subjected to TAU compared to parasites maintained in control conditions ( see below ) . After the first period of metacyclogenesis parasites were stained with the Trypan blue vital dye to study parasite viability . An aliquot of parasites were exposed to UV as a positive control of mortality . In contrast to the low percentage of survival displayed by UV irradiated parasites , survival of starved parasites was high and similar to the control ( S1B Fig ) . Since maximal stress conditions were produced in the first 2 h , we tested autophagic response at this point , by studying the presence of autophagosomes . Parasites were fixed and processed for IIF . Autophagosomes could be readily detected in parasites subjected to nutritional and thermic stress , as compared to parasites maintained in full-nutrient media at 28°C ( control conditions ) ( Fig 1A ) . Similar to the method described by Brady and coworkers , we quantified the number of Atg8 . 1 dots in parasites incubated under both conditions and empirically established a maximum threshold of two autophagosomes per parasite [26] . The percentage of parasites with more than two Atg8 . 1 positive vesicles was significantly higher in TAU ( 80 +/- 1 . 65% ) than in control conditions ( 32 . 45 +/- 4 . 35% ) ( Fig 1B ) . Further TEM analyses of stressed parasites revealed the presence of double membrane vesicles and multivesicular compartments that resembled typical autophagic structures ( Fig 1C ) . To confirm these observations , we studied the expression of TcAtg8 . 1 by semi-quantitative RT-PCR using cDNA obtained from parasites maintained under control or TAU conditions . After data normalization to 18S rRNA , we observed that starved parasites presented a significant increase in the levels of TcAtg8 . 1 cDNA , as compared to control parasites ( Fig 1D ) . Apart from the autophagosome increase , the induction of autophagy in mammalian cells is characterized by an increase in the number of lysosomes/autolysosomes required for the lysis of trapped components [27] . Therefore , we used monodansylcadaverine ( MDC ) and the self-quenched albumin ( DQ-BSA ) , which are markers of acidic and hydrolytic compartments respectively , to localize lysosomes in T . cruzi . As shown in Fig 2A and 2B epimastigotes maintained under nutrient-rich conditions displayed lower frequency of MDC or DQ-BSA labeling . In contrast , epimastigotes subjected to nutritional and thermic stress in TAU differentiation medium shown a significant increase in the number of lysosomes , as compared to controls ( Fig 2C and 2D ) . Similar differences were observed in the detection of MDC fluorescent intensity associated to parasites under each condition ( S2 Fig ) . In another set of experiments , we studied the colocalization of DQ-BSA and TcAtg8 . 1 vesicles . We observed that starved parasites displayed a higher frequency of colocalization of DQ-BSA with TcAtg8 . 1 , as compared to controls ( Mander´s overlap coefficient 0 . 87 +/- 0 . 03 vs . 0 . 42 +/- 0 . 05 ) , indicating fusion of autophagosomes with lysosomes to form autolysosomes ( Fig 2E ) . Taken together , these results evidence that nutritional and thermic stress conditions induce the autophagic pathway during the T . cruzi metacyclogenesis . Next , we studied the possible participation of mTOR and Vps34 kinases on T . cruzi autophagy . As shown in Fig 3A , the treatment of parasites with 50 ng/μl Rap under control conditions ( Diamond media at 28°C ) for 2 h , induced a significant increase in the percentage of parasites with more than two Atg8 . 1 positive vesicles , as compared to non-treated parasites . This increment in the autophagic response was similar to that obtained in the first period of differentiation in parasites incubated in TAU medium . Conversely , treatment with 100 nM Wort of parasites exposed to differentiation conditions impaired the autophagic response ( Fig 3A ) . Similar differences were observed in the content of acidic and hydrolytic vesicles detected with MDC and DQ-BSA , respectively ( Fig 3B and 3C ) . Bafilomycin ( Baf ) , which is another important autophagy inhibitor , is mainly used to study the autophagic flux [28] . The treatment of mammalian cells with Baf blocks the normal autolysosomal degradation , leading to accumulation of autophagic compartments at different maturation stages . Unexpectedly , a significant reduction in the autophagosome number was observed when parasites were incubated in TAU in the presence of Baf , as compared to TAU medium alone ( Fig 4A and 4B ) . The number of autophagosomes was lower in parasites subjected to Baf in control medium than in non-treated parasites , indicating that Baf inhibits both basal and induced autophagy . Unlike the effect observed in mammalian cells , Baf abrogated autophagosome formation in T . cruzi , resulting in a complete inhibition of the autophagic response . Polyamines ( PA ) are low molecular mass polycations that bind to acidic macromolecules such as DNA , RNA and proteins to regulate proliferation and differentiation . Ornithine decarboxylase 1 ( ODC1 ) , which is one of the rate-limiting enzymes in the polyamine biosynthetic pathway , catalyzes the conversion of L-ornithine to putrescine ( Put ) . The sequential addition of two aminopropyl groups to Put by spermidine synthase and spermine synthase generates spermidine ( Spd ) and spermine ( Spm ) , respectively . Since the role of Spd on T . cruzi autophagy has not been studied before; we analyzed the effect of Spd and also Spm in our system . Our results showed that the presence of either Spd or Spm significantly increased the detection of Atg8 . 1 ( Fig 5A ) . Quantitative data showed a significant increase of parasites with autophagic vesicles after PA treatment , as compared to control parasites ( ≈ 80% vs . ≈ 20% ) . Interestingly , the degree of autophagic response of T . cruzi in the presence of PA was similar to that obtained under TAU condition . Furthermore , addition of 1mM DFMO , which is a non-reversible inhibitor of ODC [29] , to TAU , did not modify the number of autophagosomes , as compared to TAU condition alone ( Fig 5B ) . This result was not surprising due to the lack of ODC1 in T . cruzi [30] . As demonstrated previously , T . cruzi present all components of the Atg8 conjugation system [13] . Therefore , induction of autophagy in the parasite involves the processing and lipidation of TcAtg8 . 1 to insert it in the membrane of the autophagosome . Next we performed a western blot assay to detect the two different forms ( soluble and membrane-bound ) of TcAtg8 . 1 from protein extracts obtained of parasites subjected to control or autophagic-induced conditions . As shown in Fig 5C , a double band corresponding to unprocessed TcAtg8 . 1 ( soluble ) and processed TcAtg8 . 1 ( membrane-bound ) were detected in TAU and Spd conditions while parasites maintained in control medium contain only the soluble form . Level of processed TcAtg8 . 1 normalized to tubulin was significantly increased under TAU and Spd treatments , as compared to control , thus confirming the processing of this protein when autophagy is triggered . As mentioned above , T . cruzi is unable to synthesize endogenous PA due to the lack of ODC [30] , thus relying on PA uptake from the extracellular medium by the polyamine permease TcPAT12 [23 , 30] . In our laboratory , we have previously generated a T . cruzi mutant strain coexpressing heterologous ODC and GFP ( Y-GFP-ODC ) [22] that reverted the natural polyamine auxotrophy . In this work , we have used this strain to confirm the effect of PA on parasite autophagy . To ensure ODC activity , Y-GFP-ODC parasites were maintained in the semisynthetic medium SDM79 ( containing only traces of polyamines ) . In this medium , ≈ 80% of parasites displayed more than two Atg8 . 1 positive dots ( Fig 5D and 5E ) . The number of autophagosomes increased when mutant parasites were subjected to the first period of differentiation in TAU medium at 37°C . In contrast , the treatment of parasites with 1 mM DFMO in SDM or TAU media significantly reduced the autophagic response . Taken together , these results demonstrated that PA are important inducers of autophagy in T . cruzi and that the higher basal autophagy of Y-GFP-ODC parasites was due to the activity of heterologous ODC and the increased availability of PA . Similar conclusions were obtained with a mutant strain that overexpresses the TcPAT12 transporter and displays a higher uptake of PA [31] . As expected , in SDM79 medium , only 30% of Y-GFP-PAT12 parasites displayed autophagic activity . This percentage increased significantly after addition of 100 μM Spd or 100 μM Spm to the medium or when parasites were subjected to differentiation conditions in TAU medium . Unexpectedly , the presence of DFMO reduced the basal autophagy levels in this strain , probably by an indirect effect of the drug ( S3 Fig ) . Next , we studied the effect of the modulation of autophagy on the global process of metacyclogenesis . Y-GFP strain T . cruzi epimastigotes were subjected to the complete differentiation process in either the presence of Rap or Wort in either the first ( F ) or the total ( T ) period . After this time , samples were treated with human fresh serum to induce the complement-dependent lysis of epimastigotes . Generated MT were then directly observed and counted in a Neubauer chamber ( S1 Fig , see details in Methods ) . Our results showed that the percentage of MT differentiated up to 48 h was higher in both TAU medium at the first ( TAU-F ) and the total ( TAU-T ) period of metacyclogenesis than in controls ( 12 . 1 +/- 0 . 9% and 16 . 4 +/- 1 . 7% vs . 3 . 9 +/- 0 , 6% , Fig 6A ) . Furthermore the addition of Wort to TAU during the first ( Wort-F ) or the total ( Wort-T ) period significantly reduced the number of MT in the parasite mixture , as compared to TAU-T condition . Conversely , Rap was able to increase the generation of MT when added to control medium at the first period of differentiation ( 8 . 3 +/- 0 . 2% vs . 3 . 9 +/- 0 . 6% ) . In contrast , Rap had not effect when added to the total period due to a toxic action of this compound on parasites at longer times . In another set of experiments , samples were allowed to infect Vero cells monolayers during 24 h and then fixed and processed for confocal microscopy . The degree of cell infection was considered an indirect measure of the number of MT present in the samples at different conditions , as previously described [22] . As depicted in Fig 6B , parasite nuclei were visualized in green due to the stable expression of TcH2b histone fused to GFP ( 24 ) , whereas host cells were in red due to the labeling with phalloidin-rhodamine that binds to actin cytoskeleton . The blue color depicts cell and parasite nuclei and parasite kinetoplast stained with Hoechst . Results showed that around 45% of cells were infected by TAU differentiated parasites , whereas 10% of the cells were infected by control parasites . The presence of Wort at either the first ( Wort-F ) or the whole ( Wort-T ) period of differentiation in TAU significantly reduced the infection rate . In concordance to the above result , the addition of Rap at the first period of time ( Rap-F ) induced a significant increase of infectivity ( Fig 6C ) . Taken together , these results demonstrated that autophagy has a key role during T . cruzi differentiation . Therefore , modulation of autophagy can be used to interrupt the metacyclogenesis rate in T . cruzi and to block the normal progression of the parasite biological cycle .
Metacyclogenesis is an essential process for the transmission of T . cruzi from the insect vector to mammalian hosts , which involves several metabolic and morphological changes in the parasite . Although the relationship between the rate of metacyclogenesis and a low nutritional state of the vector was described many years ago [32] , the specific cellular response to this stimulus was solely recognized upon identification of T . cruzi ATG genes [13] . In this work , we applied a previously published method of in vitro differentiation [20 , 25] to make a systematic analysis of autophagy and its possible modulators during T . cruzi metacyclogenesis . Our data show that after the first two hours of metacyclogenesis induction ( S1 Fig ) , there is an increased autophagic activity in epimastigotes , evidenced by a higher number of autophagosomes and lysosomes observed by both fluorescence and electron microscopy . A higher autophagic response of parasites is achieved during this first period of differentiation , when epimastigotes were exposed to a severe nutritional and thermic stress . Metacyclogenesis naturally occurs in the gut of the insect vector after a period of rapid multiplication of epimastigotes . Several factors are required to activate this process . Attachment of the parasite to the luminal surface of the insect’s rectum , hemolymph components , cAMP action and even the redox status of the parasite may promote metacyclogenesis [33–37] . Similarly to our in vitro method , a starvation environment caused by the higher number of replicating epimastigotes is another important inductor of metacyclogenesis . Considering that the main cellular response to starvation is autophagy , it is was not a surprise that this process was activated during parasite differentiation , as previously observed by other authors [18] . As mentioned above , metacyclogenesis is characterized by a renewal of proteins and subcellular structures required for parasite infection of a new host , while eliminating others that are no longer needed . It is expected that the pathways of degradation play a very important role and , among them , the autophagic pathway . During differentiation of the protozoan pathogen Trypanosoma brucei from the bloodstream form to the procyclic trypomastigote , the glycosomes , which are organelles that contain the enzymes of the glycolytic pathway , are significantly reduced while new organelles containing different enzymes are synthesized [38] . Further experiments carried out by the same research group have demonstrated that this efficient glycosome turnover involves autophagic degradation and confer to procyclic trypomastigotes the capacity to survive under the low glucose environment of the mosquito’s salivary glands [39] . Similarly to T . brucei , T . cruzi exhibits morphological and biochemical changes among the different stages . In this sense , previously works have evidenced a massive proteolysis during T . cruzi differentiation [40 , 41] . In agreement with these data , we observed that during the induction of metacyclogenesis , there occurs a significant expansion of the lysosomal compartment , as demonstrated by the increased number of acidic and hydrolytic vesicles . Moreover , the high colocalization levels of DQ-BSA and Atg8 . 1 indicates that the autophagic activity contribute to this process . Many virulence factors expressed in the infective forms may be modified by this increased hydrolytic activity . The trans-sialidase family member gp82 that is expressed earlier during differentiation is located in cruzipain-positive organelles at the posterior region before it is delivered towards the cell surface [42] . Autophagic degradation may also contribute to the cytostome-cytopharynx disappearance and the loss of the endocytic ability observed at the end of metacyclogenesis [43] . We next studied the effect of drugs widely used to modulate mammalian autophagy . Our data showed that rapamycin and wortmannin stimulated and inhibited parasite autophagy , respectively , indicating that the molecular targets of these drugs also exist in T . cruzi . Rapamycin is a reversible inhibitor of the kinase mTOR , a central regulator of cell growth [44] . Besides a study of mammalian TOR inhibitors as repurposed drugs against kinetoplastid parasites [45] , this is the first report that describes the cellular effects of rapamycin in T . cruzi . In T . brucei , TOR kinases are an extended family of proteins comprising the TbTOR1 and TbTOR2 that form the complexes TORC1 and TORC2 similar to mammals , and two additional TOR kinases , TbTOR3 and TbTOR4 . The latter forms a third complex that negatively regulates parasite differentiation [46] , whereas the TbTORC2 complex participates in cell growth and is sensible to rapamycin [47] . Ortholog sequences with homology to TbTOR and other genes that encode proteins with putative domains of TOR kinases were detected in T . cruzi . Three of them also contain the rapamycin recognizing domain [48] . Although the exact number and function of putative TcTOR genes is still unknown , our experimental data demonstrate a clear effect of Rap as an inducer of parasite autophagy and , as a consequence , of metacyclogenesis . Classical autophagy inhibitors like wortmannin have an effect on T . cruzi autophagy and metacyclogenesis . The presence and activities of inositol kinases in T . cruzi epimastigotes have been previously characterized [49] . The TcVps34 kinase plays an important role in osmoregulation , acidification and vesicular trafficking [50] and , as demonstrated in this work , also as an autophagy inhibitor . This enzyme is regulated by the TcVps15 catalytic activity and both TcVps15-TcVps34 form a complex that partially colocalizes to autophagosomes [51] . On the other hand , and contrarily to the autophagosome accumulation observed in mammalian cells after Baf treatment , this compound abrogated autophagosome formation in T . cruzi , resulting in a complete inhibition of the autophagic response after TAU treatment . This is probably a similar phenomenon to that observed in T . brucei . In this parasite , the acidocalcisome , which is a lysosome-related organelle characterized by acidic pH and a high content of Ca2+ and polyphosphates , has been found to regulate autophagy . Li et al . have demonstrated that the induction of autophagy in T . brucei , is accompanied by an acidification of acidocalcisomes and that drugs that impair this process , such as bafilomycin , completely inhibit the formation of autophagosomes [52] . Even though the mechanism has not been fully elucidated yet , the acidification of acidocalcisomes upon starvation of parasites , activates the synthesis of PI3P in the membrane of these organelles , a process required for autophagosome biogenesis [53] . In this work , we also verified that spermidine and spermine , that may act under some circumstances as a Spd analogue [31 , 54 , 55] , exerting a potent induction of T . cruzi autophagy under control conditions . The effect of Spd on autophagy has previously been demonstrated in yeast , flies , worms , and human immune cells [9] . In those cases , the activation of autophagy prolonged the life span of those organisms by counteracting the aging processes [10] . In contrast , on T . cruzi , and as previously shown [22] , PA activated differentiation processes like metacyclogenesis . In this work , we demonstrate that this action is mediated by the induction of parasite autophagy . Since the spermidine action is related to an increased expression of ATG genes [9] , and given the T . cruzi auxotrophy for polyamines , the acquisition of PA by the parasite should be produced before autophagy induction . For this reason , mutant parasites that expressed the heterologous ODC gene displayed high basal autophagic activity that was suppressed by DFMO , the non-reversible inhibitor of ODC . Moreover , when these mutants were incubated in TAU , autophagy was even higher , indicating an additive effect of both starvation and PA on the induction of autophagy . As expected , in contrast to ODC , the PAT12 mutant did not exhibit high basal autophagy until Spd or Spm were present in the medium ( S3 Fig ) . Surprisingly DFMO was able to inhibit basal autophagy in this mutant . An explanation of this unexpected result is the possible existence of an interference with the transport of PA ( or amino acids ) in the presence of the drug . Further studies are needed to clarify this point . Similarly to metacyclogenesis , other processes of T . cruzi differentiation may require autophagic activity . In a previous work we have found high levels of Atg8 in amastigotes located in the host cell cytoplasm , indicating the existence of increased autophagic activity in parasites at this stage [56] . Other works have also demonstrated the key role of protein degradation during amastigogenesis [16 , 57] . A detailed study of such mechanisms together with their possible inhibitors are crucial to find new drugs that interrupt the life cycle of T . cruzi . On the other hand , many trypanocidal drugs trigger autophagy in the parasite [58 , 59] . In this context , a deep knowledge of the mechanisms that regulate autophagy in this pathogen will contribute to a better understanding of the mechanisms of action of these drugs and improve the strategies for the treatment of Chagas’ disease . | In spite of its old discovery , more than one hundred years ago , Trypanosoma cruzi , the causative agent of Chagas’ disease , is still prevalent in the world , infecting more than 6 million people mostly in Latin America , where this illness is endemic . Only two approved drugs , benznidazole and nifurtimox , are currently used for the treatment of Chagas’ disease . Although efficient for the acute phase , they are poorly effective in the chronic period of the disease and they cause many undesirable side effects . There is an urgent need for therapeutic alternatives . To this end , identifying and validating novel molecular targets is critically relevant . This study describes the effect of different inhibitors on the T . cruzi autophagic pathway , a process required for parasite differentiation . Herein , we demonstrate that the regulation of parasite autophagy exhibits similarities and differences with host cell autophagy . Our study provides new insights that could be used to avoid T . cruzi cycle progression in both insect and mammalian hosts . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"cell",
"death",
"autophagic",
"cell",
"death",
"medicine",
"and",
"health",
"sciences",
"lysosomes",
"vesicles",
"cell",
"processes",
"microbiology",
"parasitic",
"diseases",
"protozoan",
"life",
"cycles",
"parasitic",
"protozoans",
"cell",
"differentiation",
"developm... | 2017 | The regulation of autophagy differentially affects Trypanosoma cruzi metacyclogenesis |
Broadly neutralizing antibodies targeting a highly conserved region in the hemagglutinin ( HA ) stem protect against influenza infection . Here , we investigate the protective efficacy of a protein ( HB36 . 6 ) computationally designed to bind with high affinity to the same region in the HA stem . We show that intranasal delivery of HB36 . 6 affords protection in mice lethally challenged with diverse strains of influenza independent of Fc-mediated effector functions or a host antiviral immune response . This designed protein prevents infection when given as a single dose of 6 . 0 mg/kg up to 48 hours before viral challenge and significantly reduces disease when administered as a daily therapeutic after challenge . A single dose of 10 . 0 mg/kg HB36 . 6 administered 1-day post-challenge resulted in substantially better protection than 10 doses of oseltamivir administered twice daily for 5 days . Thus , binding of HB36 . 6 to the influenza HA stem region alone , independent of a host response , is sufficient to reduce viral infection and replication in vivo . These studies demonstrate the potential of computationally designed binding proteins as a new class of antivirals for influenza .
The influenza envelope glycoprotein hemagglutinin ( HA ) on the surface of the influenza virus consists of a highly variable globular head domain ( HA1 ) and a more conserved stem domain ( HA2/HA1 ) [1 , 2] . Influenza viruses comprise two phylogenetic groups ( Groups 1 and 2 ) consisting of 18 HA subtypes and numerous genetic variants or strains within each subtype . Although vaccination can prevent influenza infection , current vaccines are strain specific , and provide minimal protection against drifted or shifted strains or subtypes [3–5] . New antivirals that broadly protect against a wide range of influenza variants are urgently needed to supplement the protective effects of vaccines and improve treatment options against seasonal influenza and future pandemics . Broadly neutralizing monoclonal antibodies ( bnAbs ) that bind the conserved HA stem can neutralize diverse influenza strains in vitro , suggesting that antivirals targeting the HA stem could provide similar widespread protection . BnAbs bind to the fusogenic region of the HA stem and inhibit the conformational rearrangements in HA required for membrane fusion [6–8] . Recent studies show that protection by HA-stem binding bnAbs is greatly enhanced through FcγR engagement in vivo [1 , 9] . While antibody binding to the fusogenic region is sufficient for in vitro neutralization of the virus , Fc-FcγR interaction and activation of antibody-dependent cellular cytotoxicity ( ADCC ) are critical for in vivo efficacy of stem-binding bnAbs [1 , 10] . We previously described two computationally designed small proteins that bind the HA stem region of multiple Group 1 influenza virus HA subtypes with equal or higher affinity than most bnAbs [11 , 12] . These results demonstrated the feasibility of using computational modeling to design a protein that mimics the stem binding of bnAbs in vitro , but since the designed proteins lacked an Fc , it was unclear if they would be able to afford protection against a rigorous influenza challenge in vivo . Here , we optimized one of these HA stem binding protein for tighter binding using deep mutational scanning [13] and investigated its ability to afford protection against influenza infection in vivo . We show that intranasal administration of an HA stem binding protein reduces viral replication and provides strong protection against diverse influenza strains when administered as a prophylactic or therapeutic in vivo . We further show that protection is independent of the host immune response , demonstrating that an antiviral can disrupts influenza infection in vivo via direct binding to the HA stem .
We optimized a broadly cross-reactive HA binding protein , HB36 . 5 , which is a stable , 97-residue designed protein , by increasing its affinity against multiple HA subtypes [12] . We constructed a library in which each residue was individually mutated to all other possible amino acids and carried out two rounds of yeast surface display selection against seven different Group 1 HA subtypes . We then sequenced the libraries and computed the enrichment ( or depletion ) of each individual point mutant during affinity maturation . The core of the binding interface was highly conserved in the selections against different HA subtypes ( Fig 1A , green ) , but several mutations in HB36 in the second shell of residues close to the binding interface were enriched across all seven subtypes ( Fig 1A ( red ) and 1B ) . Multiple subtype-specific substitutions were also identified around the periphery of the binding interface ( Fig 1B ) , which reflect changes to accommodate HA sequence differences near the interface ( Fig 1C ) . We made a combinatorial library of substitutions that were enriched across all subtype selections at 12 mutated positions and carried out three rounds of yeast display sorting against A/South Carolina/1/1918 ( H1N1 ) HA , which converged on a variant with nine substitutions called HB36 . 6 . Negative-stain electron microscopy revealed that HB36 . 6 binds to the designed target location on the HA ( Fig 2A , 2B and 2C ) . Biolayer interferometry showed that HB36 . 6 had higher affinity than HB36 . 5 against a wide range of Group 1 influenza subtypes , with greater than 40-fold and 10-fold affinity increases against H2 and H5 HAs , respectively ( Fig 2D and S1 Table ) . In vitro , HB36 . 6 potently neutralized a panel of genetically distinct human ( H1N1 ) and avian ( H5N1 ) influenza viruses ( range of genetic diversity between HA amino-acid sequences is 50–89% ) with a 50% effective concentration ( EC50 ) range of 0 . 18–12 . 0 μg/ml ( Fig 2E ) . This level is comparable to the EC50 range of 0 . 27–0 . 34 μg/ml observed for the monoclonal antibody , FI6v3 , which has been shown to broadly neutralize Group 1 and 2 influenza strains [6] . In addition , HB36 . 6 was more potent than ribavirin , a broad-spectrum antiviral that neutralizes influenza in vitro [14–16] and protects against influenza in mice [15 , 17] but had a higher EC50 of 15–18 μg/ml against a representative subset of the same influenza strains ( Fig 2E ) . However , HB36 . 6 did not neutralize either of the Group 2 strains tested or a Group 1 A/Hong Kong/2009 H9N2 strain , results that are consistent with computationally designed stem binders not binding Group 2 viruses [11] and FI6v3 weakly neutralizing the same H9N2 virus with an EC50 of 210 μg/ml . We next investigated the ability of HB36 . 6 to protect against influenza in mice . We administered a single intranasal ( IN ) dose of HB36 . 6 ( 6 . 0 mg/kg ) to BALB/c mice at 2 , 24 , or 48 hours prior to challenge with a lethal dose ( 10 times the 50% mouse lethal dose or 10 MLD50 ) of H1N1 A/California/04/2009 ( CA09 ) virus . CA09 is a highly virulent Group 1 pandemic influenza strain that leads to rapid weight loss and death in mice within 5–8 days post-infection ( d . p . i . ) [18] . When administered up to 48 hours before challenge , a single pre-exposure dose of HB36 . 6 afforded complete protection with 100% survival and little ( <10% ) to no weight loss , whereas all untreated controls ( Ctr ) exhibited >30% weight loss and no survival ( Fig 3A ) . Protection was specific to HB36 . 6 since a protein control ( lysozyme , 6 . 0 mg/kg ) , administered either 48 or 2 hours before CA09 challenge provided no protection and resulted in weight loss and mortality comparable to the controls ( Fig 3A ) . Protection was dependent on the IN route of delivery because the same dose of HB36 . 6 delivered intravenously ( IV ) provided no protection ( S1 Fig ) . When administered intranasally , HB36 . 6 was readily detected throughout the lung within 6 hours after administration indicating penetration into the lower respiratory tract ( S2 Fig ) . The observed prophylactic protection between 48–72 hours before challenge suggests a bioavailability range within this timeframe , although additional studies to determine the bioavailability and pharmacokinetics of HB36 . 6 will be needed . Lower doses of 1 . 0 , 0 . 1 , and 0 . 01 mg/kg administered IN two hours prior to lethal challenge with CA09 also resulted in 100% survival with little ( 0 . 1 mg/kg ) or no ( 1 . 0 mg/kg ) weight loss whereas controls exhibited rapid weight loss and succumbed to the infection within 7 d . p . i . ( Fig 3B ) . Mice that received the lowest IN dose tested ( 0 . 01 mg/kg ) exhibited weight loss , yet survived 2–3 days longer than controls and 20% of mice completely recovered . To determine if HB36 . 6 can protect against genetically distinct strains in vivo , we investigated protection against H1 and H5 viruses that are the most virulent Group 1 subtypes that infect mice and cause the majority or most severe Group 1 influenza infections in humans . We inoculated mice IN with HB36 . 6 ( 3 . 0 mg/kg ) two hours before challenge with either CA09 , A/PR8/34 ( PR8 ) , another highly virulent H1N1 mouse-adapted virus ( PR8 ) , or the H5N1 avian strain A/Duck/MN/1525/81 ( MN81 ) . The HA protein sequence of CA09 is 18% and 36% divergent from PR8 and MN81 , respectively . HB36 . 6 provided robust protection against these two genetically distinct H1N1 viruses and the highly pathogenic H5N1 virus , with 100% of the mice surviving and no weight loss ( Fig 3C ) . This result is consistent with the in vitro results showing that HB36 . 6 broadly binds and neutralizes H1 and H5 HAs ( Fig 2D and 2E ) . We next investigated HB36 . 6 for the ability to protect post-exposure . We challenged mice with CA09 and then treated with either a single IN dose of 3 . 0 mg/kg HB36 . 6 on day 0 ( 2 hours p . i . ) , +1 , +2 or +3 p . i . or four daily IN doses on days +1–4 p . i . HB36 . 6 reduced weight loss and afforded complete recovery and protection from lethality in 100% of mice when administered daily for 4 days or as a single inoculation 2 hours p . i . and 60% protection from lethality when administered as a single inoculation + 1 day p . i . ( Fig 4A ) . There was no significant difference in weight loss between mice receiving daily doses on days +1–4 p . i . or a single dose at 2 hours or day +1 p . i . , suggesting that a single dose within 1 day post-exposure is sufficient to protect from disease . Although mice that received HB36 . 6 at day +2 or +3 p . i . succumbed to infection , disease onset was delayed . The majority of these mice died at 8–9 d . p . i . , whereas 100% of the controls succumbed within 4–7 d . p . i . ( 2 d . p . i . , p = 0 . 0006; 3 d . p . i , p = 0 . 0031 compared to controls ) ( Fig 4A ) . The protection was specific for HB36 . 6 binding to the HA since daily administration of the scaffold protein ( PDB ID 1u84 ) that HB36 . 6 was modeled on provided no protection ( Fig 4A ) . We next compared a single dose of HB36 . 6 to oseltamivir [19] , an antiviral that targets influenza neuraminidase and is currently used to treat influenza in humans . We challenged mice with CA09 virus and then treated with either a single IN dose of HB36 . 6 ( 1 . 0–10 mg/kg ) on day +1 p . i . , or the recommended schedule of ten doses of oseltamivir administered by oral gavage , twice daily for 5 days starting on day +1 p . i . ( 5 mg/kg/day ) [17 , 20 , 21] . Oseltamivir afforded a modest delay in time to death , but provided no protection ( 0% ) from lethality . In contrast , escalating doses of HB36 . 6 protected mice from lethality with the highest dose ( 10 mg/kg ) affording 70% efficacy ( Fig 4B ) . Thus , a single dose of HB36 . 6 provided superior protection against a highly virulent influenza challenge when compared to ten doses ( 2 times per day for 5 days ) of a leading influenza antiviral . Furthermore , post-infection treatment with a combination of a sub-optimal single dose of HB36 . 6 ( 1–10 mg/kg ) and twice-daily doses of Oseltamivir resulted in 100% survival , indicating a synergistic effect when the two antivirals are combined ( Fig 4C ) . To determine the effects of HB36 . 6 at the respiratory sites of virus exposure , we analyzed viral replication in nasal and lung compartments in mice that received a single IN dose of HB36 . 6 ( 6 . 0 mg/kg ) either 24 hours before ( Prophylactic- Pro ) or after ( Therapeutic- Ther ) challenge with CA09 . We collected nasal washes on days 2 , 4 , and 6 post-challenge and viral titers were measured by an end-point dilution assay ( TCID50 ) . At each time-point p . i . , mice that were treated with HB36 . 6 before ( Pro ) or after ( Ther ) challenge exhibited a substantial 1–3 log-fold reduction in mean viral titer when compared to untreated controls , with the lowest viral load observed in the prophylactic group ( Fig 5A ) . We next investigated the effects of HB36 . 6 on viral replication in the lung . We treated mice IN with HB36 . 6 ( 6 . 0 mg/kg ) 1 day before or post-infection with CA09 , collected lung tissue on days 2 and 4 , and then stained for intracellular expressed influenza nucleoprotein ( NP ) to identify infected cells . Lung tissues from mice that received prophylactic or therapeutic administration of HB36 . 6 showed less viral replication in the lungs when compared to the untreated controls at day 4 p . i . ( Fig 5B ) and in situ enumeration of NP positive cells in the lung tissue confirmed significantly lower numbers of infected cells in the lung at day 4 in mice that received HB36 . 6 as a therapeutic compared to controls ( P ≤ 0 . 0263 , Fig 5C ) . The lower nasal wash viral loads in the prophylactic group ( Fig 5A ) but comparable lung viral loads in the prophylactic and therapeutic groups ( Fig 5B and 5C ) suggest that prophylaxis with HB36 . 6 likely affords protection by binding and blocking the virus at the nasal site of exposure , whereas post-exposure therapy with HB36 . 6 affords protection by containing the burst of viral replication and progeny release in the lung resulting in reduction of viral load in the lung and blunting of the inflammatory response that typically initiates within 24 hours after challenge [18] . Influenza infection results in the expression of cytokines that induce inflammation and recruit activated immune cells to clear the infection . However , this inflammatory response damages the pulmonary epithelium and increases susceptibility to secondary infections by ~100 fold [22 , 23] . To determine if HB36 . 6 protects from influenza-induced inflammation , mice were administered a single IN dose of HB36 . 6 ( 6 . 0 mg/kg ) either 24 hours before ( Pro ) or 24 hours after ( Ther ) lethal challenge with CA09 . Lungs were collected on day 2 p . i . and supernatants from lung homogenates were analyzed for the expression of inflammatory cytokines ( IL-6 , IL-10 , IL-12 ( p70 ) , TNF-α , IFN-γ ) . HB36 . 6 delivered as a prophylactic significantly lowered several cytokines , including the inflammatory cytokines IL-6 and TNF-α , when compared to controls ( P≤0 . 0012 , Fig 5D ) . HB36 . 6 delivered as a therapeutic also significantly lowered the amount of IL-12 ( p70 ) and IFN-γ when compared to controls ( P≤0 . 0007 , Fig 5D ) . These results suggest that reduction in viral load by HB36 . 6 provided an additional benefit of decreasing the cytokine responses that typically lead to increased inflammation and tissue damage . Together , these results show that HB36 . 6 blocks and interferes with viral spread , resulting in a lower viral replication , suppression of the cytokine response , and decreased lung inflammation . Furthermore , since HB36 . 6 lacks an Fc domain , these results show that engagement of the host FcγR is not required for protection in vivo . Small proteins , such as HB36 . 6 , may stimulate an immune response that could interfere with the effectiveness of a second administration or alternatively , stimulate antiviral responses that can contribute to protection [24] . Four doses of HB36 . 6 administered 2 weeks apart induced very low levels of antibody; however , 100% of mice were still completely protected when challenged with a lethal dose of CA09 1 day after the 4th dose ( S3 Fig ) . These results indicate that HB36 . 6 is poorly immunogenic and repeat administration does not interfere with the antiviral activity of subsequent doses . However , induction of even a modest antibody response after multiple doses suggested HB36 . 6 likely stimulated a host innate response . To determine if HB36 . 6 administration induces antiviral cytokine responses that could contribute to protection , cytokines were measured at different time-points post-HB36 . 6 administration . Mice either received a single IN dose of HB36 . 6 ( 6 . 0 mg/kg ) or the scaffold protein ( PDB 1u84 , 6 . 0 mg/kg ) and lungs were collected at 2 , 24 or 48 hrs post-administration . Supernatants from lung homogenates were analyzed for the expression of inflammatory cytokines ( IL-6 , IL-10 , IL-12 ( p70 ) , TNF-α , IFN-γ ) . Both HB36 . 6 and scaffold protein induced low levels of cytokines that peaked between 2–24 hrs post-administration and , by 48 hrs , the levels had dropped to pre-administration levels ( Fig 6A ) . Importantly , cytokine levels after HB36 . 6 administration were significantly lower than levels induced by scaffold protein that afforded no protection from challenge . These data suggest that , although administration with HB36 . 6 induced a low cytokine response , the levels were too transient and/or low to contribute to protection . To investigate the possibility that HB36 . 6 may induce other host responses that could contribute to protection , we tested HB36 . 6 for protection against influenza in two severe immune-deficient mouse models: NOD SCID gamma ( SCID ) and MyD88-/- mice . SCID mice lack mature T , B , and NK cells and are unable to develop an adaptive immune response [25 , 26] . MyD88-/- mice lack TLR signaling and are deficient in cytokine signaling , resulting in a severely dampened innate and adaptive immune response [27–29] . HB36 . 6 ( 6 . 0 mg/kg ) , scaffold protein ( 6 . 0 mg/kg ) , and another protein control ( lysozyme , 6 . 0 mg/kg ) were IN administered 2 hours before challenge with CA09 . HB36 . 6 protected 100% of the SCID mice and 90% of the MyD88-/- mice with only minimal weight loss ( Fig 6B ) , whereas the control SCID and MyD88-/- mice ( 1u84 , Protein , and Naïve ) exhibited significant weight loss and 0% survival . These results provide further evidence that the antiviral effect of HB36 . 6 is likely due to direct binding to the HA stem and is independent of an antiviral host response .
We showed previously that computationally designed proteins optimized for high affinity binding to a viral protein can neutralize viruses in vitro [11 , 12] . However , prior to this study , it was not known if such proteins would have sufficient stability and potency to afford protection in vivo . Here , we provide the first proof-of-concept that a novel small protein that was computationally designed to mimic bnAbs and bind the highly conserved HA stem could be developed into a highly effective antiviral capable of neutralizing and affording robust protection against diverse strains of influenza in vivo . We show that HB36 . 6 neutralizes a panel of Group 1 H1 and H5 viruses in vitro and a single intranasal dose afforded significant protection against three highly divergent H1 and H5 influenza strains in vivo . This suggests that the range of neutralizing specificity of HB36 . 6 observed in vitro translated to protection against these strains in vivo . Our studies also show that HB36 . 6 mediates protection independent of a host response . This contrasts to previous studies employing intravenous injection of a bnAb ( FI6v3 ) that HB36 . 6 binding was designed to mimic . These studies showed that engagement with the host’s FcγR and recruitment of ADCC was crucial for optimum protection by bnAb in vivo [1 , 9] . Here , we show that HB36 . 6 , which lacks an Fc , still affords robust protection against different strains of influenza in vivo . This outcome may be due to intranasal delivery of HB36 . 6 , which localizes the antiviral at the respiratory site of viral exposure and/or the ability of HB36 . 6 to bind the stem with high affinity [30] . Consistent with this possibility , Leyva-Grado et . al [30] showed that intranasal delivery of the fragment antigen-binding ( Fab ) region from the broadly neutralizing antibody , FI6v3 , afforded a similar degree of protection as we report here for HB36 . 6 . A recent study showed that a host receptor binding peptide provided prophylactic protection against lethal influenza challenge that depended on the induction of an inflammatory antiviral response [24] . The peptide did not work as a therapeutic , since antiviral cytokines are less effective after a viral infection is already established . In contrast , we found that HB36 . 6 induced only weak cytokine responses that were lower than the non-protective scaffold protein control and provided protection in two severe immune-deficient mouse models indicating a mechanism that is independent of a host antiviral cytokine immune response . Together , these results indicate that binding to the HA stem alone was sufficient for in vivo protection against influenza . These findings have implications for development of HB36 . 6 as a safe and effective alternative for protection from influenza . Here , we found that pre-exposure treatment with HB36 . 6 prevented infection without inducing an inflammatory response , hence it could be used pre-exposure to increase resistance to infection without the risk of inducing adverse inflammatory responses . Furthermore , since post-exposure inflammation mediates enhanced influenza disease and increased susceptibility to secondary infections [22] , this also suggests HB36 . 6 could be used to treat influenza without the risk of exacerbating disease due to immune effector-mediated inflammation . Finally , the ability of HB36 . 6 to mediate protection independent of the host response has further implications for protection in the immune-compromised or elderly , who comprise the majority of deaths from seasonal influenza each year [31] . The CA09 strain used in our therapeutic challenge studies is highly virulent , rapidly disseminating into the lower lung of mice within hours after challenge and causing death in control mice within 8 days [18] . Although weight loss is not seen until later time-points , the robust inflammatory response responsible for these symptoms is initiated within hours after challenge [32] . The level of protection afforded by HB36 . 6 against this strain when used as a therapeutic suggests significant potential to provide post-exposure benefit and improve treatment of influenza infection when compared to current treatments . Consistent with this possibility , we show that a single dose of HB36 . 6 administered to mice challenged with a highly virulent influenza strain outperformed a five-day , ten-dose regimen of oseltamivir , the lead antiviral approved for treatment of influenza in humans . This result is consistent with previously reported results showing that oseltamivir delayed , but did not protect from mortality in mice [33–35] . Furthermore combining sub-optimal doses of HB36 . 6 and oseltamivir resulted in synergistic protection , a result that suggests potential for use of HB36 . 6 as an approach to augment the effectiveness of existing marketed antivirals . Indeed , several studies have shown that therapeutic use of influenza antiviral combinations could increase antiviral potency , clinical effectiveness , and reduce resistance emergence [36 , 37] . Although this HA stem epitope is highly conserved the potential for emergence of resistance to HB36 . 6 will require further investigation . Previous studies with bnAbs , small molecule inhibitors , and proteins designed to bind the HA stem demonstrate that targeting the HA stem affords protection by inhibiting the low pH-induced fusion of the viral membrane with the endosomal membrane [1 , 9 , 11 , 12 , 38] . The direct binding of HB36 . 6 to the highly conserved fusion region similarly inhibits key conformational rearrangements in the HA that drive the fusion of the viral and endosomal membranes , blocking entry of the viral RNA into the cell via the endosome [11 , 12] . Intravenous delivery of bnAbs has been shown to be highly effective in mice and ferrets and is being developed for the hospital setting to treat severe and complicated influenza [1 , 6 , 30 , 39] . However , due to the route of delivery and the high cost of monoclonal antibodies , this strategy is not viable for treatment of uncomplicated influenza in the general population . An antiviral , such as HB36 . 6 , that is effective intranasally could be more widely self-administered in the general population pre- or post-exposure to prevent infection or shorten recovery from the infection . Seasonal drifted strains reduce vaccine efficacy and drug-resistant strains hinder the use of current antivirals in the prevention and treatment of influenza . These problems highlight the need for effective new antiviral drugs [40 , 41] . Overall , our results show that computationally designed proteins have potent anti-viral efficacy in vivo and suggests promise for development of a new class of HA stem-targeted antivirals for both therapeutic and prophylactic protection against seasonal and emerging strains of influenza .
Wild type HB36 . 5 and the transformed HB36 . 5 site-saturation mutagenesis ( SSM ) yeast display library was inoculated into 1mL of synthetic dextrose casamino acids ( SDCAA ) medium supplemented with carbenicillin and chloramphenicol and grown overnight at 30°C , 250rpm . Cells were pelleted by centrifugation , resuspended in 200μL of synthetic galactose casamino acids ( SGCAA ) , 40μL of the resuspended cells were inoculated into 960 additional μL of complete SGCAA and induced ~24h at 18°C , 250rpm . Cells were collected by centrifugation , washed with Phosphate Buffered Saline ( PBS ) , 0 . 1% w/v bovine serum albumin ( BSA ) ( PBSF ) , and diluted to optical density 600nm ( OD600 ) of 2 . 0 . 1 . 5x105 cells were mixed with purified biotinylated hemagglutinin ( HA ) in PBSF individually at a range of concentrations spanning the construct’s predicted kD and incubated at 22°C for 30m . After labeling with HA , the cells were collected by centrifugation , washed once with PBSF , and incubated with 0 . 6μL of FITC-labeled anti-CMyc antibody and 0 . 25μL phycoerythrin ( PE ) -labeled streptavidin on ice for 10m . Cells were collected , washed with PBSF , and resuspended into 200μL of PBSF . Fluorescence of 50 , 000 cells from each titration point was measured on an Accuri C6 flow cytometer with a 488nm laser for excitation and a 575nm band pass filter for emission . Negative controls for binding were induced cells with no HA labeling . BD Cflow software was used to measure the total PE fluorescence of the displaying cell population and a custom MATLAB non-linear curve fitting script was used to derive equilibrium binding constants for each hemagglutinin subtype . HB36 . 5 in pETCON plasmid was mutagenized individually via Kunkel’s method [42] in 86 consecutive codon positions using NNK degenerate primers purchased from Integrated DNA Technologies ( Coralville , IA ) . Primers were designed using Firnberg’s method . The theoretical library size of all 86 reactions was 1720 amino-acid sequences . Kunkel’s reactions were purified with QiaQuick columns ( Qiagen , Hilden , Germany ) and pooled in groups of 12 ( codon positions 1–12 , 13–24 , 25–36 , 37–48 , 49–60 , 61–72 , 73–86 ) . 1μL of each plasmid pool was transformed by electroporation into XL10 Gold electrocompetent cells ( Stratagene , La Jolla , CA ) with a minimum efficiency of 3x105 Colony Forming Units ( CFUs ) per pool ( >100-fold coverage ) . Liquid cultures were grown overnight in terrific broth ( TB ) and harvested using a QiaPrep miniprep kit ( Qiagen , Hilden , Germany ) . Mutated genes were amplified from each plasmid pool by adding 1μL of plasmid to 10μL of 5x Phusion Buffer , 1μL of 10mM dNTPs , 2 . 5μL of 10μM upGS primer , 2 . 5μL of 10μM downCMyc primer , and 0 . 5μL of Phusion polymerase in 50μL total volume . The reaction used 30 cycles of PCR ( 98°C 10s , 65°C 15s , 72°C 15s ) . PCR product was purified with a QiaQuick kit and transformed into EBY100 S . cerevisiae using Chao’s method [43] along with gel-purified pETCON vector digested with NdeI/XhoI ( NEB , Waltham , MA ) . A transformed HB36 . 5 SSM yeast display library was sorted in two rounds . For each round , cells were grown in 10mL of SDCAA overnight at 30°C , collected by centrifugation , and induced in SGCAA at 18°C for ~24 hours . Cells were collected by centrifugation , washed with PBSF , and ~4x106 cells labeled with purified biotinylated HA at a concentration half of the kD determined by yeast display titrations or , if no kD could be determined , at 500nM . In the first round of sorting , primary labeling proceeded for 30m at 22°C as described for titrations . In the second round , to select for mutations that improved binding but did not destabilize the protein , primary labeling was performed for 30m at 37°C . Secondary labeling was done with 1 . 2μL anti-c-Myc tag antibody conjugated to fluorescein isothiocyanate ( FITC ) and 0 . 5μL streptavidin-phycoerythrin ( SAPE ) in a total volume of 100μL PBSF on ice for 10m . In each sorted population , the top ~5% of FITC-displaying cells were collected . Plasmid DNA was prepared as previously described [11] . Genes were amplified from the plasmid by adding 36 . 5μL of purified plasmid to 10μL of 5x Phusion master mix , 1μL each of pETCON_inner_fwd and pETCON_inner_rev , 1μL of 10mM dNTPs , and 0 . 5μL of Phusion polymerase ( Thermo ) . The reaction used 30 cycles of PCR ( 98°C 10s , 58°C 15s , 72°C 15s ) . Correctly sized products were gel extracted using a Qiaquick gel extraction kit ( Qiagen ) . 10μL of gel extracted reaction product were added to 10μL of 5x Phusion master mix , 1μL each of miseq_outer_fwd and miseq_outer_rev with the correct library barcode , 1μL of 10mM dNTPs , and 0 . 5μL of Phusion polymerase in 50μL , and amplified again with 30 cycles of PCR ( 98°C 10s , 58°C 15s , 72°C 15s ) . The two primer sets have overhangs that add Illumina sequencing primer binding sites , barcode sequences , and flow cell adaptors to the gene to be sequenced . They additionally add 12 entirely degenerate bases at the beginning of the forward and reverse read , ensuring adequate diversity for the Illumina basecalling algorithms . This enabled the DNA pools to be prepared and sequenced in two runs of paired-end 251bp mode on an Illumina MiSeq ( Illumina , San Diego , CA ) using a standard MiSeq kit and protocols . Pools were mixed to have 6 . 5% of the total loaded DNA from the unselected pool , 6 . 5% of the total loaded DNA from each first-round selected pool , and 3 . 25% from each second-round selected pool , with 35% Illumina PhiX control DNA to increase diversity and data quality . Raw sequence files were processed into fastq format , split by barcode , allowing up to 1 mismatch , and adapter sequence was removed using Illumina OLB 1 . 9 . 4 . Split library sequences were processed using scripts from Enrich 0 . 2 to yield mutation counts in each library . Counts for each sorted library were converted to log2 enrichment relative to the unselected library using custom scripts . An enrichment value was calculated by linear regression of enrichment for each individual substitution at each round . The slope of the regressed line is the enrichment value [44] . Twelve positions in HB36 . 5 that contained substitutions highly enriched against many or all tested subtypes were mutated in a combinatorial library with a total sequence diversity of 108 . This library was constructed using recursive PCR assembly , as described below , with the only difference being that the assembly oligos contained degenerate codons designed using GLUE [45] to maximize enriched amino acid codon representation . This library was transformed into yeast using Chao’s method [43] with an efficiency around 107 , and sorted by three rounds of yeast display for binding to A/South Carolina/1/1918 HA until sequence convergence was achieved . The final converged sequence , HB36 . 6 , had 9 total mutations relative to HB36 . 5 . The gene for HB36 . 6 with 40bp of additional pETFLAG overlap sequence , to allow homologous recombination , was assembled via recursive PCR . Sequences were designed using DNAWorks [46] and purchased from Integrated DNA Technologies , Inc . ( Coralville , IA ) . The outermost two primers were diluted to 5μM and mixed , while the inner primers were diluted to 0 . 5μM and mixed . A 10μL volume of outer primer mix was added to 12 . 7μL of inner primer mix , along with 1μL of 10mM DNTPs , 6μL of 5x Phusion buffer , and 0 . 3μL of Phusion polymerase ( NEB , Waltham , MA ) for a final volume of 30μL . Product was assembled with 30 rounds of PCR ( 98°C 30s , 58°C 30s , 72°C 30s ) . A second round of PCR was used to further amplify correctly assembled product . A 1 . 25μL aliquot of the first PCR reaction product was added to 5μL of 5x Phusion Buffer , 0 . 75μL 10mM DNTPs , 2μL of outer primer mix , and 0 . 25μL of Phusion polymerase in 25μL . The same PCR conditions were used for a further 30 rounds and the product was purified using a QiaQuick PCR cleanup kit ( Qiagen , Hilden , Germany ) and eluted in EB . Gibson assembly [47] was used to insert the assembled gene into gel-purified pETFLAG vector digested with NdeI/XhoI ( NEB , Waltham , MA ) . A 1 . 5μL aliquot of cut vector at 20ng/μL was added to 1μL of assembled gene and 7 . 5μL of Gibson enzyme mix ( all enzymes from NEB , Waltham , MA ) . The reaction was incubated at 50°C for 1h and 2μL was transformed into 20μL of XL10 Gold chemically competent E . coli and plated onto a kanamycin agar plate . Plasmid sequences were confirmed by colony PCR and Sanger sequencing , and colonies with correctly assembled plasmid were grown in TB and harvested using a QiaPrep miniprep kit ( Qiagen , Hilden , Germany ) . HB36 . 6 in a pETFLAG vector was expressed in Rosetta2 ( DE3 ) E . coli cells . HB36 . 6 used throughout these studies contained the FLAG tag ( DYKDDDDK ) . Cells were grown in 500μL aliquots of medium salt aspartate-glucose ( MDG ) non-inducing media supplemented with kanamycin at 37°C , 250rpm overnight . Each vial was used to inoculate 500mL of ZYM-5052 auto-induction media [48] , which was grown for ~48 hours at 22°C , 250rpm . Cells were harvested by centrifugation and resuspended in 25mL of lysis buffer ( 50mM Tris , 300mM NaCl , 30mM imidazole , pH 8 . 2 ) with half of a dissolved complete , ethylenediaminetetracetic acid ( EDTA ) -free protease inhibitor tablet ( Roche , Basel , Switzerland ) and supplemented with DNAse and lysozyme at ~1mg/mL . Resuspended cells were lysed via sonication with a Qsonica Q500 ( Fisher Scientific , Hampton , New Hampshire ) at 70% power for 10 minutes ( 20s on/20s off ) on ice . Insoluble cell debris was removed by centrifugation for 30m at 40 , 000g . Supernatant was applied to gravity-flow columns containing 2 . 5mL of Ni-NTA resin ( Qiagen , Hilden , Germany ) pre-equilibrated with lysis buffer . Protein was washed with 25mL wash buffer ( 50mM Tris , 300mM NaCl , 75mM imidazole , pH 8 . 2 ) and eluted with 10mL elution buffer ( 50mM Tris , 300mM NaCl , 300mM imidazole , pH 8 . 2 ) . Protein was concentrated to ~20mg/mL using a Vivaspin 10kD MWCO centrifugal concentrator ( Sartorius Stedim , Goettingen , Germany ) at 4000g at 4°C . Imidazole was removed by dialysis ( 2x 4L buffer ) into 50mM Tris , 300mM NaCl , pH 8 . 2 at 4°C . Concentration was determined by absorbance at 280nm on a NanoDrop spectrophotometer ( Thermo Scientific , Waltham , Massachusetts ) using extinction coefficients calculated from amino acid sequences . For in vivo experiments , proteins were further processed by incubating with magnetic bacterial endotoxin removal beads . Beads were removed with a magnet and then samples were centrifuged to ensure complete removal . Endotoxin lipopolysaccharide ( LPS ) levels were reduced to <110 EU/mg ( Miltenyi Biotec , San Diego , California ) . A larger scale purification procedure for HB36 . 6 was used to produce the protein for the mouse study comparing HB36 . 6 to oseltamivir . Rosetta2 ( DE3 ) E . coli cells carrying the HB36 . 6 pETFLAG vector were cultured in 50 ml of ZY Broth ( 10 g/l tryptone , 5 g/l yeast extract ) with 15 ug/ml kanamycin at 37°C , at 250rpm overnight . The overnight culture was used to seed ( 3 ml starter culture per bottle ) multiple airlift fermentation bottles , each containing 2 liters of autoinduction media ZYP-5052 [48] supplemented with 15 ug/ml kanamycin , and the cultures were sparged with air for 72 hours at 25°C using an airlift LEX Bioreactor System ( Harbinger Biotech , Ontario Canada ) [49] . Cells were harvested at 4°C from culture media in 2 liter centrifuge buckets , pelleted at 4000g using a Sorvall RC 12 BP centrifuge fitted with an H-12000 swinging-bucket rotor . The masses of the cell pastes were measured to verify proper growth ( usually in the range of 20–30 g per 2 liter culture ) . The pelleted cells were flash frozen in liquid nitrogen and stored at -80°C . Frozen cell pellets from the equivalent of 2 liters of autoinduction culture were thawed by addition of 20 ml of Lysis Buffer ( 25 mM HEPES pH 7 . 0 , 500 mM NaCl , 5% Glycerol , 0 . 5% CHAPS , 30 mM Imidazole , 10 mM MgCl2 , 1 mM TCEP , 250 ug/ml AEBSF , and 0 . 025% Azide ) and 15 minute incubation at 37°C in a water bath , followed by addition of 0 . 01g of lysozyme and incubation for an additional 15 minutes in the 37°C water bath . The thawed cell pellet is gently resuspended with a spatula and transferred into a 500 mL beaker inside an ice bath filled with 180 ml of Lysis Buffer . The resuspended pellet is lysed in the ice bath by sonication for 30 minutes at 100W with “10 s ON and 20 s OFF” cycle . After sonication , the crude lysate is clarified with 20ul ( 25 units/ul ) of Benzonase and incubated at room temperature for 40 minutes in 250 ml centrifuge bottles using a Stuart SRT1 rotating mixer . The clarified crude extract is then centrifuged at 14 , 000 rpm for 1 hour at 4°C using a Sorvall SLA-1500 Rotor and the supernatant is transferred into a clean reservoir . Using an ÄKTAexplorer 100 ( GE Healthcare ) , the supernatant with soluble protein is pumped at 5 ml/min aver a 5 ml Ni-NTA His-Trap FF column ( GE Healthcare ) pre-equilibrated with Wash Buffer ( 25 mM HEPES pH 7 . 0 , 500 mM NaCl , 5% Glycerol , 30 mM Imidazole , 1 mM TCEP , and 0 . 025% Azide ) . The column is washed with 20 column volumes ( CVs ) of Wash Buffer and the bound protein is then eluted with 20 ml of Elution Buffer 1 ( Wash Buffer + 250 mM imidazole ) , followed by 20 ml of Elution Buffer 2 ( Wash Buffer + 320 mM imidazole ) . Each imidazole elution pool was then further purified by preparative size exclusion chromatography using an ÄKTAexplorer 100 , and a HiLoad 26/60 Superdex 75 preparative-grade column ( GE Healthcare ) equilibrated with SEC Buffer ( 25 mM HEPES , 0 . 5 M NaCl , 5% Glycerol , 1 mM TCEP , pH 7 ) [50] . Following SDS-PAGE analysis , the peak SEC fractions of HB36 . 6 were pooled together with additional SEC Buffer such that a HB36 . 6 was at final concentration of 1–2 mg/ml . Using a Millipore Tangential Flow Filtration System ( Millipore Labscale system ) , the pooled HB36 . 6 was concentrated to approximately 5 mg/mL using a 5kDa MWCO PES 10cm2 membrane , then diafiltered at a constant volume against Storage Buffer ( 20mM Tris-HCl , 300mM NaCl pH 8 ) for five diavolumes at 2-8C . Transmembrane pressure was held to 20psi . Concentrated HB36 . 6 samples were analyzed by SDS-PAGE to verify purity of the samples at >98% and mass spectrometry to verify identity . Endotoxin levels were also measured using the Endosafe PTS reader ( Charles River Laboratories ) , with a result of 17 . 2 EU per mg of HB36 . 6 . Finally the purified HB36 . 6 was dispensed into 2 ml vials and flash frozen in liquid nitrogen followed by storage at -80°C . Purified proteins were tested for folding and denaturation temperature using an Aviv 420 circular dichroism ( CD ) spectrometer ( Aviv Biomedical , Lakewood , NJ ) . Protein was diluted to 0 . 6mg/mL in PBS and CD absorbance was measured at 205nm at 25°C . Absorbance was characteristic of a structured α-helical protein . To test the thermal denaturation temperature of HB36 . 6 , absorbance at the 222nm α-helix peak was measured in 2° increments between 15° and 95°C . Proteins to be tested were diluted to 1mg/mL in PBS and Gibco 0 . 25% trypsin/phenol red ( Life Technologies , Carlsbad , CA ) was diluted to 0 . 005% in PBS . Equal parts protein solution and trypsin were mixed and incubated at 37°C . Time points were taken by removing 6μL aliquots of solution , mixing with 6μL of 2x SDS loading buffer ( 100mM Tris , 4% SDS , 0 . 2% bromphenol blue , 20% glycerol , pH 6 . 8 ) and incubating for 2m at 85°C . Denatured aliquots were stored at -20°C until being loaded on a 12% NuPage Bis-Tris gel ( Life Technologies , Carlsbad , CA ) and run at 150V in 1x MES buffer ( 50 mM MES , 50 mM Tris Base , 0 . 1% SDS , 1 mM EDTA , pH 7 . 3 ) . Negative controls were one lane with undigested protein and one lane with trypsin but no test protein . Gels were scanned and bands were quantified using ImageJ . Band size as a percentage of the undigested negative control was fit by non-linear regression using a custom MATLAB script to derive protein digestion half-lives . Titrations were performed on an OctetRed96 BLI system ( ForteBio , Menlo Park , CA ) using streptavidin-coated probe tips . Tips were equilibrated for 10m in 50x kinetics buffer , designed to reduce nonspecific binding ( PBS , pH 7 . 4 , 0 . 5% w/v BSA , 0 . 05% v/v Tween 20 ) and loaded with 25-40nM biotinylated HA of one of six subtypes in 50x kinetics buffer for 15m . Following a 10m wash and 10m baseline reading in the 50x kinetics buffer , association rates were measured by incubating each tip for 30m in different concentrations of purified HB36 . 5 or HB36 . 6 protein spanning the predicted kD for the given HA subtype , diluted in 50x kinetics buffer . Dissociation was measured by then incubating the tips in 50x kinetics buffer for a further 30m . HB36 . 6 did not show well-defined off-rates , so equilibrium binding constants were computed from the maximum steady-state response reached during the association phase . The limit of detection for the Octet instrument used in these experiments is around 1nM and the equilibrium binding experiments on the Octet yielded values of HB36 . 5 and HB36 . 6 against SC1918 at that threshold and within a standard error of each other . However , HB36 . 6 is a stronger binder against this strain with lower koff , though not a different kon . Unfortunately , due to equipment limitations , a specific measurement of how much stronger was not possible . The kD values for HB36 . 6-SC1918 were as low as 1pM but , since the dissociation curves are shallow , any minor deviation would cause enormous changes in the calculated kD . Thus equilibrium values are reported . Nonlinear regression curve fitting was done with a custom MATLAB script . MDCK ( Madin Darby canine kidney ) from American Type Culture Collection ( ATCC , Manassas , VA ) were grown in Growth medium comprising minimum essential medium ( MEM ) with non-essential amino acids , 5% FBS and 0 . 22% NaHCO3 . Influenza A/California/07/2009 ( H1N1 ) , A/Puerto Rico/08/1934 ( H1N1 ) , A/New Caledonia/20/1999 ( H1N1 ) , A/Hong Kong/213/2003 ( H5N1 ) , A/Nanchang/933/1995 ( H3N2 ) , A/Brisbane/10/2007 ( H3N2 ) , and A/Hong Kong/33982/2009 ( H9N2 ) , were obtained from the Center for Disease Control ( Atlanta , GA ) . Influenza A/Duck/MN/1525/81 was kindly provided by Robert Webster ( St . Jude Children’s Research Hospital , Memphis . TN ) . The viruses were prepared in Madin Darby canine kidney ( MDCK ) cells , placed in ampules and frozen at -80°C . Cells are seeded to 96-well flat-bottomed tissue culture plates at the proper cell concentration to establish confluent cell monolayers and incubated overnight at 37°C . Various dilutions of test compound were added to each well . Ribavirin ( 1-β−D-ribofuranosyl-1 , 2 , 4-triazole-3-carboxamide ) , FI6v3 , and HB36 . 6 were tested in half-log increments from 320 μg/ml and below . Virus was added to test compound wells and to virus control wells at about 50–100 cell culture infectious dose per ml . The virus titer was determined by a prior titration , where the most diluted virus stock causes 100% cytopathic effect ( CPE ) in all wells at the particular virus dilution . Test medium without virus was added to all toxicity control wells and to cell control wells . The plates were incubated at 37°C for 72 hours . Sterile neutral red ( 0 . 034% in saline solution ) was then added to each well . After two hours at 37°C , all medium was removed and the cells washed with PBS and inverted to drain . Neutral red was extracted from the cells by adding an equal volume mixture of absolute ethanol and Sörensen’s citrate buffer , pH 4 . 2 . The contents of each well are mixed gently and the optical density ( O . D . ) values of each well are obtained by reading the plates at 540 nm with a microplate reader . Complexes of HA and HB36 . 6 were prepared for electron microscopy studies by diluting to 2 . 1 μg/ml in Tris buffered saline and applied to freshly glow discharged carbon coated 400 mesh copper grids for 20 seconds . Two rounds of a 3 μl droplet of 2% uranyl formate were applied and immediately blotted followed by a third 3 μl droplet blotted after 1 minute . Grids were viewed using the FEI Tecnai T12 electron microscope operating at 120 kV accelerating voltage at 52 , 000 x magnification resulting in a pixel size 2 . 05 Å at the specimen level . Images were acquired on a Tietz 4k x 4k complementary metal-oxide-semiconductor ( CMOS ) camera using Leginon [51 , 52] MSI-raster 3 . 0 software package at a defocus of ~1 . 0 μm . Microscope magnifications were calibrated using a catalase crystal prior to data collection . Particles were picked automatically using DoG Picker [53] and boxed into 96x96 pixel boxes and aligned using Xmipp CL2D clustering alignment [54] . Ten ab initio models of each complex were created using EMAN2CL [55] with C3 symmetry and based on 17 2D class averages of PR8 in complex with HB36 . 6 . Initial models of complexes were then refined against 10 , 005 raw particles using EMAN [56] . The resolution of the final model was determined to be ~22 Å using an FSC cut-off of 0 . 5 . The UCSF Chimera “Fit in Map” function was used to dock structural models into the EM maps . Animal studies approved by the University of Washington and Utah State University Institutional Animal Care and Use Committee . Female , 6–8 week-old BALB/c mice were randomly separated in to groups , anesthetized and intranasally administered protein binder ( HB36 . 6 ) at concentrations varying from 0 . 01 to 6 . 0 mg/kg . Two to forty-eight hours later , the mice were anesthetized with 2 . 5% isoflurane and challenged IN with 3–10 MLD50 ( fifty percent mouse lethal dose ) of A/California/04/09 ( H1N1 ) ( CA09 ) , A/PR/8/34 ( H1N1 ) ( PR8 ) or A/Duck/MN/1525/81 ( H5N1 ) ( MN81 ) . In a therapeutic setting , mice received the protein binder 0 ( 2 hours post-infection ) , +1 , +2 , +3 , or +4 days post infection . The mice were monitored daily for weight loss and survival until 14 days post-infection . Animals that lost more than 30% of their initial body weight were euthanized by carbon dioxide in accordance with our animal protocols . Oseltamivir-treated mice received 2 . 5 mg/kg of oseltamivir ( Roche , Palo Alto , CA ) twice daily for 5 days ( total of 10 doses ) by oral gavage . Oseltamivir was dissolved in water prior to administration . The SCID ( Non-Obese Diabetic ( NOD ) , Severe Combined Immunodeficiency ( SCID ) gamma , strain NOD . Cg-Prkdcscid Il2rgtm1Wjl/SzJ ) mice and the MyD88-/- ( strain B6 . 129P2 ( SJL ) -Myd88tm1 . 1Defr/J ) mice were purchased from Jackson Laboratory . At least five mice per group were used for each experiment . All mice used for the experiments are included for analyses . For mouse experiments , researchers were not blinded to animal identity . Nasal wash samples were collected by making an incision in the trachea and washing the nasal passages with 0 . 2 ml sterile PBS ( pH 7 . 2 ) . Supernatants from lung homogenates were collected by mincing whole lungs in 500μl MEM media , freeze thawing twice on dry ice , and then centrifuging at 13 , 000rpm for 10m . The viral titers in the nasal washes and supernatants from lung homogenates were determined using the TCID50 , as described previously [57] . In brief , monolayers of MDCK cells were inoculated with tenfold serial dilutions of mouse nasal washes in quadruplicate ( three total replicates per sample ) . One hour after inoculation , the supernatants were removed and replaced with MEM media plus antibiotics and 1 μg/ml TPCK-trypsin ( Sigma , St . Louis , MO ) . The viral cytopathic effect was observed for 3 days before viral infectivity in MDCK cells was measured using a hemagglutination assay with 0 . 33% turkey erythrocytes . The tissue viral titers were calculated using the Reed and Muench method [58] and expressed as log10 TCID50/g of tissue . HB36 . 6-specific IgG antibody levels in mouse serum were assessed by ELISA . Maxisorp ( Thermo Scientific-Nunc ) were coated with 100 ng/well of HB36 . 6 in PBS overnight at 4°C . Plates were blocked with 5% nonfat milk powder in PBS for 1h at room temperature , and then washed three times with wash buffer ( PBS-T; phosphate-buffered saline containing 0 . 05% Tween 20 ) . Two-fold serial dilutions of samples were added to the wells and plates were incubated for 1hr at room temperature . Following three washes with PBS-T , plates were incubated with horseradish-peroxidase conjugated goat anti-mouse IgG ( 1/3 , 000 dilution ) secondary antibodies ( Thermo Scientific Pierce ) for 1h at room temperature . After five washes with PBS-T , TMB substrate ( KPL ) was added to the wells for 30 min at room temperature . Color development was stopped by the addition of TMB Stop solution ( KPL ) , and the plates were read at 450nm to measure relative optical densities ( O . D . ) . The concentrations of cytokines in lung tissue were measured . On days 2 and 4 post-infection , 8 mice per group were sacrificed and whole lung tissue was collected and immediately frozen . Lungs were thawed , weighed and lysed using the Bio-Plex Cell Lysis Kit ( Bio-Rad , Hercules , CA ) . The levels of interleukin ( IL ) -6 , IL-10 , IL-12 ( p70 ) , interferon ( IFN ) -γ , and tumor necrosis factor ( TNF ) -α in the lysate were measured using a Bio-Plex multiplex bead array kit ( Bio-Rad , Hercules , CA ) . The Bio-Plex assay was performed in accordance with the manufacturer’s instructions . During in vivo challenge experiments , lungs were removed from mice and immersed in 10% neutral buffered formalin . Following fixation , tissues were removed from formalin and placed in paraffin . Immunohistochemical staining was performed on the Leica Bond Automated Immunostainer . Sections were deparaffinized in Leica Bond Dewax Solution ( Leica Cat No . AR922 ) and rehydrated through 100% EtOH . After antigen retrieval with EDTA buffer pH 9 . 0 ( Lieca Bond Epitope Retrieval Solution 2 , Cat No AR9640 ) at 100°C for 20m , blocking endogenous peroxidase activity with 3 . 0% H2O2 for 5m , and blocking with 10% normal donkey serum in TBS for 20m , the sections were incubated with goat anti-influenza A virus , ( Meridian Life Science Inc . Cat No . B65141G ) at 1:2000 or normal goat IgG , isotype control , ( Invitrogen Cat No . 02–6202 ) at ( 1:5000 dilution ) both in Bond Primary Antibody Diluent ( Leica Cat No . AR9352 ) for 30m at room temperature . Sections were then incubated with rabbit anti-goat IgG ( Jackson ImmunoResearch Cat . No . 305-005-045 ) 1:1500 + 5% normal donkey serum for 8 minutes at RT followed by incubation with goat anti-rabbit poly-HRP polymer secondary detection ( Leica Cat No DS9800 ) for 8m at room temperature . Sections were then incubated with Leica Bond Mixed Refine DAB substrate detection for 10 minutes at room temperature . ( Leica Cat No DS9800 ) . After washing with DIH2O , the sections were counter stained with Hematoxylin solution ( Leica Bond Refine Kit ) dehydrated through 100% EtOH , cleared in Xylene and mounted with synthetic resin mounting medium and 1 . 5 coverslip . All of the analyses were performed using Graphpad Prism version 5 . 01 . A Student's t test ( to compare two samples ) and analysis of variance ( ANOVA ) ( to compare multiple samples ) were used for statistical analysis . Survival analyses were performed by using the Kaplan-Meier log-rank test . A P value of <0 . 05 was considered to be significant . For mice , the minimum group size was determined using weight loss data with 100% of control mice becoming infected with CA09 . Based on a standard deviation of 2% in weight loss , a group size of n = 5 yields >80% power to detect a minimum of a 10% difference between groups in weight loss using a two-sized t-test with an alpha value of 0 . 05 . | Influenza is a major public health threat , and pandemics , such as the 2009 H1N1 outbreak , are inevitable . Due to low efficacy of seasonal flu vaccines and the increase in drug-resistant strains of influenza viruses , there is a crucial need to develop new antivirals to protect from seasonal and pandemic influenza . Recently , several broadly neutralizing antibodies have been characterized that bind to a highly conserved site on the viral hemagglutinin ( HA ) stem region . These antibodies are protective against a wide range of diverse influenza viruses , but their efficacy depends on a host immune effector response through the antibody Fc region ( ADCC ) . Here we show that a small engineered protein computationally designed to bind to the same region of the HA stem as broadly neutralizing antibodies mediated protection against diverse strains of influenza in mice by a distinct mechanism that is independent of a host immune response . Protection was superior to that afforded by oseltamivir , a lead marketed antiviral . Furthermore , combination therapy with low doses of the engineered protein and oseltamivir resulted in enhanced and synergistic protection from lethal challenge . Thus , through computational protein engineering , we have designed a new antiviral with strong biopotency in vivo that targets a neutralizing epitope on the hemagglutinin of influenza virus and inhibits its fusion activity . These results have significant implications for the use of computational modeling to design new antivirals against influenza and other viral diseases . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"innate",
"immune",
"system",
"medicine",
"and",
"health",
"sciences",
"immune",
"physiology",
"cytokines",
"pathology",
"and",
"laboratory",
"medicine",
"influenza",
"pathogens",
"immunology",
"microbiology",
"orthomyxoviruses",
"viruses",
"animal",
"models",
"developmen... | 2016 | A Computationally Designed Hemagglutinin Stem-Binding Protein Provides In Vivo Protection from Influenza Independent of a Host Immune Response |
Long-read sequencing and novel long-range assays have revolutionized de novo genome assembly by automating the reconstruction of reference-quality genomes . In particular , Hi-C sequencing is becoming an economical method for generating chromosome-scale scaffolds . Despite its increasing popularity , there are limited open-source tools available . Errors , particularly inversions and fusions across chromosomes , remain higher than alternate scaffolding technologies . We present a novel open-source Hi-C scaffolder that does not require an a priori estimate of chromosome number and minimizes errors by scaffolding with the assistance of an assembly graph . We demonstrate higher accuracy than the state-of-the-art methods across a variety of Hi-C library preparations and input assembly sizes . The Python and C++ code for our method is openly available at https://github . com/machinegun/SALSA .
Genome assembly is the process of reconstructing a complete genome sequence from significantly shorter sequencing reads . Most genome projects rely on whole genome shotgun sequencing which yields an oversampling of each genomic locus . Reads originating from the same locus are identified using assembly software , which can use these overlaps to reconstruct the genome sequence [1 , 2] . Most approaches are based on either a de Bruijn [3] or a string graph [4] formulation . Repetitive sequences exceeding the sequencing read length [5] introduce ambiguity and prevent complete reconstruction . Unambiguous reconstructions of the sequence are output as “unitigs” ( or often “contigs” ) . Ambiguous reconstructions are output as edges linking unitigs . Scaffolding utilizes long-range linking information such as BAC or fosmid clones [6 , 7] , optical maps [8–10] , linked reads [11–13] , or chromosomal conformation capture [14] to order and orient contigs . If the linking information spans large distances on the chromosome , the resulting scaffolds can span entire chromosomes or chromosome arms . Hi-C is a sequencing-based assay originally designed to interrogate the 3D structure of the genome inside a cell nucleus by measuring the contact frequency between all pairs of loci in the genome [15] . The contact frequency between a pair of loci strongly correlates with the one-dimensional distance between them . Hi-C data can provide linkage information across a variety of length scales , spanning tens of megabases . As a result , Hi-C data can be used for genome scaffolding . Shortly after its introduction , Hi-C was used to generate chromosome-scale scaffolds [16–20] . LACHESIS [16] is an early method for Hi-C scaffolding that first clusters contigs into a user-specified number of chromosome groups and then orients and orders the contigs in each group independently to generate scaffolds . Thus , the scaffolds inherit any assembly errors present in the contigs . The original SALSA1 [21] method first corrects the input assembly , using a lack of Hi-C coverage as evidence of error . It then orients and orders the corrected contigs to generate scaffolds . Recently , the 3D-DNA [20] method was introduced and demonstrated on a draft assembly of the Aedes aegypti genome . 3D-DNA also corrects the errors in the input assembly and then iteratively orients and orders contigs into a single megascaffold . This megascaffold is then broken , identifying chromosomal ends based on the Hi-C contact map . There are several shortcomings common across currently available tools . They are sensitive to input assembly contiguity and Hi-C library variations and require tuning of parameters for each dataset . Inversions are common when the input contigs are short , as orientation is determined by maximizing the interaction frequency between contig ends across all possible orientations [16] . When contigs are long , there are few interactions spanning the full length of the contigs , making the true orientation apparent from the higher weight of links . However , in the case of short contigs , there are interactions spanning the full length of the contig , making the true orientation have a similar weight to incorrect orientations . Biological factors , such as topologically associated domains ( TADs ) , also confound this analysis [22] . SALSA1 [21] addressed some of these challenges , such as not requiring the expected number of chromosomes beforehand and correcting assemblies before scaffolding them with Hi-C data . We showed that SALSA1 worked better than the most widely used method , LACHESIS [16] . However , SALSA1 often did not generate chromosome-sized scaffolds . The contiguity and correctness of the scaffolds depended on the coverage of Hi-C data and required manual data-dependent parameter tuning . Building on this work , SALSA2 does not require manual parameter tuning and is able to utilize all the contact information from the Hi-C data to generate near optimal sized scaffolds permitted by the data using a novel iterative scaffolding method . In addition to this , SALSA2 enables the use of an assembly graph to guide scaffolding , thereby minimizing errors , particularly orientation errors . SALSA2 is an open source software that combines Hi-C linkage information with the ambiguous-edge information from a genome assembly graph to better resolve contig orientations . We propose a novel stopping condition , which does not require an a priori estimate of chromosome count , as it naturally stops when the Hi-C information is exhausted . We show that SALSA2 produces fewer orientation , ordering , and chimeric errors across a wide range of assembly contiguities . We also demonstrate its robustness to different Hi-C libraries with varying levels of intra-chromosomal contact frequencies . When compared to 3D-DNA , SALSA2 generates more accurate scaffolds across most conditions . To our knowledge , this is the first method to leverage assembly graph information for scaffolding Hi-C data .
Hi-C methods first crosslink a sample ( cells or tissues ) to preserve the genome conformation . The crosslinked DNA is then digested using multiple restriction enzymes ( targeting in this case the restriction sites GATC and GANTC ) . The single-stranded 5’-overhangs are then filled in causing digested ends to be labeled with a biotinylated nucleotide . Next , spatially proximal digested ends of DNA are ligated , preserving both short- and long-range DNA contiguity . The DNA is then purified and sheared to a size appropriate for Illumina short-read sequencing . After shearing , the biotinylated fragments are enriched to assure that only fragments originating from ligation events are sequenced in paired-end mode via Illumina sequencers to inform DNA contiguity . Hi-C paired end reads are aligned to contigs using the BWA aligner [26] ( parameters: -t 12 -B 8 ) as single end reads . First , the reads mapping at multiple locations are ignored as they can cause ambiguities while scaffolding . Reads which align across ligation junctions are chimeric and are trimmed to retain only the start of the read which aligns prior to the ligation junction . After filtering the chimeric reads , the pairing information is restored . Any PCR duplicates in the paired-end alignments are removed using Picard tools [27] . Read pairs aligned to different contigs are used to construct the initial scaffold graph . The suggested mapping pipeline is available at http://github . com/ArimaGenomics/mapping_pipeline . As any assembly is likely to contain mis-assembled sequences , SALSA2 uses the physical coverage of Hi-C pairs to identify suspicious regions and break the sequence at the likely point of mis-assembly . We define the physical coverage of a Hi-C read pair as the region on the contig spanned by the start of the leftmost fragment and the end of the rightmost fragment . A drop in physical coverage indicates a likely assembly error . In SALSA1 , contigs are split when a fixed minimum coverage threshold is not met . A drawback of this approach is that coverage can vary , both due to sequencing depth and variation in Hi-C link density . Fig 2 sketches the new contig correction algorithm implemented in SALSA2 . Instead of the single coverage threshold used in SALSA1 , a set of suspicious intervals is found with a sweep of thresholds . For a sweep of thresholds , we find the continuous stretches of regions which have lower physical coverage . Note that there can be multiple intervals for a particular threshold that have multiple stretches of low coverage . In such case , we only consider the interval of the maximum size . These intervals denote the regions of potential misassembly on the contig . Using the collection of these intervals as an interval graph , we find the maximal clique , which is the maximal set of intervals intersecting at any location along the contig . This maximal clique represents the region of the contig which had low coverage for the majority of the tested cutoffs . This can be done in O ( NlogN ) time , where N is the number of intervals . For a maximal clique , the region between the start and end of the smallest interval in the clique is flagged as a mis-assembly and the contig is split into three pieces—the sequence to the left of the region , the junction region itself , and the sequence to the right of the region . The intuition behind choosing the smallest interval is to accurately pinpoint the location of assembly error . Note that this algorithm finds only one misassembly per contig . For more rigorous misassembly detection , same algorithm can be run multiple times on each contig until no more drops in physical coverage are found . For our experiments , we use the unitig assembly graph produced by Canu [28] ( Fig 1 ( C ) ) , as this is a more conservative assembly output than contig sequences that represent various traversals of this graph . SALSA2 requires only a GFA format [25] representation of the assembly . Since most long-read genome assemblers such as FALCON [29] , miniasm [25] , Canu [28] , and Flye [30] provide assembly graphs in GFA format , their output is compatible with SALSA2 for scaffolding . The scaffold graph is defined as G ( V , E ) , where nodes V are the ends of contigs and edges E are derived from the Hi-C read mapping ( Fig 1B ) . The idea of using contig ends as nodes is similar to that used by the string graph formulation [4] . Modeling each contigs as two nodes allows a pair of contigs to have multiple edges in any of the four possible orientations ( forward-forward , forward-reverse , reverse-forward , and reverse-reverse ) . The graph then contains two edge types—one explicitly connects two different contigs based on Hi-C data , while the other implicitly connects the two ends of the same contig . As in SALSA1 , we normalize the Hi-C read counts by the frequency of restriction enzyme cut sites in each contig . This normalization reduces the bias in the number of shared read pairs due to the contig length as the number of Hi-C reads sequenced from a particular region are proportional to the number of restriction enzyme cut sites in that region . For each contig , we denote the number of times a cut site appears as C ( V ) . We define edges weights of G as: W ( u , v ) = N ( u , v ) C ( u ) + C ( v ) where N ( u , v ) is the number of Hi-C read pairs mapped to the ends of the contigs u and v . By the ends , we mean the first and second half of the contig if divided at the midpoint along its length . We observed that the globally highest edge weight does not always capture the correct orientation and ordering information due to variations in Hi-C interaction frequencies within a genome . To address this , we defined a modified edge ratio , similar to the one described in [20] , which captures the relative weights of all the neighboring edges for a particular node . The best buddy weight BB ( u , v ) is the weight W ( u , v ) divided by the maximal weight of any edge incident upon nodes u or v , excluding the ( u , v ) edge itself . Computing best buddy weight naively would take O ( |E|2 ) time . This is computationally prohibitive since the graph , G , is usually dense . If the maximum weighted edge incident on each node is stored with the node , the running time for the computation becomes O ( |E| ) . We retain only edges where BB ( u , v ) > 1 . This keeps only the edges that are the best incident edge on both u and v . Once used , the edges are removed from subsequent iterations . Thus , the most confident edges are used first but initially low-scoring edges can become best in subsequent iterations . For the assembly graph , we define a similar ratio . Since the edge weights are optional in the GFA specification and do not directly relate to the proximity of two contigs on the chromosome , we use the graph topology to establish this relationship . Let u ¯ denote the reverse complement of the contig u . Let σ ( u , v ) denote the length of shortest path between u and v . For each edge ( u , v ) in the scaffold graph , we find the shortest path between contigs u and v in every possible orientation , that is , σ ( u , v ) , σ ( u , v ¯ ) , σ ( u ¯ , v ) and σ ( u ¯ , v ¯ ) . With this , the score for a pair of contigs is defined as follows: S c o r e ( u , v ) = min x ′ ∈ { u , u ¯ } - { x } , y ′ ∈ { v , v ¯ } - { y } σ ( x ′ , y ′ ) min x ∈ { u , u ¯ } , y ∈ { v , v ¯ } σ ( x , y ) where x and y are the orientations in which u and v are connected by a shortest path in the assembly graph . Essentially , Score ( u , v ) is the ratio of the length of the second shortest path to the length of the shortest path in all possible orientations . Once again , we retain edges where Score ( u , v ) > 1 . If the orientation implied by the assembly graph differs from the orientation implied by the Hi-C data , we remove the Hi-C edge and retain the assembly graph edge ( Fig 1D ) . Computing the score graph requires |E| shortest path queries , yielding total runtime of O ( |E|* ( |V| + |E| ) ) since we do not use the edge weights . Once we have the hybrid graph , we lay out the contigs to generate scaffolds . Since there are implicit edges in the graph G between the beginning and end of each contig , the problem of computing a scaffold layout can be modeled as finding a weighted maximum matching in a general graph , with edge weights being our ratio weights . In a weighted maximum matching , a set of edges from a graph is chosen in such a way that they have no endpoints common and the sum of edge weights is maximized . In the case of scaffolding , a maximum weighted matching implies a layout of contigs , where no end can be used twice , that is maximally consistent with the data being used for scaffolding ( Hi-C in our case ) . If we find the weighted maximum matching of the non-implicit edges ( that is , edges between different contigs ) in the graph , adding the implicit edges to this matching would yield a complete traversal . However , adding implicit edges to the matching can introduce a cycle . Such cycles are prevented by removing the lowest-weight non-implicit edge . Computing a maximal matching takes O ( |E||V|2 ) time [31] . We iteratively find a maximum matching in the graph by removing nodes found in the previous iteration . Using the optimal maximum matching algorithm this would take O ( |E||V|3 ) time , which would be extremely slow for large graphs . Instead , we use a greedy maximal matching algorithm which is guaranteed to find a matching within 1/2-approximation of the optimum [32] . The greedy matching algorithm takes O ( |E| ) time , thereby making the total runtime O ( |V||E| ) . The algorithm for contig layout is sketched in Algorithm 1 . Fig 1 ( D ) –1 ( F ) show the layout on an example graph . Contigs which were not scaffolded are inserted in the large scaffolds with the method used in SALSA1 . If unitigs are used as an input and the layout of unitigs along contigs is provided as an input to SALSA2 , it can replace unitig sequences by contigs in the final scaffolds . Algorithm 1 Contig Layout Algorithm E: Edges sorted by the best buddy weight M: Set to store maximal matchings G: The scaffold graph while all nodes in G are not matched do M* = {} for e ∈ E sorted by best buddy weights do if e can be added to M* then M* = M* ∪ e end if end for M = M ∪ M* Remove nodes and edges which are part of M* from G end while Since the contig layout is greedy , it can introduce errors by selecting a false Hi-C link which was not eliminated by our ratio scoring . These errors propagate downstream , causing large chimeric scaffolds and chromosomal fusions . We examine each join made within all the scaffolds in the last iteration for correctness . Any join with low spanning Hi-C support relative to the rest of the scaffold is broken and the links are blacklisted for further iterations . We compute the physical coverage spanned by all read pairs aligned in a window of size w around each join . For each window , w , we create an auxiliary array , which stores −1 at position i if the physical coverage is greater than some cutoff δ and 1 , otherwise . We then find the maximum sum subarray in this auxiliary array , since it captures the longest stretch of low physical coverage . If the position being tested for a mis-join lies within the region spanned by the maximal clique generated with the maximum sum subarray intervals for different cutoffs ( Fig 2 ) , the join is marked as incorrect . The physical coverage can be computed in O ( w + N ) time , where N is the number of read pairs aligned in window w . The maximum sum subarray computation takes O ( w ) time . If K is the number of cutoffs ( δ ) tested for the suspicious join finding , then the total runtime of mis-assembly detection becomes O ( K ( N + 2*w ) ) . The parameter K controls the specificity of the mis-assembly detection , thereby avoiding false positives . The algorithm for mis-join detection is sketched in Algorithm 2 . When the majority of joins made in a particular iteration are flagged as incorrect by the algorithm , SASLA2 stops scaffolding and reports the scaffolds generated in the penultimate iteration as the final result . Algorithm 2 Misjoin detection and correction algorithm Cov: Physical coverage array for a window size w around a scaffold join at position p on a scaffold A: Auxiliary array I: Maximum sum subarray intervals for δ ∈ {min_coverage , max_coverage} do if Cov[i] ≤ δ then A[i] = 1 else A[i] = −1 end if sδ , eδ = maximum_sum_subarray ( A ) I = I∪{sδ , eδ} end for s , e = maximal_clique_interval ( I ) if p ∈ {s , e} then Break the scaffold at position p end if
We created artificial assemblies , each containing contigs of same size , by splitting the GRCh38 [33] reference into fixed-sized contigs of 200 to 900 kbp . This gave us eight assemblies . The assembly graph for each input was built by adding edges for any adjacent contigs in the genome . So the simulated assembly graph was linear with edges between two adjacent contigs for each contig in the graph . For real data , we use the recently published NA12878 human dataset sequenced with Oxford Nanopore [34] and assembled with Canu [28] . We use a Hi-C library from Arima Genomics ( Arima Genomics , San Diego , CA ) sequenced to 40x Illumina coverage ( SRX3651893 ) . Table 1 shows the statistics for this library . We compare results with the original SALSA ( commit—833fb11 ) , SALSA2 with and without the assembly graph input ( commit—68a65b4 ) , and 3D-DNA ( commit—3f18163 ) . We did not compare our results with LACHESIS because it is no longer supported and is outperformed by 3D-DNA [20] . SALSA2 was run using default parameters , with the exception of graph incorporation , as listed . For 3D-DNA , alignments were generated using the Juicer alignment pipeline [35] with defaults ( -m haploid -t 15000 -s 2 ) , except for mis-assembly detection , as listed . A genome size of 3 . 2 Gbp was used for contiguity statistics for all assemblies . For evaluation , we also used the GRCh38 reference to define a set of true and false links from the Hi-C graph . We mapped the assembly to the reference with MUMmer3 . 23 ( nucmer -c 500 -l 20 ) [36] and generated a tiling using MUMmer’s show-tiling utility . For this “true link” dataset , any link joining contigs in the same chromosome in the correct orientation was marked as true . This also gave the true contig position , orientation , and chromosome assignment . We masked sequences in GRCh38 that matched known structural variants from a previous assembly of NA12878 [37] to avoid counting true variations as scaffolding errors . Table 2 lists the metrics for NA12878 scaffolds . We include an idealized scenario , using only reference-filtered Hi-C edges for comparison . As expected , the scaffolds generated using only true links had the highest NGA50 value and longest error-free scaffold block . SALSA2 scaffolds were generally more accurate and contiguous than the scaffolds generated by SALSA1 and 3D-DNA , even without use of the assembly graph . The addition of the graph further improved the NGA50 and longest error-free scaffold length . We also evaluated the assemblies using Feature Response Curves ( FRC ) based on scaffolding errors [40] . An assembly can have a high raw error count but still be of high quality if the errors are restricted to only short scaffolds . FRC captures this by showing how quickly error is accumulated , starting from the largest scaffolds . Fig 5 ( D ) shows the FRC for different assemblies , where the X-axis denotes the cumulative % of assembly errors and the Y-axis denotes the cumulative assembly size . The assemblies with more area under the curve accumulate fewer errors in larger scaffolds and hence are more accurate . SALSA2 scaffolds with and without the graph have similar areas under the curve and closely match the curve of the assembly using only true links . The 3D-DNA scaffolds have the lowest area under the curve , implying that most errors in the assembly occur in the long scaffolds . This is confirmed by the lower NGA50 value for the 3D-DNA assembly ( Table 2 ) . Apart from the correctness , SALSA2 scaffolds were highly contiguous and reached an NG50 of 112 . 8 Mbp ( cf . GRCh38 NG50 of 145 Mbp ) . Fig 6 shows the alignment ideogram for the input contigs as well as the SALSA2 assembly . Every color change indicates an alignment break , either due to error or due to the end of a sequence . The input contigs are fragmented with multiple contigs aligning to the same chromosome , while the SALSA2 scaffolds are highly contiguous and span entire chromosomes in many cases . Fig 7 ( A ) shows the contiguity plot with corrected NG stats . As expected , the assembly generated with only true links has the highest values for all NGA stats . The curve for SALSA2 assemblies with and without the assembly graph closely matches this curve , implying that the scaffolds generated with SALSA2 are approaching the optimal assembly of this Arima-HiC data . We also evaluated the ability of scaffolding short-read assemblies for both 3D-DNA and SALSA2 . We did not include SALSA1 in this comparison because it is not designed to scaffold short-read assemblies . We observed that use of the assembly graph when scaffolding significantly reduced the number of orientation errors for SALSA2 , increasing the scaffold NGA50 and largest chunk almost two-fold . When compared to 3D-DNA without input assembly correction , SALSA2 with the assembly graph generates scaffolds of much higher NGA50 ( 7 . 99 Mbp vs . 1 . 00 Mbp ) . The number of scaffolding errors across all the categories was much lower in SALSA2 compared to 3D-DNA . We computed the CPU runtime and memory usage for both the methods while scaffolding long and short read assemblies with Arima-HiC data . We excluded the time required to map reads in both cases . While scaffolding the long-read assembly SALSA2 was 30-fold faster and required 3-fold less memory than 3D-DNA ( 11 . 44 CPU hours and 21 . 43 Gb peak memory versus 3D-DNA with assembly correction at 318 CPU hours and 64 . 66 Gb peak memory ) . For the short-read assembly , the difference in runtime was even more pronounced . SALSA2 required 36 . 8 CPU hours and 61 . 8 Gb peak memory compared to 2980 CPU hours and 48 . 04 Gb peak memory needed by 3D-DNA without assembly correction . When run with assembly correction , 3D-DNA ran over 14 days on a 16-core machine without completing so we could not report any results . We next tested scaffolding using two libraries with different Hi-C contact patterns . The first , from [42] , is sequenced during mitosis . This removes the topological domains and generates fewer off-diagonal interactions . The other library was from [43] , are in vitro chromatin sequencing library ( Chicago ) generated by Dovetail Genomics ( L1 ) . It also removes off-diagonal matches but has shorter-range interactions , limited by the size of the input molecules . As seen from the contact map in Fig 8 , both the mitotic Hi-C and Chicago libraries follow different interaction distributions than the standard Hi-C ( Arima-HiC in this case ) . Table 1 shows the mapping statistics for these libraries . We ran SALSA2 with defaults and 3D-DNA with both the assembly correction turned on and off . For mitotic Hi-C data , we observed that the 3D-DNA mis-assembly correction algorithm sheared the input assembly into small pieces , which resulted in more than 25 , 000 errors and more than half of the contigs incorrectly oriented or ordered . Without mis-assembly correction , the 3D-DNA assembly has a higher number of orientation ( 250 vs . 81 ) and ordering ( 215 vs . 54 ) errors compared to SALSA2 . The feature response curve for the 3D-DNA assembly with breaking is almost a diagonal ( Fig 5 ( B ) ) because the sheared contigs appeared to be randomly joined . SALSA2 scaffolds contain longer stretches of correct scaffolds compared to 3D-DNA with and without mis-assembly correction ( Fig 7 ( B ) ) . SALSA1 scaffolds had a similar error count to SALSA2 but were less contiguous . For the Chicago libraries , 3D-DNA without correction had the best NGA50 and largest correct chunk . However , the scaffolds had more chimeric join errors than SALSA2 . SALSA2 outperformed 3D-DNA in terms of NG50 , NGA50 , and longest chunk when 3D-DNA was run with assembly correction . 3D-DNA uses signatures of chromosome ends [20] to identify break positions which are not prominently present in Chicago data . As a result , it generated more chimeric joins compared to SALSA2 . However , the number of order and orientation errors was similar across the methods . Since Chicago libraries do not provide chromosome-spanning contact information for scaffolding , the NG50 value for SALSA2 is 5 . 8 Mbp , comparable to the equivalent coverage assembly ( 50% L1+L2 ) in [43] but much smaller than Hi-C libraries . Interestingly , SALSA1 was able to generate scaffolds of similar contiguity to SALSA2 , which can be attributed to the lack of long range contact information in the library . SALSA2 is robust to changing contact distributions . In the case of Chicago data it produced a less contiguous assembly due to the shorter interaction distance . However , it avoids introducing false chromosome joins , unlike 3D-DNA , which appears tuned for a specific contact model . To evaluate the effectiveness of SALSA2 on a non-model organism , we used Hi-C data from recently published Anopheles funestus genome assembly which was scaffolded using an independent method ( Phase Genomics or LACHESIS [16] ) and manually curated using Illumina mate-pair support as well as FISH information [44] . This genome had high heterozygosity as the data was sequenced from a colony of mosquitoes rather than a single individual . Due to this , the assembly had a high duplication rate and was almost double the expected genome size . We scaffolded both the full assembly and the assembly after running purge haplotigs [45] using SALSA2 and 3D-DNA . For the post purge assembly , 3D-DNA generated an assembly with higher continuity but with more errors and a similar NA50 to SALSA2 . ( S3 Table ) . However , neither method performed well for the full assembly . SALSA2 was more contiguous than 3D-DNA ( S4 Table ) but was still very fragmented and much larger than the expected genome size . We conclude that heterozygous genome scaffolding remains a challenge and assemblies must either be de-duplicated beforehand or improved algorithms for scaffolding , such as [46] are needed .
In this work , we present the first Hi-C scaffolding method that integrates an assembly graph to produce high-accuracy , chromosome-scale assemblies . Our experiments on both simulated and real sequencing data for the human genome demonstrate the benefits of using an assembly graph to guide scaffolding . We also show that SALSA2 outperforms alternative Hi-C scaffolding tools on assemblies of varied contiguity , using multiple Hi-C library preparations . SALSA2’s misassembly correction and scaffold misjoin validation can be improved in several ways . The current implementation does not detect a misjoin between two small contigs with high accuracy , mainly because Hi-C data does not have enough resolution to correct such errors . Also , we do not account for any GC bias correction when using the Hi-C coverage for detecting misjoins . Addressing these challenges in misjoin detection and misassembly correction is the immediate next step to improve the SALSA2 software . The human genome is relatively homozygous compared to many other species . Assembly of many species is further complicated by DNA input requirements which necessitates pooling multiple individuals . SALSA2 does not remove duplication present in an input assembly and thus requires pre-processing by another tool , such as Purge Haplotigs [45] or haplomerger [47] . Once contigs are classified into the “primary” and “haplotig” sets , SALSA2 could be run on each of the sets independently . Hi-C scaffolding has been historically prone to inversion errors when the input assembly is highly fragmented . The integration of the assembly graph with the scaffolding process can overcome this limitation . Orientation errors introduced in the assembly and scaffolding process can lead to incorrect identification of structural variations . On simulated data , more than 50% of errors were due to inversions , and integrating the assembly graph reduced these by as much as 3 to 4 fold . We did not observe as much improvement with the NA12878 test dataset because the contig NG50 was much higher than in the simulation . However , it is not always possible to assemble multi-megabase contigs . In such cases , the assembly graph is useful for limiting Hi-C errors . Most existing Hi-C scaffolding methods also require an estimate for the number of chromosomes for a genome . This is implicitly taken to be the desired number of scaffolds to output . As demonstrated by the Chicago , mitotic , and replicate [48] Hi-C libraries , the library as well as the genome influences the maximum correct scaffold size . It can be impractical to sweep over hundreds of chromosome values to select a “best” assembly . Since SALSA2’s mis-join correction algorithm stops scaffolding after the useful linking information in a dataset is exhausted , no chromosome count is needed as input . Obtaining the chromosome-scale picture of the genome is important and there is a trade-off between accuracy and continuity of the assembly . However , we believe that manual curation to remove assembly errors is an expensive and involved process that can often outpace the cost of the rest of the project . Most of the assembly projects using Hi-C data have had a significant curation effort to date [19 , 49] . Thus , we believe that not introducing errors in the first place is a better strategy to avoid the later burden of manual curation of small errors in chromosomes . The Hi-C data can be used with other independent technologies , such as optical mapping or linked-reads to reach accurate chromosome-scale scaffolds . 3D-DNA was recently updated to not require the chromosome count as input but the algorithm used has not been described . Interestingly , it no longer generates single-chromosome scaffolds in our experiments , a major claim in [20] , supporting a conservative scaffolding approach . Even while scaffolding short-read assemblies , we observed that SALSA2 generated more accurate scaffolds than 3D-DNA , indicating the utility of SALSA2 in scaffolding existing short-read assemblies of different genomes with the newly generated Hi-C data . As the Genome10K consortium [50] and independent scientists begin to sequence novel lineages in the tree of life , it may be impractical to generate physical or genetics maps for every organism . Thus , Hi-C sequencing combined with SALSA2 presents an economical alternative for the reconstruction of chromosome-scale assemblies . | Hi-C technology was originally proposed to study the 3D organization of a genome . Recently , it has also been applied to assemble large eukaryotic genomes into chromosome-scale scaffolds . Despite this , there are few open source methods to generate these assemblies . Existing methods are also prone to small inversion errors due to noise in the Hi-C data . In this work , we address these challenges and develop a method , named SALSA2 . SALSA2 uses sequence overlap information from an assembly graph to correct inversion errors and provide accurate chromosome-scale assemblies . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"infographics",
"sequencing",
"techniques",
"sequence",
"assembly",
"tools",
"genome",
"sequencing",
"chromosome",
"mapping",
"genome",
"analysis",
"molecular",
"biology",
"techniques",
"research",
"and",
"analysis",
"methods",
"sequence",
"analysis",
"genomic",
"librarie... | 2019 | Integrating Hi-C links with assembly graphs for chromosome-scale assembly |
Active-sensing systems abound in nature , but little is known about systematic strategies that are used by these systems to scan the environment . Here , we addressed this question by studying echolocating bats , animals that have the ability to point their biosonar beam to a confined region of space . We trained Egyptian fruit bats to land on a target , under conditions of varying levels of environmental complexity , and measured their echolocation and flight behavior . The bats modulated the intensity of their biosonar emissions , and the spatial region they sampled , in a task-dependant manner . We report here that Egyptian fruit bats selectively change the emission intensity and the angle between the beam axes of sequentially emitted clicks , according to the distance to the target , and depending on the level of environmental complexity . In so doing , they effectively adjusted the spatial sector sampled by a pair of clicks—the “field-of-view . ” We suggest that the exact point within the beam that is directed towards an object ( e . g . , the beam's peak , maximal slope , etc . ) is influenced by three competing task demands: detection , localization , and angular scanning—where the third factor is modulated by field-of-view . Our results suggest that lingual echolocation ( based on tongue clicks ) is in fact much more sophisticated than previously believed . They also reveal a new parameter under active control in animal sonar—the angle between consecutive beams . Our findings suggest that acoustic scanning of space by mammals is highly flexible and modulated much more selectively than previously recognized .
The importance of “active sensing , ” by which an animal actively interacts with the environment to adaptively control the acquisition of sensory information , is fundamental to perception across sensory modalities [1]–[7] . Echolocating bats emit ultrasonic signals and analyze the returning echoes to perceive their surroundings . Bat echolocation , an active sensory system , enables an acoustic representation of the environment through precise control of outgoing sonar signals . Laryngeal bats control many aspects of their sensory acquisition: they determine the timing of acquisition and the information flow [8]–[11] , they control the intensity of the emission as well as its direction [12]–[17] , and they control the spectral and temporal resolution of the acquired data [18]–[23] . Another acoustic parameter potentially under active control by echolocating bats is the pattern of the sonar beam . It has been debated whether bats can actively adjust the width of the sonar beam in response to task conditions , but empirical studies have not yet adequately addressed this question . It seems likely that bats would benefit greatly from the ability to control the beam pattern . They could for instance narrow the beam in order to concentrate energy onto a certain object , or they could widen the beam to increase the size of the sector that is being scanned . Studying the bat's active control over the shape and directionality of sonar emissions is technically difficult because reconstruction of the beam pattern requires a large circumferential ultrasonic microphone array in a setting where a free-flying bat engages in sonar tasks . A recent study suggests that laryngeal echolocating bats can change the space covered by their beam through adjustments in their call spectrum [24] . Here , we aimed to examine a very different mechanism by which echolocating bats might control the effective space they scan , namely adjustments in the angle between sequentially emitted sonar clicks . We studied this question in lingual echolocating bats . Lingual echolocation is exhibited by one family of fruit bats , Rousettus , and has been historically considered to be more rudimentary than laryngeal echolocation [25] . The primary reason behind this notion was that these bats were believed to have very little control over their sonar emissions . In contrast , we recently demonstrated that the lingual echolocator Rousettus aegyptiacus ( Egyptian fruit bat ) uses a sophisticated strategy for beam-steering: This bat emits sonar clicks in pairs , and it directs the maximum slope of each sonar beam towards the target , rather than directing the center of the beam , thereby optimizing stimulus localization in the horizontal plane [15] . Here , we further tested Egyptian fruit bats' active control over their echolocation-based sensory acquisition . To this end , we tracked the flight trajectories of Egyptian fruit bats in a large room , and recorded their echolocation behavior when performing a landing task under different levels of environmental complexity . We found that lingual echolocation allows much more selective control over sonar signal parameters than previously believed . We discovered that Egyptian fruit bats alter the intensity of their emissions as they approach and lock the sonar beam onto a target , and that emission intensity changes with environmental complexity . Moreover , we found that Egyptian fruit bats apply a novel strategy to change the spatial region , or “field-of-view” that they scan: They increase the angle between the beam axes of sonar click-pairs , to effectively increase spatial scanning . Such a strategy has never been observed before in any bat species , and therefore comprises a new dimension of active control in lingual bat echolocation .
In the first set of experiments—the “one-object experiments”—bats were trained to detect , localize , and land on a 10-cm diameter sphere , similar in size to fruit eaten by this bat species , such as mango . The sphere was the only object in an empty flight room ( Figure 1A ) , and it was randomly moved between trials . Recordings were taken in complete darkness , forcing the bats to rely only on echolocation ( see Materials and Methods ) . The echolocation of Egyptian fruit bats is comprised of pairs of clicks with a short inter-click time interval ( ∼20 ms ) and a longer inter-pair interval ( ∼90 ms in complete darkness ) [26] , [27] . The bats direct their sonar beam axes Left-Right→Right-Left , maintaining a certain angle between the sequential clicks of a pair ( Figure 1A–B ) [15] . When approaching the target , bats significantly increased the inter-click angle by 6 . 8±0 . 4 degrees , on average ( mean ± s . e . m . ; t test of inter-click angle before locking versus after locking , when pooling all data together: p<10−5 ) . This increase in inter-click angle occurred abruptly , coinciding with the time when the bats locked on the landing target , i . e . the time when the average direction of the click-pair coincided with the direction to the target ( Figure 1C; see Materials and Methods ) [15] . The increase occurred in all individual bats ( Figure S1 ) , and on average across all bats the change represented a 15% widening in the inter-click angle ( post-locking compared to pre-locking ) . Population analysis of 236 trials ( Figure 1D ) confirmed that the increase of the inter-click angle was abrupt; in fact , it could occur within 2 click-pairs , i . e . as fast as 200 ms ( Figure 1C–D ) . This abrupt increase in inter-click angle may result from the bat's need to increase the field-of-view; or it may represent the animal's attempt to position the maximum slope of its sonar beam onto the target [15] . To further elucidate the possible roles of this abrupt change in inter-click angle , we conducted additional experiments that aimed to challenge the bat's scanning behavior . To this end , we manipulated the spatial complexity ( number of objects ) that the bat encountered within its field-of-view as it flew towards the landing sphere . In the next set of experiments , we manipulated the complexity of the environment , and examined how this influenced the Egyptian fruit bat's echolocation behavior . We hypothesized that when introducing a set of objects ( obstacles ) in the vicinity of the landing-point , which increases the environmental complexity , the bats would alter their scanning behavior to inspect several objects—thus increasing their field-of-view . To test this hypothesis , we studied the bats' behavior in two new setups ( Figure 2A ) : ( i ) Open room condition: In 56 trials ( 8–12 trials per bat ) we removed the sphere where the bats were trained to land . These trials were randomly introduced in between one-object trials; hence the bats reacted by vigorously searching for the target while flying around the room . We shall refer to this setup as the “no-object” experiment ( Figure 2A , left ) . ( ii ) Environmentally complex condition: In 54 trials ( 8–11 per bat ) we added two nets that were spread between four poles on both sides of the target , creating a relatively narrow ( 0 . 6–1 . 6 m ) corridor for accessing the target ( Figure 2A , right ) . The width of the corridor , its angle relative to the walls of the room , and the position of the landing sphere within the corridor were all randomly varied between trials . This setup mimics natural situations , in which a bat has to negotiate fruitless branches ( the nets ) , before landing on a branch with a fruit ( the target ) . We refer to this setup as the “multiple-object” experiment , because the bats consistently negotiated some or all of the five objects in the room—the single landing sphere ( Figure 2A , right , closed gray circle ) and the four poles ( open circles ) . In all illustrations , bat's trajectory is depicted by a gray line and the direction of the beam's peak by a black line . Egyptian fruit bats increased the angle between sequential clicks when environmental complexity increased ( Figure 2A , bottom ) . The angular separation between the beam axes of sonar click pairs in the “no-object” setup was the narrowest; it increased in the one-object setup by 9 . 2±0 . 4 degrees ( after locking , “L” ) , and increased even further in the multiple-object setup , widening on average by 12 . 3±0 . 6 degrees compared to the “no object” setup ( Figure 2A bottom , “multiple object , ” after locking ) . This behavioral pattern was consistent across all the individual bats that we tested ( Figure S2 ) . Statistical analysis showed that the increase in inter-click angle was highly significant ( Figure 2A , bottom: one-way ANOVA: F>71 , p<10−8; post-hoc t tests: p<10−11 for comparing one-object experiments after locking versus no-object experiments; p<10−6 for multiple-object experiments after locking versus one-object after locking ) . In the multiple-object setup , the bats increased the inter-click angle significantly beyond the point of maximum slope ( i . e . , the maximum slope of the beam was lateral to the target; t tests: p<10−3 for comparing one-object experiments after locking versus multiple-object experiments after locking ) . This suggests that , at least in this case , the inter-click angle plays another role in addition to placing the maximum slope on target for optimizing localization . We propose that widening the angle between the beam axes of sonar click pairs serves to modulate the bat's field-of-view . During the last time-bin before landing ( Figure 2B , right-most point ) , the inter-click angle has increased on average by 14 . 5 degrees , compared to the mean angle in no-object experiments . When doing so , the point in the beam that was pointed to the center of the target was 2 . 5 degrees medial to the maximum slope . In the multiple-object experiment ( Figure 2B ) , unlike in the one-object setup ( Figure 1D ) , it seemed that the bats did not increase the inter-click angle abruptly ( when we used the same locking criterion ) , but instead began the approach to the landing sphere with a large inter-click angle , and gradually increased even further after the final locking onto the landing target ( Figure 2B ) . However , this gradual change may have been a result of temporal smearing that is specific to the multiple-object condition , and which is due to the difficulty in defining the exact time of “locking” in the multiple-object experiments: Although we defined sonar beam locking with reference to the landing sphere ( i . e . , when the average of the click pair was directed towards the landing target ) , the bats often locked onto the net's poles before locking onto the landing target ( the 10-cm sphere ) . This means that they could have been in a “locked” sonar mode ( locked onto a pole ) when we defined them as un-locked relative to the landing target ( see more details in the Discussion ) . We therefore tested an alternative sonar locking criterion for the multiple-object experiments , defining locking as the moment when the bats entered a corridor between the nets . This criterion revealed a clearer picture of the inter-click angle dynamics in the multiple-object situation ( Figure 2C ) : Well before passing between the nets , the bats used an intermediate inter-click angle ( 5 . 8±0 . 7 degrees wider than no-object ) , which is between the locked and un-locked one-object situations . When the bats approached closer to the net corridor , they rapidly increased the inter-click angle to nearly its final value; subsequently , after the bats entered the net corridor , another slight increase was observed , which brought the inter-click angle to an average value that was 14 . 5±2 . 0 degrees wider than in the no-object experiments . At the plateau , the center of the target was ∼2 . 5 degrees beyond the maximum slope ( t tests: p<10−3 for comparing one-object experiments after locking versus multiple-object experiments after locking ) . Maintaining such a high inter-click angle could possibly allow the bat to track both the target and the off-axis objects ( distal poles ) as the animal approaches landing—providing a potential strategy for target landing while avoiding collisions . Interestingly , when further analyzing data from the one-object experiments , we found that in some trials , especially when the bats flew a long trajectory before landing , the bats sometimes locked their sonar on the landing sphere , then redirected the beam away and later performed the “final” locking when starting the final approach . The “E-L” bar ( dark gray ) in Figure 2A represents the inter-click angle during these “early locking” instances . It shows that the bats increased the inter-click angle even when they only transiently locked onto the target ( during early locking , “E-L” ) . The widening of the inter-click angle in these instances was not as salient as in the final locking , probably because this beam-angle adjustment occurred for rather short periods of time ( only a few click-pairs ) , and when the bats were rather far from the target ( >1 . 5 m ) . In addition to the increase in inter-click angle , we found that Egyptian fruit bats decrease their emission intensity along the approach to landing ( Figure 3A–C ) . We always refer here to peak intensity ( see Materials and Methods ) , but since the duration of the sonar clicks is very constant , this is also highly correlated to the click's total energy . Because the bats in this experiment were free to choose the trajectory of landing , it was not always relevant to analyze the bat's distance to the target: for instance when a bat circles the target , it could be very close to it in terms of distance but very far in terms of time-to-landing ( and may in fact be echolocating in a different direction ) . We therefore examined the intensity versus time-to-locking ( Figure 3A–B ) , as well as intensity versus distance-to-target in trials in which the distance decreased nearly monotonically as the bat approached the landing sphere ( Figure 3C ) . Figure 3C shows six examples in which the bat flew directly to the target , exhibiting a salient reduction in intensity , with a 4–6 dB decrease with halving of the distance-to-target during the final approach ( Figure 3C , gray line , close to target ) . These results are consistent with reports in other bat species [28] , [29] . Interestingly , this decrease in intensity began only 80–100 cm before landing—similar to what was observed in laryngeal echolocators [28] . Thus , the intensity dynamics along the approach seem to be shared by clicking and laryngeal bats . In addition to increasing the inter-click angle , bats also increased the intensity of their clicks with environmental complexity . The intensity increased by 6 . 5±0 . 6 dB on average in the one-object experiments compared with the no-object experiments , and further increased by 2 . 6±0 . 8 dB on average in the multiple-object experiments—that is , a total intensity increase of 9 . 1 dB in the multiple-object versus no-object condition ( Figure 3D ) . These modulations of intensity could be used by the bat to maintain fixed signal energy directed towards the region of interest , compensating for changes in signal-to-noise ratio due to a widening field-of-view ( see Figure 4 , and next section ) . These differences in intensity were highly significant ( one-way ANOVA: F>108 , p<10−9; post-hoc t tests: p<10−33 for t test of one-object versus no-object; p<10−16 for multiple-object versus one-object; here we pooled together data from the approach phases before and after locking ) . Since we used a planar rather than a 3-D microphone array , and could not calculate the absolute emitted intensity , we performed explicit tests to control for the effects of bats' height , the distance from the microphones , and flight pitch ( see Materials and Methods ) . The increase in intensity , together with the increase in inter-click angle , both contribute to an increase in the effective area that is sampled by the bats via a single click-pair ( see next section and Discussion ) . Our two main findings—that Egyptian fruit bats increase their inter-click angle and also increase the click intensity with increased environmental complexity—suggest that the field-of-view scanned by the bat is under active control and adapted to the environment . These adaptive sonar signal changes served to increase the bat's field-of-view when the environment became more complex ( i . e . , contained more objects ) . To examine this notion further , we quantified the field-of-view scanned by the bat , assuming a constant ensonification-intensity level and calculating the change in the angle of the sector covered by the bat's beam . When we used the intensity at the crossing point of the two beams in the one-object setup as reference ( Figure 4 dashed lines , normalized intensity 1 , see Materials and Methods ) , we found that the angle of the sector scanned by the bat with a single click-pair increased by a factor of 2 . 18 in the one-object experiments in comparison to the no-object ( from 44 to 96 degrees ) , and by a factor of 2 . 73 in the multiple-object experiments in comparison to the no-object setup ( from 44 to 120 degrees , see Figure 4C versus 4A ) . Interestingly , the same intensity ( corresponding to a normalized intensity of 1 in Figure 4 ) is directed towards the crossing point of the two beams ( where the object of interest is positioned ) in both the multiple- and one-object setups , and it is the peak intensity ( directed forwards ) in the no-object setup . These modulations might thus reflect the bat's attempt to maintain a fixed energy impinging on the region of interest , compensating for the changes in signal-to-noise ratio due to the changes in field-of-view . The maximum distance ( range ) scanned by the bats also increased with environmental complexity , because detection range increases as the fourth root of the increase in intensity [30] . Thus , the 3-D “sensory volume” of space [31] that was scanned by the bats has increased at least 3-fold in the multiple-object versus the no-object experiment . We further examined several additional echolocation parameters in this set of experiments , and the results are summarized here . ( i ) We did not find any significant change in the beam width of the single clicks in the different environments . ( ii ) The bats did not significantly change the click repetition-rate in the multiple-object experiments in comparison to the one-object experiments ( i . e . , the intra-pair interval remained 23 ms on average and inter-pair interval was 93 ms on average ) . However , in the no-object experiments there was a small but significant decrease in the repetition rate , whereby the inter-pair interval increased by 8 ms ( 101 ms on average , t test of no-object versus one-object , for intervals >40 ms: p<10−10; Figure S3 ) . ( iii ) We tested the spectral content of the echolocation clicks ( recorded with a wide-band microphone ) only in the one-object experiments; hence we cannot exclude changes in the spectra of the clicks . However , spectral changes seem physiologically unlikely , considering the tongue-production mechanism of the brief lingual clicks . Thus , the most salient changes that we observed were changes in inter-click angle , and changes in click intensity . These two parameters changed in opposite directions along the approach path to an object ( inter-click angle increased while click intensity decreased during the approach ) , and both of these parameters increased substantially with environmental complexity .
The research findings presented here suggest that lingual ( click-based ) echolocation allows more adaptive control than previously reported . Egyptian fruit bats performing a landing task changed both their emission intensity and inter-click angle as they approached a target , in a manner that depended on both the environmental complexity and the behavioral phase . The increase in inter-click angle might serve two different functions: ( i ) Pointing the maximum-slope to the target: In the one-object setup , the increase in inter-click angle coincided with the moment of locking ( Figure 1C–D ) , thus representing a behavioral phase-transition that could serve the function of directing the maximum slope to the center of the landing sphere , in order to optimize stimulus localization [15] . In comparison , in the no-object setup , the bats aimed most of the energy forward , in the direction of the flight , by decreasing the inter-click angle ( Figure 2A , bottom ) . The narrow inter-click angle before locking , in the one-object situation , is very similar to the angle in the no-object situation , and might thus represent a narrow , forward focused field-of-view that is used before the final approach to the target . ( ii ) Changing the field-of-view: When shifting from the one-object to the multiple-object setup , the increase in inter-click angle was likely caused by the need to increase the field-of-view . In the multiple-object setup , the bats had to land on a specific target that was placed in the vicinity of other obstacles ( e . g . poles , nets ) . In this situation , the bat's own motion created very large and rapid angular changes in the directions to nearby objects , and hence the bats would need to increase the field-of-view in order to track these objects . Interestingly , the bats also decreased the emitted intensity while approaching the landing target ( within a given level of environmental complexity ) . Such intensity decrease was not reported in a previous study of Rousettus echolocation [27] , probably because they did not record bat signals during landing in that study . In our study , we observed a decrease in click intensity only during the last 80–100 cm before landing ( Figure 3C ) , which suggests that the intensity decrease is initiated only when the bat actually approached the landing sphere . Thus , Rousettus bats increase the field-of-view and concurrently reduce the emitted intensity when approaching landing . A similar behavior is exhibited by approaching laryngeal echolocators [24]: The calls in the terminal group of these bats have more energy in low frequencies , and thus a wider beam , but also lower peak intensity . Lingual echolocators ( e . g . , Egyptian fruit bats ) seem to have developed an alternative way to increase the effective beam width , which does not require them to change the spectral content of their emission . Instead , they change the scanning width by adjusting the angle between the axes of these two beams , and may treat the echoes returning from two consecutive clicks as a single “information unit” [15] . Such a strategy , which is based on adjusting the angular separation between two consecutive sonar emissions within a click-pair , has never been reported in any bat species to date , and it suggests an alternative adaptive mechanism in bat echolocation to sample a wider spatial region . Laryngeal echolocators are also known to steer their beams [13] and could thus also adjust the directional aim of successive sonar calls to control spatial sampling . However , there is no evidence for any laryngeal echolocator that constantly emits pairs of signals , similar to the Egyptian fruit bat; and accordingly , there is no evidence for any laryngeal echolocating bat that regards pairs of signals as their basic “sonar unit . ” In addition , Egyptian fruit bats are probably able to achieve such quick changes in beam steering by rapid tongue movements [15] , while changes in beam steering in laryngeal echolocators would probably require head movements , and would thus be slower than 20 ms . Thus , the field-of-view control strategy , suggested here for Egyptian fruit bats , might be a unique phenomenon among echolocating animals . The increase in emission intensity in the different environmental setups may represent an attempt by the bat to maintain fixed energy directed towards the region of interest , thus compensating for changes in signal-to-noise ratio due to changes in field-of-view . Figure 4A–B shows that the region of interest ( i . e . , the crossing-point of the right and left beams ) has approximately the same intensity in the one-object as in the multiple-object setups ( horizontal dashed lines ) . Interestingly , the peak intensity that is being directed towards the direction of interest in the no-object setup is identical to the crossing-point-intensity in the one-object and multiple-object setups ( Figure 4C , horizontal dashed line shows normalized intensity 1 ) . This could be interpreted as a principle of “conservation of signal-to-noise” in lingual bat echolocation , and can explain the seemingly paradoxical behavior of decreasing emission intensity when performing a search task ( in the no-object setup ) . In a previous report [15] , we described a trade-off between detection and localization in the Egyptian fruit bat , whereby detection is maximized by pointing the peak of the beam towards an object , while localization is optimized by pointing the maximum-slope towards the object . This tradeoff predicts that the bat will direct its sonar beam towards an object of interest at an angle that rests between the peak and the maximum slope . In our current multiple-object experiment , the bats deviated from this principle by consistently increasing the inter-click angle such that they directed the beam towards the target at points beyond the maximum slope of the beam ( Figure 4A , see “+” ) . This finding is surprising , because it means that target localization cues were now likely diminished . In light of these results , as well as the other results presented in this article , we believe that a new dimension has to be considered , thus introducing a three-way tradeoff between ( i ) detection , ( ii ) localization , and ( iii ) angular scanning ( modulated via changes in field-of-view ) . We suggest that in a complex environment , the need to scan the area around the landing-point , and to increase the field-of-view , is sufficiently important for the bats to reduce localization accuracy . Note that detection was actually not reduced by the increase in the inter-click angle in the more complex environments , because the bats also increased the click intensity , possibly as a compensatory mechanism ( Figure 4A–C dashed lines ) . What is the functional relevance of the sonar field-of-view ? All the previous studies that were conducted on beam steering in laryngeal echolocating bats suggested that , despite their broad emission beams ( 60–70° width at −3 dB [24] , [32] ) , these bats carefully direct the center of their beam towards the object of interest [13] , [14] , [24] , [32] . Our previous study of sonar beam steering in Egyptian fruit bats showed that these lingual echolocators direct the center of their beam-pair onto the target [15] , reminiscent of the individual calls of laryngeal echolocators . The behavior observed in the current study suggests that Egyptian fruit bats collect sensory information also from their acoustic periphery . In the multiple-objects experiments , the bats exhibited a wide repertoire of behaviors before landing on the target ( see details in Materials and Methods ) . In many cases ( ∼30% of trials ) the Egyptian fruit bats only locked onto one of the poles , or occasionally did not lock on any of the poles while entering the corridor between the nets . We cannot completely exclude the possibility that the bats were relying on spatial memory ( see Materials and Methods ) , but data from these trials imply that the bats can localize an object to some extent without the need to point the center of the beam-pair towards it . Increasing the field-of-view in order to follow objects near the landing target thus makes perfect sense from the bat's point of view . In summary , our findings reveal two new aspects of adaptive control in lingual bat echolocation , namely the ability to change emission intensity as well as changing the inter-click angle between sequential emissions . The ability of lingual bats to change the inter-click angle reveals a new strategy for bats to actively control the field-of-view that they scan . Adjustment in field-of-view could also theoretically be exploited by laryngeal echolocators through movements of the head , mouth opening , and spectral changes in sonar emissions . The Egyptian fruit bat's directional aim of tongue click pairs demonstrates a new parameter of acoustic control in animal sonar . We suggest that environment-associated changes in emission intensity seem to be related to changes in field-of-view , and can compensate for decreases in signal-to-noise ratio due to changes in field-of-view . Further , our results suggest a three-way trade-off between three goals that a bat has to fulfill with its echolocation in a target-landing task: The detection of an object of interest , its accurate localization , and controlling the field-of-view that is being scanned by the bat . We believe that further studies of sensory trade-offs in echolocating bats will shed new light on bat echolocation—and more generally , on sensory constraints in active-sensing systems .
All experimental procedures were approved by the Institutional Animal Care and Use Committees of the Weizmann Institute of Science and the University of Maryland . Five adult Egyptian fruit bats ( Rousettus aegyptiacus ) were trained to detect , localize , and approach a polystyrene sphere ( 10-cm diameter ) that was mounted on a vertical pole positioned inside a large flight-room ( 6 . 4×6 . 4×2 . 7 m; Figure 1A ) . The target's size mimics the size of some fruits eaten by these bats in nature , such as mango . To minimize sound reverberations , the walls of the room were covered with acoustic foam and the pole was covered with felt . In order to ensure that the bats were relying solely on echolocation to perform the task , we took the following precautions: ( i ) To exclude the possibility of using visual cues , the target was painted black and the room was in complete darkness ( illuminance <10−4 lux ) . The experimenter inside the room wore night-vision-goggles with infrared illumination . ( ii ) To prevent use of olfactory cues , the bats were food-rewarded only after landing on the target . The target was also cleaned with soap and water after every three trials to remove any possible odors that remained on it due to the contact with the bat . ( iii ) After every trial , the target was randomly re-positioned inside the room , both in the horizontal and in the vertical planes ( the pole had a telescopic mechanism that allowed changing the target height ) . It took the bats ∼4 wk in order to learn the task and once they learned it they always succeeded in landing on the target . The basic setting included only the landing target ( 10-cm polystyrene sphere ) in the flight room . We also tested two alternative settings: ( i ) In 56 randomly interspersed trials we removed the landing target from the room , which made the bats eagerly fly in search for the target . We call these experiments the “no-object” experiments . ( ii ) In 54 trials , we added two nets mounted on 4 poles on both sides of the landing target ( Figure 2A top , right ) . The distance between the nets randomly varied between trials ( in the range of 0 . 6–1 . 6 m ) and so did the position of the landing target and the angle of the nets in relation to the target . The bats learned to correctly land on the target between the nets—within 3–4 trials ( which were not counted within these 54 trials ) ; nevertheless , bats still occasionally landed on the poles even after many more trials . They were only rewarded for landing on the original target ( sphere ) . Because these experiments involved five salient objects ( 1 target+4 poles ) , they were termed here “multiple-object” experiments . The bats exhibited a wide behavioral repertoire in the multiple-object experiments: In some trials , they behaved similarly to the behavior described for the laryngeal echolocator , Eptesicus fuscus [14] . In those previous experiments , E . fuscus were trained to fly through a hole in a net , and they typically scanned both sides of the hole ( pointing the peak of the beam to each edge of the hole ) before flying through it . In the equivalent trials in the current study , Egyptian fruit bats locked the center of their click-pairs on both poles that outlined the opening of the net corridor , and only subsequently they flew through the corridor . In other trials , the Egyptian fruit bats either locked onto one of the poles before landing or did not lock on any of them . Because the bats had the opportunity to fly around the poles and nets before approaching them , and could thus learn their spatial locations , we could not completely exclude their relying on spatial memory . However , it is not likely that this was the only factor facilitating their approach , because the location and layout of the setup was always randomly changed between trials , and the bats did not always scan the setup before approaching the landing target . The bat's average flight speed was negatively correlated with the environmental complexity ( 1 . 2±0 . 9 m/s in the multiple-object experiment , 1 . 9±1 . 2 m/s in the one-object experiment , and 2 . 4±1 . 0 m/s in the no-object experiment; mean ± s . d . ; p<0 . 001 for all three t test comparisons , and F>880 , p<10−10 in a one-way ANOVA test ) . This difference in flight speed remained when we analyzed the speeds only for pre-locking or only for post-locking epochs . We believe that the changes in flight speed were a result of the different maneuverability situation , due to the difference in the environmental complexity . In the multiple-object setup , the flight-speed likely decreased also because of the need to slow down in order to allow more time to scan the setup ( in the multiple-object experiments , the bats typically slowed their flight before entering the net corridor , or when scanning the poles ) . Since the bats had the possibility to pre-scan the room , they could potentially adjust their speed to the expected maneuverability conditions , and this is likely why we did not see a change in flight speed between the pre-locked and post-locked situations . The bats' echolocation behavior was recorded with an array of 20 microphones spaced 1-m from each other around a rectangular supporting frame ( 5 . 3×5 . 2 m ) , at a height of 90 cm above the floor ( Figures 1A and 2A , top: black dots around the circumference of the room show microphone locations ) [32] . The signal from each microphone was amplified and fed into a band-pass filter centered around 35 kHz , with a frequency response that matches the frequency content of the Rousettus sonar click ( see details in ref . [15] ) . Next , the signal was fed to an electronic circuit which extracted the envelope of this band-passed signal . The envelope was then low-pass filtered and digitized into a data-acquisition computer . Finally , the maximum value of this signal was translated into a dB scale in which analysis was performed . In order to control for changes in click spectra , in ∼20 trials of the one-object experiment we have recorded the audio using three wideband ultrasonic microphones positioned on the floor ( sampled at 250 kHz/channel ) . To ensure that we were only using high-quality data , we included only clicks that were clearly above noise level in at least five microphones of the array . In addition , we excluded beam measurements that were either too wide or too narrow relative to the overall distribution of >5 , 000 beam patterns recorded during >300 trials , because deviant widths led us to suspect a recording artifact due to temporary noise in some of the channels . To this end , we measured the width of the beams [15] , and accepted only clicks with: 30°<beam width<120° . This resulted in exclusion of ∼6% of the clicks . In total , we analyzed here 5 , 144 sonar clicks from 346 behavioral trials in 5 bats ( 56 no-object trials , 236 one-object trials , and 54 multiple-object trials ) . We only analyzed clicks that occurred more than 250 ms before landing , because later clicks were emitted when the bat was too close to the target ( closer than 15 cm on average ) , where any angular calculation of direction-to-target would suffer from very high error . This typically corresponded to excluding the last two click-pairs in the trial . All 20 signals ( from 20 microphones ) were first segmented to include vocalizations and exclude echoes . Then , the intensity at each microphone was corrected for spherical loss and atmospheric attenuation according to the measured position of the bat and the temperature and humidity in the flight room [32] . The click intensity was then taken as the maximum of these 20 intensity values . In order to calculate the beam direction , we averaged the direction of all microphones that recorded intensities of at least 0 . 8 of the maximum intensity or higher . This was done after smoothing the raw beam intensities with a 3rd-degree Golay-Savitzky filter [15] . Taking into account the system's noise and our beam estimation method , the error in beam-direction estimate was ∼5 . 5° ( see ref . [15] ) . The inter-click angle was taken as the difference between two consecutive beam directions within a pair of clicks . The pairs are easy to recognize and can be mathematically defined as two clicks with a time-interval of less than 35 ms between them ( Figure S3 ) . Two high-speed digital video cameras ( Photron , set with a frame rate of 125 frames per second ) , synchronized with the ultrasonic array , were used to record the flight of the bats . The direct-linear-transform algorithm was used to measure the three-dimensional location of the bat and other objects in the room , using the two camera views . We defined a “locked” click-pair as a pair in which the vector-average direction of its two clicks was <30° relative to the target ( see example in Figure 1A; locking time is denoted by arrow ) . The 30° criterion was chosen since it corresponds to twice the asymptotic standard deviation of all click-pair vector averages , just before landing [15] . This is the same locking criterion as used in our previous study [15] . We tested two additional criteria for locking threshold ( 20° and 40° , unpublished data ) , which did not affect the results . Because our microphone-array was planar , we could not estimate the absolute emission intensity ( sound pressure level ) . In order to be sure that the differences we found in the emitted intensity were not a result of some recording artifact , we tested whether the measured intensity of echolocation clicks is correlated with several flight-trajectory parameters: ( i ) distance from microphones ( r = −0 . 03; n . s . ) ; ( ii ) height of flight ( r = 0 . 02; n . s . ) ; ( iii ) flight pitch ( r = 0 . 06; n . s . ) . None of these parameters showed any correlation with the emission intensity . We could not control for the head's pitch angle , but an examination of the raw videos did not reveal any tendency of the bats to systematically change head pitch in an environment-dependent manner . The sensitivity of the array could not have changed between setups because the multiple-object and the no-object experiments were interspersed in time between the one-object experiments . To control for possible sound-occlusion effects due to the specific layout of objects in the room ( e . g . , the target may have blocked a specific microphone and thus may have artificially enlarged the measured inter-click angle ) , we re-ran the entire analysis , taking for the direction of the beam the direction of the single microphone that recorded the peak intensity ( rather than weighing over several microphones ) . This analysis did not affect our findings . It should be noted that such an artifact is not likely for other reasons as well: ( i ) If the angle increase was a result of an “occlusion artifact , ” the angle should have increased gradually ( rather than abruptly ) in the one-object experiments . ( ii ) If it were an artifact , we would not have observed a widening of the angle when the bat was far from the target in the pre-locked situations ( “E-L” bar in Figure 2A , bottom ) . In order to verify that the nets were not blocking sound waves and possibly causing some acoustic artifacts , we estimated the attenuation caused by the nets , by comparing the emission recorded from a test speaker without nets to that recorded through the nets; no difference was found for an impinging angle of 90° ( i . e . , when emission was perpendicular to the nets ) . Because each bat produced its individual typical emission intensity and unique inter-click angle , we always normalized data from each bat separately before averaging across all bats . This means that we first calculated the average ( intensity or inter-click angle ) in the no-object setup and then calculated the average change relative to this value in the different setups ( one-object and multiple-object ) or different behavioral phases ( unlocked versus locked ) . We next calculated the average normalized change for all bats in each of the experimental paradigms . Unless stated otherwise , all the data were normalized in comparison to the one-object condition ( rather than to no-object condition ) , because we had almost 5 times more data-trials for the one-object experiments , which provided us with a smooth , robust baseline to compare to . | Most sensory systems have an active component , i . e . driven by an animal's behavior , which contributes directly to its perception . For example , eye movements are important for visual perception , sniffs are crucial for olfactory percepts , and finger movements for touch percepts . A classic example of an active-sensing system is bat echolocation , or biosonar . Echolocating bats actively emit the energy with which they probe their surroundings , and they can control many aspects of sensory acquisition , such as the temporal or spectral resolution of their signals . A key open question in bat echolocation concerns bats' ability to actively change the area scanned by their emitted beam . Here , we used a large microphone array to study the echolocation behavior of Egyptian fruit bats . We found that these bats apply a new strategy to alter the area scanned by their beam; specifically , bats changed their acoustic field-of-view by changing the direction of consecutively emitted beams . Importantly , they did so in an environment-dependent manner , increasing the scanned area more when there were more objects in their surroundings . They also increased their field-of-view when approaching a target . These findings provide the first example for active changes in sensing volume , which occur in response to changes in environmental complexity and target-distance , and they suggest that active sensing of space is more flexible than previously thought . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"biology",
"sensory",
"perception",
"neuroscience"
] | 2011 | Active Control of Acoustic Field-of-View in a Biosonar System |
One of the most conserved features of the invasion process in Apicomplexa parasites is the formation of a moving junction ( MJ ) between the apex of the parasite and the host cell membrane that moves along the parasite and serves as support to propel it inside the host cell . The MJ was , up to a recent period , completely unknown at the molecular level . Recently , proteins originated from two distinct post-Golgi specialised secretory organelles , the micronemes ( for AMA1 ) and the neck of the rhoptries ( for RON2/RON4/RON5 proteins ) , have been shown to form a complex . AMA1 and RON4 in particular , have been localised to the MJ during invasion . Using biochemical approaches , we have identified RON8 as an additional member of the complex . We also demonstrated that all RON proteins are present at the MJ during invasion . Using metabolic labelling and immunoprecipitation , we showed that RON2 and AMA1 were able to interact in the absence of the other members . We also discovered that all MJ proteins are subjected to proteolytic maturation during trafficking to their respective organelles and that they could associate as non-mature forms in vitro . Finally , whereas AMA1 has previously been shown to be inserted into the parasite membrane upon secretion , we demonstrated , using differential permeabilization and loading of RON-specific antibodies into the host cell , that the RON complex is targeted to the host cell membrane , where RON4/5/8 remain associated with the cytoplasmic face . Globally , these results point toward a model of MJ organization where the parasite would be secreting and inserting interacting components on either side of the MJ , both at the host and at its own plasma membranes .
Invasion by Apicomplexa is an essential step of the pathologies associated with these protozoan parasites that include Plasmodium spp . , the causative agents of malaria , as well as Toxoplasma gondii , responsible for human and animal toxoplasmosis . The invasive stages of these parasites share a highly conserved architecture , including a cytoskeleton-associated original pellicular complex , and two types of vesicular apical organelles ( micronemes and rhoptries ) that participate to the invasion process through the exocytosis of their contents in a sequential manner [1] . Host cell invasion has been well described at the ultrastructural level , but the precise molecular interactions and the specific role of the exocytosed parasite proteins are still poorly understood . Proteins located on the surface of the parasite probably mediate the initial interaction with the target cell . This is followed by an intimate contact between the apical tip of the parasite and the host cell membrane , called the moving junction ( MJ ) [2] . This singular structure , likely linked to the subpellicular cytoskeleton motor of the parasite , might serve as a support to propel the parasite into the parasitophorous vacuole ( PV ) that forms inside the host cell . To do so , the MJ rapidly turns into a ring that is moved backward along the parasite during invasion and ends up at the posterior end of the invaded parasite at the end of the process . Despite a number of investigations having led to the discovery of a variety of putative parasite adhesive molecules secreted from micronemes , and of an original acto-myosin based motor for gliding motility [3] , the process of invasion itself ( i . e . MJ-dependent host cell entry ) , remains a major conundrum . Indeed , although the morphological features of the process have been described 30 years ago [2] , the MJ was , up to a recent period , completely unknown at the molecular level . The major reason for this was its transient nature , since host cell invasion is a very rapid process ( a few seconds ) , and therefore isolating the structure was not possible . Rhoptries are elongated organelles composed of a bulbous body that tapers into a thin duct-like neck . Rhoptries empty their contents apically during the invasion process , after microneme exocytosis , and their contribution to invasion was considered mostly as providing building material for the developing PV , since proteins of the bulb of the rhoptry ( ROPs ) were found associated with the nascent vacuole membrane ( for a review see [4] ) . Recently , an unexpected function of the rhoptries in MJ formation arose from the discovery that one rhoptry neck protein ( RON4 ) was associated to the MJ [5] , [6] . It was proposed that MJ formation would derive from a cooperation between i ) newly discovered RONs located in the rhoptry neck and ii ) the micronemal protein AMA1 [5] . Numerous lines of evidence suggest that the conserved AMA1 protein plays a central role during invasion of Apicomplexa . For instance , AMA1 has been shown to be essential for Plasmodium merozoites and Toxoplasma tachyzoites [7] , [8] . In T . gondii , AMA1 and RON4 have been found to be associated in a complex in vitro and they localize precisely at the MJ during cell invasion , although a direct association of the two proteins has not been demonstrated in vivo [5] , [6] . The isolation of RON4 from parasite extracts by affinity purification led us to the simultaneous purification of the rhoptry neck protein RON2 and protein TwinScan_4705 ( annotated also 583 . m00636 ) [6] , which was later shown to be also a RON ( RON5 , P . Bradley personal communication ) . Like AMA1 , RONs are conserved throughout the Apicomplexa including Plasmodium spp . , and they are not found outside this phylum . AMA1 and RONs are stored in two distinct compartments that release their content sequentially during invasion . Cross-linking experiments on invading parasites showed that the interaction of AMA1 with RONs takes place during invasion and is not the result of non-specific or indirect binding occurring in the parasite lysate during IP [5] . One intriguing question is how the micronemal protein AMA1 and the complex of rhoptry neck proteins RON2/RON4/RON5 avoid interacting in the secretory pathway . Another important question is how these proteins are organized at the MJ . The microneme protein AMA1 has been characterized structurally and appears to be translocated as a type-1 transmembrane ( TM ) protein in the tachyzoite plasma membrane [9] , [10] . On the contrary , the topology of the RONs at the MJ is still obscure and several important questions remain unanswered . Are RONs directly or indirectly linked to the parasite surface ? Could they be binding to a host cell receptor or , as we speculated previously [6] , are they directly inserted into the host cell membrane to serve as a receptor for AMA1 ? Here , we describe an additional partner of the previously characterized AMA1/RON2/4/5 complex named RON8 . We also show that the complex may be assembled as pro-proteins but that a distinct timing of biosynthesis between MICs and RONs precludes the association of RONs with AMA1 before secretion . Furthermore , we demonstrate that RONs are exported to the host cell membrane , RON/4/5/8 being exposed to the host cell cytosol and RON2 being probably an integral membrane protein that displays a privileged interaction with AMA1 . These results provide an important clue to understand how such a crucial structure for the invasive and developmental processes of the parasite is built and organized .
In order to further refine the molecular characterization the MJ complex of T . gondii , we searched for additional proteins co-immuno-purified ( IP ) by the anti-RON4 antibody matrix , as previously described [6] . The RON4-associated proteins were subjected to mass spectrometric analysis . As in our first analysis , we detected two principal bands at ∼120 kDa and ∼100 kDa , which corresponded to RON2 , RON4 and RON5 ( Figure 1A ) . In addition to the proteins of the MJ already known to be associated with each other ( RON2 , RON4 , RON5 and AMA1 ) , mass spectrometry analysis identified peptides from proteins that are described in Table S1 . Peptides from proteins originated from the secretory organelles involved in invasion ( microneme and rhoptry ) have retained our attention . First , peptides from two microneme proteins MIC1 [11] and MIC3 [12] were detected . However , Western blot and reverse IP analysis using anti-MIC3 or anti-MIC1 antibodies did not confirm a specific interaction of these proteins with the MJ complex proteins ( data not shown ) . Second , we found peptides from TwinScan_0092 ( 80 . m02161 ) and TwinScan_2001 ( 541 . m00141 ) , two predicted Toxoplasma proteins that had also been detected in the proteomic analysis of the rhoptries [13] . TwinScan_0092 predicts a protein of 49 kDa that is not localising at the MJ but was instead found to be a new dense granule protein [14] . Concerning TwinScan_2001 , a previous study using an antibody raised against a specific peptide had localised it to the apicoplast by IFA [13] , although it does not possess any bona fide plastid-targeting element in its amino acid sequence . We then decided to reassess its subcellular localization by generating a specific polyclonal antiserum directed against a recombinant TwinScan_2001 protein corresponding to the central part of the protein ( Figure 1D ) . This antibody ( anti-Tw2001 ) reacted on Western blot with a major band of about to 250 kDa and several minor bands of lower molecular mass ( Figure 1C ) , that were also detected with an additional serum raised against another region of the protein ( not described here ) , but were absent when probed with the pre-immune serum ( Figure 1C , Figure S1A ) . By IFA , the anti-Tw2001 serum recognized an antigen co-localized with RON4 in intracellular parasites ( Figure 1B ) , suggesting that TwinScan_2001 was a new rhoptry neck protein that we named RON8 . To further verify that RON8 is associated with the AMA1/RON2/4/5 complex , we performed an IP using the anti-Tw2001 serum ( referred to as anti-RON8 throughout the manuscript ) , as described previously for RON4 [6] and showed co-purification of RON8 , AMA1 and RON4 ( Figure 1C ) . The formation of a stable complex in 1% NP40 and 1 M NaCl conditions , containing RON2/4/5/8 and AMA1 , was further confirmed by co-IP of all members after affinity chromatography using either of the specific anti-RONs ( data not shown ) . The complete coding sequence of RON8 was determined ( GenBank accession number ACK57540 ) and showed that it coded potentially for a 2979 amino acids-long protein , with a theoretical molecular mass of 329 kDa . A putative signal peptide was found at position 1–29 . PROSITE search yielded no obvious sequence motifs . A search of the GenBank non-redundant database and ApiDB showed that RON8 is unique to Toxoplasma and Neospora among Apicomplexa ( in contrast to other MJ proteins ) and is not found in other organisms . We have previously shown that RON4 is associated with the MJ during invasion [6] , here we examined if RON2 , RON5 and RON8 would also follow the MJ . We first generated antisera specific of RON2 , RON4 and RON5 . For RON2 , two sera were prepared against different regions of the protein produced as recombinant proteins named RON2n and RON2c ( see Figure 1D ) . An anti-serum against the N terminal part of RON4 ( RON4n ) was also produced . The specificity of the sera was first analyzed by Western blot on whole tachyzoite lysates ( Figure S1A ) . All sera recognized in non-reduced condition a band at about the predicted size ( RON2: 155 kDa , , RON5: 179 kDa , RON8: 329 kDa ) . No detection was observed with pre-immune sera . In reduced condition , the anti-RON2c , anti-RON2n , anti-RON4n , and anti-RON5 recognized proteins that migrated faster , indicating that , as previously shown for RON4 ( [6] and Figure S1A ) , RON2 and RON5 are also sensitive to reduction of disulfide bonds ( discussed later ) . The sera were also analyzed by IFA on intracellular parasites . As shown in Figure 2A , all the anti-RONs labelled the neck of the tachyzoites rhoptries , as indicated by co-localisation with RON4 and RON8 . Throughout the study , the anti-RON4 T5 4H1 [15] and the anti-AMA1 CL22 mAbs were used systematically , except when specified . On invading parasites , in permeabilization conditions optimized to detect only the material secreted by the parasite [1] , we showed that anti-RON5 and anti-RON8 recognized exclusively the characteristic ring-shaped MJ ( Figure 2B ) . In contrast , in these conditions both anti-RON2 antibodies failed to react ( Figure 2B , lower panel ) . We have shown previously that when the PVM has pinched off the host cell , the MJ can still be detected at the posterior pole of the parasite for a few hours and is characterized by a dot-like signal with anti-RON4 mAb [6] . Again , we showed using specific antibodies that RON5 and RON8 could be found together at the same location , but not RON2 ( data not shown ) . Since cytochalasin D ( Cyt-D , an inhibitor of actin polymerization ) -treated parasites form a “static” junction that is labelled by anti-RON4 [6] but not translocated to the posterior end of the tachyzoites [16] , [17] , we tested if all the RONs could be immunolocalized at the junction in these conditions . We found that after Cyt-D treatment , in addition to RON4 [6] , all the proteins of the complex , this time including RON2 ( yet only with the anti-RON2c antibody ) , could be detected at the same location ( Figure 2C ) . The detection of RON2 in these conditions could be explained by the fact that the Cyt-D treatment had improved the accessibility of the protein to the antibody , either because it destabilised some link of RON2 with Cyt-D sensitive structures of the host or , more simply , that it blocked the junction in an early stage where the protein is more accessible . Overall , this is strengthening the idea that RON2 , RON4 , RON5 and RON8 are present together in the MJ complex during invasion . The generation of antisera against the individual members of the MJ complex allowed us to analyse more precisely the RONs and their interactions by IP using different conditions for solubilization of the parasite . After lysing the parasites in 1% NP40 ( the condition used to immunopurify the complex [6] ) , all members of the complex were recovered using each of the anti-sera available , as exemplified in Figure 3A with an IP using the anti-RON2n serum followed by Western blot analysis of each member of the MJ complex . We then tested the stability of the complex upon tachyzoite lysis in 0 . 6% SDS followed by heat denaturation . In these conditions , only the interaction between AMA1 and RON2 was maintained after IP with either anti-RON2n or anti-AMA1 ( Figure 3A ) and no interaction between the others RONs was observed ( ie using anti-RON4n , anti-RON5 and anti-RON8 , Figure S1B ) . These results were confirmed by comparing the profiles obtained after metabolic labelling of intracellular parasites with [35S]-methionine/[35S]-cysteine , and lysis in either 0 . 6% SDS or 1% NP40 , followed by IP with anti-RONs antibodies ( Figure 3B ) . A similar profile in which all members of the complex were detected was obtained after IP in 1% NP40 whatever the antibody used ( left panel ) . In contrast in 0 . 6% SDS , while the anti-RON4 , anti-RON5 and anti-RON8 immunopurified only the corresponding protein , the anti-RON2n and anti-AMA1 immunopurified both AMA1 and RON2 ( right panel ) . These results clearly indicated that the whole complex was not maintained in 0 . 6% SDS , but that AMA1 and RON2 proteins interact together particularly strongly , independently of the other MJ proteins . Most T . gondii rhoptry bulb proteins described so far are synthesized as pro-proteins that are subjected to removal of their N-terminal pro-region by proteolytic cleavage during traffic to the organelle . To determine whether the RONs are also processed , we studied their biosynthesis and maturation by pulse-chase metabolic labeling with [35S]-methionine/[35S]-cysteine followed by IP with anti-RONs antibodies ( Figure 4A ) . The infected cells were lysed and boiled in 0 . 6% SDS to avoid co-purification of the whole complex . For RON2 , after 20 minutes of pulse , a protein of ∼150 kDa ( reduced ) was immunoprecipitated , which is consistent with the predicted size of RON2 after the removal of the signal peptide; a minor band was also found at ∼120 kDa . After one hour of chase , the 150 kDa disappeared and the 120 kDa band was the major one detected . A 65 kDa band after chase and a slightly slower migrating one in the pulse corresponded to AMA1 ( as described above ) and proAMA1 respectively ( see below ) . In non-reduced condition , the 150 kDa band was detected both in pulse and chase fractions , indicating that RON2 is processed and that the two fragments might be linked by internal disulfide bonds ( several cysteines are present in both fragments ) . For RON5 , after 20 minutes of pulse , a major protein of ∼180 kDa ( unreduced ) was immunoprecipitated , which is consistent with the predicted size of the protein . After one hour of chase , the 180 kDa product almost disappeared and a ∼150 kDa band was detected instead . In reduced condition , the 180 kDa form was also detected in pulse , while a ∼110 kDa form was immunoprecipitated after one hour of chase . A band of ∼30 kDa was also detected by Western blot on whole tachyzoites in reduced condition ( data not shown ) and was recovered by IP ( Figure S2 ) . These results indicated that RON5 is cleaved at least at two sites , one processing event resulting in removal of a pro-sequence ( as for many ROP proteins ) , and another processing event yielding two polypeptides possibly bound by a disulfide bond ( as for RON2 ) . Concerning RON8 , a processing event was also detected in reduced and non-reduced conditions , indicating that RON8 was also subjected to removal of a pro-sequence . Pulse-chase experiment for RON4 also showed that it is expressed as a pro-protein ( ∼120 kDa reduced condition and ∼145 kDa unreduced ) that is cleaved to yield a mature protein of ∼110 kDa ( reduced ) or ∼120 kDa ( unreduced ) . One additional minor band of lower molecular mass was also sometimes present . The persistence of the immature form of RON4 after one hour of chase indicated that RON4 was only partially matured . This could be linked to the fact that , as shown before [6] , [13] , part of RON4 is secreted in the PV ( arrow in Figure 2A ) and therefore avoids the rhoptry-specific processing compartment . Serendipitously , the generation of a transgenic parasite cell line expressing a Ty-tagged version of RON4 ( see Text S1 ) that was , for unknown reasons , entirely secreted in the vacuolar space ( Figure S3A ) and remained entirely unprocessed ( Figure S3B ) , strengthened this hypothesis . In order to determine in which compartment the RONs were processed , we then generated antibodies directed against the RON8 pro-peptide . As for all rhoptry proteins described so far , this latter was assumed to be located N-term and cleaved by the protease TgSUB2 [18] . Three putative TgSUB2 cleavage sites were found in RON8 , two in RON5 and one in RON2 ( Figure 1D ) . We therefore raised antibodies against a peptide spanning RON8 AA 1-91 , located before the first SFVE motif of the RON8 sequence ( Figure 1D ) . IP using anti-proRON8 demonstrated the specificity of the anti-proRON8 for the immature form of RON8 ( Figure S4A ) , whereas IFA showed reactivity restricted to the characteristic pre-rhoptry compartment ( Figure S4B ) , which corresponds to the nascent rhoptries of daughter parasites during endodyogeny . The cleavage of all RONs beyond the ER was showed by pulse/chase analysis in the presence of the Golgi transport-inhibiting drug brefeldin A ( BFA ) , pro-RONs remaining the only forms of the proteins at the end of the chase ( data not shown and Figure 4B ) . Since RONs undergo a proteolytic maturation , we analyzed if they could bind as immature proteins or if processing was required for this binding . To this end , we analyzed the MJ complex by pulse-chase experiments , followed by IP in lysis conditions known to preserve the association of the complex ( using 1% NP40 ) . As shown in Figure 4B , after a 15 min pulse , the immature forms of RON2 , RON8 and RON5 could be recovered after IP with the anti-RON4 monoclonal . Similarly , and as a complementary approach , IP of the pro-forms of the other MJ partners was achieved using anti-RON2n , anti-RON5 or anti-RON8 ( data not shown ) . To confirm these data , we checked for the association of the complex in BFA-treated cells that would express only immature radiolabeled RONs . As expected , in presence of the drug , proRON2 ( 150 kDa ) , proRON8 ( 329 kDa ) , proRON5 ( 180 kDa ) and proRON4 ( 120 kDa , reduced ) were the only species found at the end of the chase and co-precipitated together ( Figure 4B ) . Pulse-chase with anti-AMA1 serum confirmed that AMA1 was also processed during traffic [19] , and that proAMA1 could also associate to pro-RONs ( Figure 4B , right panel ) , a result which was also observed in pulse chase experiment using anti-RON2n ( Figure 4A , left panel ) . Overall , these results show that all known members of the AMA1-RONs complex could associate together as pro-proteins in vitro . Since AMA1 and RON2/4/5/8 could interact together as pro-proteins in vitro and would follow the same secretory pathway ( i . e . rough ER and Golgi apparatus ) before being packaged in their respective compartments in the parasite , we raised the question of how the micronemal protein AMA1 and the complex of rhoptry neck proteins RON2/4/5/8 could avoid interacting before secretion . One possibility would be that they are synthesized sequentially and never coexist in the same compartment . We checked this hypothesis by IFA . Since maturation of both MICs and RONs occurs with rapid kinetics ( MICs mature within 15-60 min , [20] and Figure 4A ) and the pro-sequence is only transiently detected by IFA , detection using anti-propeptide antibodies faithfully reflects the timing of their synthesis . We therefore performed double IFA with the rabbit polyclonal anti-pro-RON8 and the antiserum raised by Hehl et al . [9] against a peptide corresponding to the AMA1 pro-sequence ( data not shown ) . Unfortunately , this anti-proAMA1 serum gave a very low signal/noise ratio and no significant data could be obtained with this probe . Since we had previously studied the fate of other microneme prodomains and showed for example that 80% of the parasites co-expressed both pro-forms of microneme proteins M2AP and MIC3 simultaneously [21] , we reasoned that AMA1 could follow the same scheme and therefore decided to use the mouse anti-proMIC3 and the rabbit anti-proM2AP sera instead . Interestingly , no colocalization of proRON8 and proMIC3 was ever found and both markers were only observed simultaneously in ∼7% of the parasites ( ±2% , mean±SEM of 3 independent experiments; usually in very large parasites in mid-stages of endodyogeny ) ( Figure 4C , upper panels ) . We then took advantage that mAb anti-RON4 did not label the mature rhoptries but only the pre-rhoptries of dividing parasites upon formaldehyde fixation and triton permeabilization [6] , to compare the timing of synthesis of RON4 with that of M2AP using rabbit anti-proM2AP . Dual staining using rabbit anti-proM2AP serum and anti-RON4 showed that pre-rhoptry RON4 staining was almost never associated with a proM2AP staining in the same parasite ( 8 . 1%±3% , mean±SEM of 3 independent experiments ) and apparently not in the same compartment , while RON2 , RON5 and RON8 were systematically detected simultaneously with RON4 in pre-rhoptries in the same conditions ( data not shown ) , confirming that RONs and MICs biosynthesis are asynchronous . This would allow MICs and RONs to reach their correct destination without interacting before secretion . As reported above , RON 2 , 4 , 5 , and 8 are found at the MJ by IFA , but their precise location respectively to the parasite or host cell membrane is not known . We thus sought to determine which membrane these RONs were associated with . First , we observed that in IFA on parasites invading cells in the presence of Cyt-D , the rhoptry protein ROP1 was sometimes detected on cells in the absence of any surrounding parasite ( Figure 5A ) , suggesting that it would either correspond to an abortive invasion after secretion of the rhoptry content or that the parasite has been mechanically removed by the washes during the experimental procedure . Abortive invasion has been previously documented in a recent mathematical model showing that approximately 55% of the parasites detach within 5 min of initial attachment , but this paper did not conclude on whether the moving junction was built or not before detachment [22] . We thus assessed the presence of RONs in this particular situation . Dual IFA showed that all RONs could usually be detected as a punctuate signal , near the point of contact where the rhoptry content had apparently been initially discharged , as detected by anti-ROP1 staining in the host cell permeabilized with saponin ( Figure 5A ) . RON2 was only detected with anti-RON2c , as reported above . In contrast , no signal was obtained with anti-MIC2 and neither with anti-SAG1 ( directed against the major surface antigen of T . gondii ) , indicating that the signal obtained with anti-RONs was not due to the presence of residual membrane fragments of the parasite ( data not shown ) . Instead , this appeared to reflect a specific association of the RONs with the host cell membrane . It is also to note that we could only rarely detect an AMA1 signal in these conditions . Indeed , quantitative analysis on three independent experiments showed that 89%±2% of the ROP1 evacuoles observed without parasites were RONs-positive while only 8 . 5%±2% of the ROP1 evacuoles observed without parasites were positive using the anti-AMA1 ectodomain B3 . 90 mAb [9] ( Figure 5A ) , strongly suggesting that the RONs and AMA1 associate with different membranes . Second , IFA of HFF cells pulse-invaded for 15 min showed the presence of the PVM marker ROP1 on empty vacuoles ( Figure 5B ) . The same labelling was observed with anti-ROP2 that labelled another associated PVM rhoptry protein ( data not shown ) . Empty vacuoles represented 6%±2% ( mean±SEM of 7 independent experiments ) of the total vacuole numbers when invasion was synchronized using a K+ buffer shift [23] and corresponded mainly to early egress of the parasite . We then checked the presence of the MJ scar on the PVM of these empty vacuoles . The association of the RONs with the PVM was systematically observed by the immunodetection of RON4/5/8 ( but not RON2 ) on empty PVs labelled with the PVM marker ROP1 but devoid of any parasite ( Figure 5B and Figure S6 ) . Since the PV derives from the host cell membrane , this also shows that RONs are associated with the host cell membrane during invasion and probably maintained together as a complex , even after the PVM has pinched off from the host cell membrane . We sought to address the topology of the RONs at the MJ . To this end , we first analyzed by IFA if the characteristic ring-like pattern of the RONs on invading parasites could be detected in the absence of any permeabilization ( which was verified by the absence of labelling of the PVM with anti-ROP2 or anti-ROP1 sera ) . Since we had observed that the use of formaldehyde to fix parasite during invasion could result in partial permeabilization of the host cell membrane ( data not shown ) , we stopped the invasion process on ice instead and performed the IFA on unfixed cells at 4°C . In these conditions , the MJ complex could not be detected unambiguously with any of the anti-RONs sera . The lack of detection of the epitopes by the antibodies could suggest either a lack of accessibility within the junction , or spatial and conformational constraints or , finally , a localization of these epitopes on the cytoplasmic side of the host cell membrane . We thus addressed the possible association of RONs with the cytoplasmic face of the host plasma membrane . To this end , we examined the topology of the RONs at the MJ remnant in fully invaded parasites by differential permeabilization . Note that since RON2 was not detected at this residual junction , this approach did not allow defining the topology of RON2 in the host cell membrane . In streptolysin-O ( SLO ) -treated infected cells , the host cell plasma membrane was selectively permeabilized without affecting the PVM ( Beckers et al . , 1994 ) , allowing the selective detection of exposed cytosolic domains of PVM-associated protein ( Figure 6A ) . These experiments were carried out with the transgenic GRA5-HA strain [24] , where the HA tagged C-terminal end of the PVM marker GRA5 , is exposed to the vacuolar space: hence , the absence of C-terminal labelling of GRA5 was used as control of the integrity of the PVM ( Figure 6A ) . In addition , anti-SAG1 antibodies were used as additional control of the integrity of the PVM and to distinguish intracellular parasites ( SAG1-negative ) from the extracellular ones that remained attached to the cells ( SAG1-positive ) . We first controlled that the RON scar was not detectable in the absence of streptolysin showing that it is inside the cell and not on the surface ( data not shown ) . On samples SLO-permeabilized 15 min after invasion , RON4 , 5 , 8 ( RON2 could not be detected ) were found to be exposed toward the host cell cytoplasm ( Figure 6A ) . This was also confirmed with the detection of RON4 , RON5 and RON8 at the surface of isolated intact parasite-containing vacuoles ( Text S1 , Figure S5 ) . Since these two approaches allowed the detection of RONs when the parasites had fully invaded , we could not exclude that the topology of the RONs had changed during the closure of the MJ . Thus , we decided to assess the topology of these proteins during the course of invasion . To this end , we pre-loaded the host cells with anti-RONs antibodies by mechanical glass beads loading and subsequently infected the cells with Toxoplasma tachyzoites . In these conditions , RONs would only be detected by the pre-loaded antibodies if they were secreted into the host cell cytoplasm . The cells were then fixed during invasion , permeabilized and subjected to fluorescent secondary antibody detection . The results clearly showed the detection of the ring-shaped MJ with anti-RON4 , 5 , 8 ( Figure 6B ) , whereas no signal was detected in the absence of permeabilization ( data not shown ) . When cells were loaded using this technique with anti-RON2 or anti AMA1 antibodies , these proteins could not be detected at the MJ . It is to note that this approach could not be used to assess the inhibitory effect of the anti-RONs on the invasion , as the amount of antibody loaded in the cells is variable and cannot be quantified . Taken together , our results show that the parasites can secrete the RONs complex directly into the host cell cytoplasm , RON4 , RON5 and RON8 remaining associated with the cytoplasmic side of the plasma membrane/PVM during invasion , after which they persist there for a few hours as a residual structure .
The export of parasite material to the host membrane described here , particularly facing the cytoplasm of the host cell , is perfectly compatible with the thickening of the inner leaflet of the host cell membrane bilayer that has been observed at the MJ by electron microscopy [2] . It is unclear how RONs are exported into the host cell and how they could insert into , or bind to , the host cell membrane , but a secretion of rhoptry material through a transient pore in the host cell membrane has been proposed [4] . Proteins from the bulb of the rhoptry ( ROPs ) are known to be secreted in association with vesicles ( e-vacuoles ) into the host cell cytoplasm [16] , but this is likely to occur after junction formation and RONs are not found in e-vacuoles . RONs must therefore be translocated at a very early stage , likely corresponding to the transient spike of conductance detected by patch clamp study of T . gondii invasion [25] . An association with lipids might also facilitate membrane insertion . RON4 and RON8 , which are not predicted to possess a TM domain , are exposed to the cytoplasmic face of the host cell . RON5 , which contains only one putative TM domain in its N-terminal end is , at least partially , exposed on this cytoplasmic face . Although we showed unambiguously that RON2 ( which possesses three putative TM domains ) is also a component of the MJ associated with the host cell , its precise topology at the membrane will need further studies . Indeed , whereas RON4/5/8 proteins are easily detected at the MJ , RON2 is only observed at a very early stage of junction formation or when the actin cytoskeleton is destabilized with Cyt-D , using a serum directed against a very short sequence located between the last two TM domains . This may reflect a direct or indirect interaction of this domain with the sub-plasmalemmal cytoskeleton , although invasion has been shown to critically depend on actin filaments of the parasite but not of the host cell [26] . Interestingly , only the last two TM domains of RON2 are unambiguously recognised by the prediction programs we used and the loop between these TM domains is particularly well conserved between all Apicomplexa RON2 orthologs , suggesting a conserved function . T . gondii develops within a vacuole that derives from the host cell membrane . A fascinating phenomenon in Apicomplexa invasion is the selective restricted access of host proteins to the forming vacuole in which the parasite develops . This molecular sieving takes place at the MJ [2] , [27] . Indeed , the presence of RONs at the cytoplasmic face of the host cell could also be involved in the exclusion of host plasma membrane proteins from the PV membrane; they would constitute a selective and protective physical barrier that would prevent protein candidates , which could mediate the fusion of the PV with the endo-lysosomal system , to be incorporated and therefore creating a non-fusigenic compartment in which the parasite could develop . These results call for functional studies to assess the respective roles of the RONs in mediating a successful invasion . A firm attachment of the parasite to the host cell membrane is necessary to propel it inside the PV . To do so , several scenarios might be envisaged . First , parasite ligands might be binding to host cell receptors . Another possibility is that parasite proteins , which would be directly inserted into the host cell membrane , could serve as receptors for the parasite through their extracellular domain . Our data fit perfectly with the latter scenario . Indeed , the parasite targets proteins on both membranes of the MJ; on the parasite surface for AMA1 via its TM domain [10] and on the host cell membrane counterpart , for RON2/4/5/8 ( this study ) . In addition , we have shown that AMA1 can interact directly with RON2 in vitro . The domains interacting together still remain to be mapped . The precise function of AMA1 is not yet known , but several previous works showed a role of the protein in establishing close contact with the host cell suggesting that AMA1 could be involved in a receptor-ligand interaction [7] , [28] . However direct binding of Plasmodium or T . gondii AMA1 to the target cell has not been proven unambiguously . Here we propose that the interaction of AMA1 with the host cell could be mediated by a RON2 receptor inserted into the host cell membrane ( Figure 7 ) . This model may account for the conserved mechanism of invasion by Apicomplexa . This type of secretion by a pathogen of a receptor for its own invasion machinery is reminiscent of the translocated intimin receptor ( TIR ) exported by enteropathogenic Escherichia coli [29] , but it would be the first one to be characterised for a eukaryotic pathogen . One crucial question is which protein is dragging the MJ backward during invasion ? AMA1 does not possess the critical tryptophan that is necessary for interaction of its C-terminus with the sub-membranous motor [30] . In addition , during invasion , AMA-1 is present at the MJ but the majority of AMA1 is clearly on the parasite surface , on both sides of this adhesion zone [5] , [19] , implying that at least part of the AMA1 pool is not translocated posteriorly as opposed to other microneme proteins . Consistent with this notion , it is possible that part of the AMA1 pool serves as the anchor for the junction , but that another microneme TM protein interacts with the complex once assembled to propel it backwards in a glideosome-dependent motion . Apicomplexan parasites show a wide range of host cell specificities that may depend on the expression of the MIC repertoire that differs greatly between parasites or stages of the same parasite; we hypothesize that the conserved process of invasion itself ( i . e . MJ-dependent host cell entry ) would be rather mediated by the specific protein complex described here , which is mostly conserved among Apicomplexa . However in this study we have characterized a new member of the MJ complex , RON8 , which , in contrast to other MJ members , is specific to T . gondii and N . caninum . This highlights the fact that the MJ complex could have a different composition in several Apicomplexa and would suggest that some MJ partners could also account for driving the specificity to the host cell type . The invasion by apicomplexan parasites is a well-orchestrated mechanism involving the targeting of interacting proteins from two distinct compartments . One intriguing question is how the micronemal protein AMA1 and the RONs complex , which move through a conventional eukaryotic secretory pathway involving the rough ER , the Golgi apparatus and endosome-like structures , avoid interacting before secretion . Here we showed that even if they could physically interact as pro-proteins in cell extracts , it is probably not the case in these intermediate compartments because of a distinct timing of biosynthesis between MICs and RONs . Indeed , all the RONs ( in addition to the ROPs , data not shown ) are present at the same time in the pre-rhoptries in dividing parasites , whereas newly-synthesized MIC3 and M2AP ( their immature forms ) are not yet detected in these parasites . They are synthesized later , when the rhoptry compartment has been fully loaded . This is the first study showing unambiguously that MICs and RONs are not expressed at the same time , which indicates that the biogenesis of rhoptries and micronemes is asynchronous in T . gondii , as previously suggested by ultrastructural analysis of other Apicomplexa such as P . berghei [31] . This distinct timing of biogenesis for proteins destined to two secretory organelles could be a general mechanism of segregation used by the parasite for interacting proteins , which would allow interaction only after secretion and during invasion . In summary , this study extends significantly our understanding of the MJ formation and composition . The finely-tuned rhoptry and micronemal protein biosynthesis , the cooperation of these proteins originating from two different secretory organelles and the secretion of MJ components directly into the host cell , highlight the sophisticated strategies driving the active invasion of the Apicomplexa .
All T . gondii tachyzoites were grown in human foreskin fibroblasts ( HFF ) or Vero cells grown in standard condition . Tachyzoites of the RH hxgprt- strain of T . gondii deleted for hypoxanthine guanine phosphoribosyl transferase ( ΔHX strain ) [32] and GRA5-HA [24] were used throughout the study . Total RNA was isolated using RNAqueous ( Ambion ) , according to the manufacturer's instructions . cDNA was synthetized from RH hxgprt- parasites using random hexamers and SuperScript II ( Invitrogen ) or using the SMART RACE cDNA Amplification Kit ( Clontech Laboratories , Inc ) . cDNA fragments of TwinScan_2001 were amplified using a set of primers ML208/ML162 , ML209/ML165 , ML211/ML210 , ML212/ML213 and ML214/ML215 , and cloned into the pCR-Blunt II-TOPO vector or into the pCR-2 . 1-TOPO vector ( Invitrogen ) . After sequencing , the complete open reading frame of RON8 was reconstituted from the overlapping cDNA sequences . For IFA of intracellular parasites , confluent HFF monolayers were infected with RH tachyzoites for 24 h , then fixed for 30 min in 4% paraformaldehyde ( PAF ) in PBS . For methanol fixation , monolayers were immersed in methanol for 6 min at −20°C before IFA . After three washes in PBS , cells were permeabilized with 0 . 2% Triton X-100 in PBS for 10 min , blocked with 10% fetal bovine serum in PBS ( PFBS ) for 30 min . The cells were stained with primary antibody diluted in 10% PFBS for one hour , washed and then incubated with secondary antibody coupled to Alexa 594 or Alexa 488 ( Sigma ) . IFA of invading parasites were obtained by synchronisation of invasion at 4°C [6] or using a K+ buffer shift [23] . Invasion was carried out for 2 min30 and was stopped and fixed by adding an excess volume of 4% PAF in PBS . After three washes in PBS , cells were permeabilized with 0 . 05% saponin ( w/v , Sigma ) in PBS for 10 min , then IFA was performed as described above . Alternatively , invasion was stopped on ice , and live cells were incubated for 1 h on ice with primary antibodies , before fixation in 4% PAF , saponin permeabilization and incubation with conjugates . When needed , invasion was blocked with Cyt-D treatment by incubation of extracellular parasites with 1 µM of drug for 20 min at 37°C before invasion and then incubation of parasites for 20 min at 37°C in the presence of the drug . SLO permeabilization was conducted as described previously [33] . The coverslips were mounted onto microscope slides using Immunomount ( Calbiochem ) . Observations were performed on a Leica DMRA2 microscope equipped for epifluorescence and images were recorded with a COOLSNAP CCD camera ( Photometrics , Tucson , AZ ) driven by Metaview ( Universal Imaging Co . , Downington , PA ) . Image acquisition was performed on workstations of Montpellier RIO imaging facility . Loading of antibodies into the host cell was done as described previously [34] . Acid-washed 150–212 µm glass beads ( Sigma ) were washed 3 times with distilled water . 0 . 1 mg of beads were then resuspended in 300 µl of the appropriate medium containing the antibody of interest ( ie . hybridoma culture supernatant , or antiserum diluted 1/30 ) . HFF cultures growing on coverslips in a 24 wells-plate were washed twice with Hanks' Balanced Salt Solution ( HBSS ) before the antibodies-beads solution was put into each well . The beads were rolled onto the coverslip by tilting the plate ∼10 times , until evenly distributed over its surface . The coverslip was then transferred to another well where it was washed 3 times with HBSS and returned to DMEM culture medium and allowed to recover at 37°C and 5% CO2 for 30 minutes . Invasion assays were then carried out by allowing T . gondii tachyzoites to sediment on the HFF for 20 minutes at 4°C and subsequently warming them during 2–5 min at 37°C to trigger invasion . Invasion was stopped and cells were fixed by adding an excess volume of 4% PAF in PBS . The extracellular portion of the tachyzoites was labelled with mAb T3 1E5 specific for the surface protein SAG1 . Parasites and cells were then permeabilized with saponin and incubated with anti-RONs or anti-AMA1antibodies . Heavily infected HFF monolayers were incubated in methionine and cysteine-free DMEM ( Invitrogen ) containing 4% dialyzed FCS for 30 min at 37°C in a 5% CO2 incubator prior to the addition of 50 µCi/ml [35S]-methionine/[35S]-cysteine ( 700 Ci/mM; Perkin Elmer ) with or without BFA ( 5 µg/ml ) . The infected monolayers were then labelled for 15 or 20 min , rinsed with complete DMEM containing 10% FCS , and either processed or incubated in this medium complemented or not with BFA ( 5 µg/ml ) for 1 h chase prior to IP . Parasite solubilization in 1% NP40 or in 0 . 6% SDS and immunosoption procedures were done as described previously [6] , [35] . Elution was performed during 5 min at 95°C with electrophoresis sample buffer . After SDS-PAGE , the gel was impregnated with Amplify ( Amersham ) , dried , and exposed to Biomax film ( Kodak ) at −80°C . Individual bands from Coomassie stained SDS-PAGE gels were excised , treated with trypsin , and extracted for analysis by nanoflow HPLC-nano-electrospray ionization on a Bruker Esquire 3000+ ion trap mass spectrometer coupled with a LC-Packings HPLC as described previously [6] . | A unique feature of apicomplexan parasites is the formation of an intimate contact between the apex of the parasite and the host cell membrane called the moving junction that moves along the parasite during invasion . Proteins originated from two distinct secretory organelles , the microneme for AMA1 and the rhoptry neck for RON2/4/5 proteins , are associated to form the junction . Here , we have furthered the characterization of the MJ complex by describing RON8 , an additional protein component . AMA1 has previously been shown to be inserted into the parasite membrane upon secretion . Our study demonstrates that all the RON proteins are translocated into the host cell , where RON4/5/8 remain associated with the cytoplasmic face of the host cell plasma membrane . Furthermore , we identified a privileged interaction between transmembrane MJ proteins AMA1 and RON2 in vitro . Overall , this led us to propose the first model describing the putative MJ organisation at the interface between the host cell and Toxoplasma . In this original concept , the parasite would export its own receptor ( RON2 ) and ligand ( AMA1 ) on either side of the MJ . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"cell",
"biology",
"microbiology/cellular",
"microbiology",
"and",
"pathogenesis",
"microbiology/parasitology"
] | 2009 | Export of a Toxoplasma gondii Rhoptry Neck Protein Complex at the Host Cell Membrane to Form the Moving Junction during Invasion |
Mosquito-borne viruses encompass a range of virus families , comprising a number of significant human pathogens ( e . g . , dengue viruses , West Nile virus , Chikungunya virus ) . Virulent strains of these viruses are continually evolving and expanding their geographic range , thus rapid and sensitive screening assays are required to detect emerging viruses and monitor their prevalence and spread in mosquito populations . Double-stranded RNA ( dsRNA ) is produced during the replication of many of these viruses as either an intermediate in RNA replication ( e . g . , flaviviruses , togaviruses ) or the double-stranded RNA genome ( e . g . , reoviruses ) . Detection and discovery of novel viruses from field and clinical samples usually relies on recognition of antigens or nucleotide sequences conserved within a virus genus or family . However , due to the wide antigenic and genetic variation within and between viral families , many novel or divergent species can be overlooked by these approaches . We have developed two monoclonal antibodies ( mAbs ) which show co-localised staining with proteins involved in viral RNA replication in immunofluorescence assay ( IFA ) , suggesting specific reactivity to viral dsRNA . By assessing binding against a panel of synthetic dsRNA molecules , we have shown that these mAbs recognise dsRNA greater than 30 base pairs in length in a sequence-independent manner . IFA and enzyme-linked immunosorbent assay ( ELISA ) were employed to demonstrate detection of a panel of RNA viruses from several families , in a range of cell types . These mAbs , termed monoclonal antibodies to viral RNA intermediates in cells ( MAVRIC ) , have now been incorporated into a high-throughput , economical ELISA-based screening system for the detection and discovery of viruses from mosquito populations . Our results have demonstrated that this simple system enables the efficient detection and isolation of a range of known and novel viruses in cells inoculated with field-caught mosquito samples , and represents a rapid , sequence-independent , and cost-effective approach to virus discovery .
Arthropod-borne viruses ( arboviruses ) encompass a range of veterinary and medically significant viral pathogens belonging to five antigenically distinct families of RNA viruses . These families can be separated according to their genome type: those with positive-sense single-stranded RNA ( ( + ) ssRNA ) genomes , the Togaviridae , Flaviviridae; negative-sense RNA ( ( - ) ssRNA ) genomes , Rhabdoviridae and Bunyaviridae; and the double-stranded RNA ( dsRNA ) viruses of the Reoviridae family . These viruses cycle between haematophagous arthropod vectors and reservoir/amplifying vertebrate hosts . Occasionally humans and livestock can become incidental hosts for these viruses and may develop encephalitic or haemorrhagic disease . New and more virulent strains of these viruses are continually emerging and expanding their geographic range [1 , 2] . As a result many arthropod populations are routinely surveyed in an attempt to assess the risk of arboviruses and identify emerging pathogens . The co-circulation of insect-specific viruses , such as the divergent insect-specific flaviviruses ( ISFs ) , adds another layer of complexity to the spread and distribution of arboviruses in mosquito populations [3] . While not of direct affect to the health of humans and animals , our lab and others have shown that ISFs circulating in mosquito populations may suppress or enhance the replication of pathogenic arboviruses such as the encephalitogenic West Nile virus ( WNV ) [4–6] . Surveillance for arboviruses and detection of new mosquito-borne viruses currently relies on antigenic , molecular or deep sequencing based approaches [7–11] . However , these methods are often expensive or limited by genus-specificity and divergent viruses such as ISFs are often missed . We have developed a novel assay system based on two unique monoclonal antibodies ( mAbs ) that recognise an antigen in cells infected with a wide range of viruses . This system provides a streamlined and economical approach for virus detection and discovery . Here we characterise the antigen recognised by these novel mAbs and show that this system provides a streamlined method for detecting infection with viruses from at least three of five conventional arboviral families as well as a new species in the novel family Mesoniviridae .
Hybridomas secreting mAbs 3G1 and 2G4 were generated by immunising an adult female BALB/c mouse with concentrated supernatant from C6/36 cell cultures infected with Palm Creek virus as previously described [6] . Antibodies were isotyped using Mouse Typer Isotyping kit ( BioRad ) as per the manufacturer’s instructions and then confirmed by electron microscopy . Monoclonal antibodies 2G4 and isotype control 3 . 1112G [12] were purified using GE IgM purification HP column ( 1 ml ) ( 17–5110–01 ) as per manufacturer’s instructions with 1 M ammonium sulfate using AKTA FPLC system . Antibody 3G1 was unable to be purified by this method . In order to obtain a crudely purified stock of this mAb , 3G1 hybridoma cell culture supernatant was concentrated using a high molecular weight ( 300K ) MWCO Spin-X UF concentrator column ( Corning ) . Preparations of purified 2G4 and concentrated 3G1 were diluted to a final concentration of 5 μg/ml in PBS . 4 μl samples of the antibody preparations was pipetted onto glow-discharged carbon-formvar films supported on 400-mesh copper-grids and negative staining was performed using 1% uranyl acetate . Samples were imaged at 59000x magnification using a Tecnai F30 FEG-TEM ( FEI ) operating at 300 kV with a 4k lens-coupled camera ( Direct Electron ) . C6/36 ( Aedes albopictus ) cells [13 , 14] were maintained in RPMI-1640 supplemented with 5% fetal bovine serum ( FBS ) and grown at 28°C . Vero cells ( African Green Monkey Kidney ) [15] were maintained in Dulbecco’s modified Eagle’s medium ( DMEM ) containing 2–5% FBS , while DF-1 cells ( chicken embryo fibroblast ) [16] were grown in DMEM with 5% FBS . Hybridoma cell lines used in this study were grown in hybridoma serum free medium ( Invitrogen ) initially supplemented with 20% FBS and then weaned to serum-free culture for antibody production . All vertebrate cells were incubated at 37°C with 5% CO2 . All media were supplemented with 50 μg/ml streptomycin , 50 U/mL penicillin and 2 mM L-glutamine . The viruses used in this study were West Nile virus subtype Kunjin strain MRM61C ( WNVKUNV ) , West Nile virus strain New York 99 ( WNVNY99 ) , dengue virus serotype 1 ( DENV-1 ) , dengue virus serotype 2 strain New Guinea C ( DENV-2 NGC ) , Yellow Fever virus strain 17D ( YFV 17D ) , Murray Valley encephalitis virus strain 151 ( MVE 151 ) , Palm Creek virus 56 ( PCV ) , Ross River virus T48 ( RRV ) , Barmah Forrest virus ( BFV ) , Akabane virus strain A661 ( AKAV; obtained from Peter Kirkland , EMAI ) , bovine ephemeral fever virus strain F704 ( BEFV; from Peter Kirkland , EMAI ) , Blue Tongue virus isolate D870 ( BTV; obtained from Peter Kirkland , EMAI ) . All virus stocks were prepared from the cell culture supernatant of virus-infected Aedes albopictus ( C6/36 ) or baby hamster kidney ( BHK ) for those that did not grow in C6/36 cells ( AKAV , BEFV , BTV ) as per previously described protocol [17] . The virus titre was determined as 50% tissue culture infective dose ( TCID50 ) by fixed-cell ELISA using virus specific mAbs [17] . C6/36 cells were seeded in 96-well plates and infected the following day with WNVKUNV at a multiplicity of infection ( MOI ) of 1 or mock infected . At day 8 post seeding , a subset of mock infected cells were incubated at 41°C for 3 hours to induce heat shock . Immediately post heat shock treatment all cells were fixed with acetone fixative buffer ( 20% acetone with 0 . 02% bovine serum albumin ( BSA ) in phosphate buffered saline ( PBS ) ) . Control mock-infected and WNV-infected cells were maintained at 28°C for the duration of the study . Fixed-cell ELISA was performed as per previously published methods with slight modifications [18] . Briefly , infected or mock-infected cell monolayers were fixed with acetone fixative buffer . Plates were blocked with 150 μl per well ELISA blocking buffer ( 0 . 05 M Tris/HCl ( pH 8 . 0 ) , 1 mM EDTA , 0 . 15 M NaCl , 0 . 05% ( v/v ) Tween 20 , 0 . 2% w/v casein ) for 1 hour at room temperature before probing with 50 μl/well respective mAbs diluted in blocking buffer as needed . Plates were incubated with primary antibody for 1 hour at 37°C before washing 4x with PBS containing 0 . 05% tween-20 ( PBST ) . Secondary antibody goat anti-mouse HRP ( DAKO ) was diluted 1/2000 in blocking buffer and added at 50 μl/well . After another incubation at 37°C for 1 hour , plates were washed 6x with PBST . Finally , 100 μl/well substrate solution [1 mM 2 , 2-azino-bis ( 3-ethylbenzthiazoline-6-sulfonic acid ) ( ABTS ) , 3 mM H2O2 in a buffer prepared by mixing 0 . 1 M citric acid with 0 . 2 M Na2HPO4 to give a pH of 4 . 2] was added per well and plates were incubated in the dark at room temperature for 1 hour . Absorbance was measured at 405 nm . All cells were grown on glass coverslips , mammalian cells were seeded at 98% confluency one day prior to infection . C6/36 cells were seeded at 25% confluency to allow for longer incubation times . Monolayers were infected or transfected as indicated and fixed with ice cold 100% acetone for 5 minutes and air dried before storing at -20°C . Prior to staining , coverslips were blocked with 1% BSA/PBS for one hour at room temperature . For single antibody staining , coverslips were incubated for 1 hour at 37°C with 3G1 and 2G4 hybridoma cell culture fluid or virus specific mAbs as follows: mAb 3 . 1112G , anti-non-structural ( NS ) protein 1 ( reactive with WNV , IgM ) hybridoma culture fluid; pan-flavivirus mAb 4G2 anti-envelope ( E ) ( IgG ) [19] hybridoma culture fluid; mAb r847 anti-BTV hybridoma culture fluid ( obtained from Peter Kirkland , EMAI ) ; mAb J2 , anti-dsRNA used at 1:200 ( IgG , English and Scientific Consulting , Hungary ) [20] . Coverslips were washed 3x with PBS and stained with Alexafluor 488-conjugated goat anti-mouse IgG ( H+L ) ( Invitrogen ) diluted 1:500 in 1%BSA/PBS and Hoechst 33342 nuclear stain ( Invitrogen ) for 1 hour or 5 minutes at room temperature respectively . Following another 3 washes with PBS , the coverslips were mounted onto glass microscope slides using ProLong Gold Anti-fade ( Invitrogen ) . For dual staining , coverslips were first incubated for 1hr RT with virus specific antibodies as follows: anti-non-structural ( NS ) protein 3 [21] ( diluted 1:100 , polyclonal rabbit sera , flavivirus-reactive ) ; mAb 5H1 anti-NS5 antibody hybridoma cell culture supernatant [22] ( IgG , strong reaction with WNVKUN strains , weak for WNVNY99 , non-reactive with WNV strains from other lineages ) ; mAb 4G4 anti-NS1 [18] ( IgG , pan-flavivirus ) ; mAb G8 anti-RRV E2 antibody hybridoma cell culture supernatant [23]; mAb r5412 anti-AKAV hybridoma cell culture supernatant ( obtained from Peter Kirkland , EMAI ) ; mAb r1459 anti-BEFV hybridoma cell culture supernatant ( obtained from Peter Kirkland , EMAI ) . Coverslips were then washed 3x with PBS and stained with Alexafluor 488-conjugated goat anti-mouse IgG ( H+L ) , or 3x with PBST and stained with Texas Red-conjugated goat anti-rabbit IgG ( H+L ) ( Invitrogen ) for anti-NS3 sera for 1 hour RT . Coverslips were again washed 3x with PBS and incubated with 3G1 or 2G4 hybridoma culture supernatant for 1 hour 37°C . Finally , coverslips were washed 3x with PBS and stained with 594-conjugated goat anti-mouse IgM ( μ chain ) or 488-conjugated goat anti-mouse IgM ( μ chain ) ( for monolayers stained with rabbit sera ) ( Invitrogen ) for 1 hour at 37°C followed by Hoechst 33342 nuclear stain ( Invitrogen ) for 5 minutes before washing and mounting with ProLong Gold Anti-fade . All coverslips were viewed under the ZEISS LSM 510 META confocal microscope . For isolation of viral RNA from supernatant of inoculated cells , RNA extraction was performed using Machery-Nagel viral RNA isolation kit , while Tri Reagent ( Sigma ) was used to harvest lysates of virus-infected C6/36 cells in 24 well plates and RNA was extracted according to the manufacturer’s instructions . RT-PCR for identification and confirmation of known viruses was performed using virus-specific primers ( Flavivirus ( FU2: GCTGATGACACCGCCGGCTGGGACAC , cFD3: AGCATGTCTTCCGTGGTCATCCA ) ; Alphavirus ( F1: TTTAAGTTTGGTGCGATGATGAAGTC , R: GCATCTATGATATTGACTTCCATGTT ) ; Sindbis virus ( SINVF7: GCCAGAGTGTGTCTCAAGCA , SINVR7: TCGATTCGTTCCTTCCACTT ) ) in the SuperScript III One-Step RT-PCR System with Platinum Taq DNA polymerase ( Invitrogen ) [24–26] . Quantitative RT-PCR to assess RNA extracted from DENV-2 and WNVKUNV-infected C6/36 lysates was performed using Taqman RT-PCR assays as described by Warrilow et al ( 2002 ) and van den Hurk et al ( 2014 ) [27 , 28] . Screening was performed by inoculating filtered mosquito homogenate onto monolayers of C6/36 cells in 96-well microplates and incubating at 28°C for 5–7 days . Cells were monitored over 7 days for cytopathic effect ( CPE ) . At 7 days post-infection culture supernatant ( 200 μL ) was collected and stored at −80°C for further analysis and cells were then fixed using acetone fixative buffer . ELISA was then performed on fixed cells as specified previously using a cocktail of mAbs 3G1 and 2G4 . Anti-poly ( I:C ) ELISA was performed as per Schonborn et al ( 1991 ) [20] . Briefly , a 96-well plate was coated with 150 μl/well 2% w/v protamine sulfate ( Sigma-Aldrich ) in PBS and incubated at 37°C for 2 hours . Plates were then incubated with 100 μl/well ( 0 . 25–200 ng ) Poly ( I:C ) diluted in TE Buffer ( 100 mM Tris , 1 mM EDTA , pH 8 ) for a further 2 hours at 37°C . ELISA was performed as previously described using 3G1 and 2G4 hybridoma culture fluid and purified anti-dsRNA antibody J2 ( diluted at 1:500 in ELISA blocking buffer , English and Scientific Consulting , Hungary ) . Capture ELISA was performed by coating purified mAbs at a pre-determined optimal range ( 0 . 5–1 μg ) in PBS and incubating overnight 4°C . Small biotinylated dsRNAs ( 30 , 40 and 50 bp designed to KUNV NS1 , IDT ) were diluted ( 0 . 25–50 pmol ) in TE buffer , added 50 μl/well and incubated 2 hours 37°C . Streptavidin conjugated with horse radish peroxidase ( HRP ) was diluted 1/2000 in PBST and added 50 μl/well . Cells were seeded to 98% confluency on coverslips in 24-well plates . The following day , cells were washed 3x with sterile PBS and topped up with media without Pen/Strep . 2 μl lipofectamine 2000 ( Invitrogen ) and 2 . 5 μg Poly ( I:C ) were added to each well with OptiMEM media . Transfections were incubated for 6 hours and then fixed with ice cold 100% acetone . Shortcut RNAse III ( New England BioLabs ) digestion was performed as per manufacturer’s suggestion . Briefly , 150 μl enzyme preparation ( 2U/100 μl ) or buffer without enzyme ( mock digestion ) was added to acetone-fixed cell monolayers on coverslips and incubated at 37°C for 2 hours before removing and staining for IFA as above . C6/36 cells at 80–90% confluency were infected with WNVKUNV , RRV or mock-infected using serial dilutions of virus ( 10-1 to 10-12 ) . Infected cells were fixed at 24 and 48 hours post infection and 5 days post-infection using acetone fixative buffer . Fixed cell ELISA was performed using mAbs 3G1 and 2G4 and pan-flavivirus anti-E protein mAb 4G2 or mAb G8 ( anti-RRV E2 ) . ELISA using each antibody was performed on triplicate plates and the titre of virus detected by each antibody was determined using the Reed-Meunch calculation .
Monoclonal antibodies ( mAbs ) 3G1 and 2G4 were raised while generating a panel of antibodies to the ISF Palm Creek virus ( PCV ) [6] . Electron microscopy analysis revealed that the antibodies formed cyclic complexes approximately 30 nm in diameter with five radially protruding arms consistent with the features of pentameric IgM antibodies ( Fig . 1 ) [29 , 30] . Our analysis showed that these antibodies reacted strongly with PCV-infected Aedes albopictus ( C6/36 ) cells in fixed-cell ELISA , as well as against C6/36 cells infected with a number of pathogenic flaviviruses compared to mock-infected cells ( Table 1 ) . Further analysis showed mAbs 3G1 and 2G4 also strongly reacted in fixed-cell ELISA against C6/36 cells infected with the alphaviruses Ross River virus ( RRV ) and Barmah Forrest virus ( BFV ) ( Table 1 ) . The ability of these antibodies to detect infection with viruses from two antigenically distinct families suggested that the antibody target was unlikely to be of specific viral origin and may be a factor produced as the result of infection with RNA viruses . In order to assess whether mAbs 3G1 and 2G4 were recognising a general stress-response protein , their reactivity against heat-shocked C6/36 cells was assessed . In fixed-cell ELISA , mAbs 3G1 and 2G4 reacted strongly against WNVKUNV-infected cells , but did not show a significant increase in reactivity against heat-shocked cells ( Fig . 2 ) . Anti-heat shock protein 70 ( HSP70 ) mAb showed significant reactivity against both heat-shocked and virus-infected cells ( Fig . 2 ) , indicating that HSP70 up-regulation was induced in these cells by both heat-shock treatment and viral infection as previously shown [31 , 32] . Given the wide range of viruses detected by 3G1 and 2G4 in mosquito cells and the ability of these viruses to infect other cell types , we decided to assess whether these mAbs would react against WNVKUNV-infected mammalian ( Vero ) and avian ( DF-1 ) cells in IFA ( Fig . 3 ) . Staining in infected DF-1 and Vero cells by both mAbs appeared in the perinuclear region . At 48 hours post infection with WNVKUNV , strong punctate circles were obvious in the perinuclear region of Vero cells stained with 3G1 and 2G4 ( Fig . 3 ) . Similar patterns were also observed in cells stained with isotype control antibody 3 . 1112G which is specific to WNV non-structural protein 1 ( anti-NS1 ) . This circular perinuclear staining observed for both 3G1 and 2G4 in WNVKUNV-infected Veros is consistent with staining for flavivirus RNA replication complexes formed inside vesicle packets in the endoplasmic reticulum of infected cells [21 , 33] , suggesting that the antigen was associated with viral RNA replication . Both antibodies exhibited some nuclear staining in uninfected Vero cells , which appeared to be associated with the nucleolus ( Fig . 3 ) . To further investigate the involvement of the mAb-reactive antigen with flavivirus replication , an antibody-binding assay was performed to look at the timing of its induction during infection . This assay revealed that binding of both mAbs 3G1 and 2G4 to WNVKUNV-infected C6/36 cells peaked at 4 days post-infection ( Fig . 4 ) . This binding pattern mirrored the detection of viral NS1 protein which peaked at 3 days post-infection and envelope ( E ) protein which peaked at 4 days post-infection ( Fig . 4 ) . We then performed dual staining on WNVKUNV-infected Vero cells using antibodies specific to flavivirus proteins known to be involved in replication of viral RNA ( anti-NS1 antibody 4G4 [18] , anti-NS5 antibody 5H1 [34] and anti-NS3 rabbit sera [21] ) . Dual staining revealed that the antigen recognised by 3G1 and 2G4 co-localised with all three proteins at these perinuclear regions in infected cells consistent with a component of the virus replication complex ( Fig . 5 ) . Whilst cross-reactivity of these mAbs with divergent viruses suggested that their target antigen was unlikely to be a specific viral protein , the data thus far indicated it was specifically associated with viral infection . The co-localisation of mAb 3G1 and 2G4 staining with proteins of the flavivirus RNA replication complex suggested that the double-stranded replicative intermediate of flaviviral RNA may be a potential candidate for the target antigen . To investigate this we treated acetone-fixed cells with the double-stranded RNA ( dsRNA ) -specific nuclease RNAse III . This treatment abolished binding for both mAbs 3G1 and 2G4 as well as binding by positive control commercial antibody J2 ( Fig . 6 ) that is specific for dsRNA [20 , 35] . Binding of NS1-specific antibody 3 . 1112G was not affected by this treatment ( Fig . 6 ) . The results thus implicated dsRNA as the primary target for 3G1 and 2G4 antibodies To investigate whether mAbs 3G1 and 2G4 bound dsRNA in cells in a non-sequence specific manner , we tested both antibodies in IFA against acetone-fixed cells which had been transfected with Poly ( I:C ) [35] . Both 3G1 and 2G4 showed cytoplasmic binding in transfected cells , with only some nuclear staining occurring in mock-transfected cells ( Fig . 7 ) . Staining was again abolished when cells were treated with RNAse III . To further confirm that the mAbs were binding specifically to Poly ( I:C ) , and not to a Poly ( I:C ) -induced cellular component , we assessed antibody reactivity in vitro using Poly ( I:C ) -coated 96-well plates in ELISA [20] . Again , both 3G1 and 2G4 showed strong reactivity to Poly ( I:C ) in this format ( Fig . 8 ) . The inability of mAbs 3G1 and 2G4 to bind dsRNA following RNAse III digestion suggested that these mAbs require the dsRNA target to be greater than 18–25 base pairs ( bp ) , since this is the length of molecule produced by RNAse III digestion [36] . In order to further characterise the minimal size of dsRNA recognised by mAbs 3G1 and 2G4 , we designed a number of short biotinylated dsRNA molecules ( Integrated DNA Technologies ) . Antibody reactivity was assessed in ELISA by capture of small dsRNA molecules by antibodies coated to plates in solid phase . Purified 2G4 captured both 50 bp and 40 bp molecules strongly with reactivity saturated at 62 . 5 pmol/ml . Optical density readings showed lower levels of capture for the 30 bp molecule ( Fig . 9A ) . Concentrated antibody 3G1 captured the 50 bp molecule strongly , but appeared less effective for smaller molecules ( Fig . 9A ) . Capture ELISA was also performed on plates coated with a negative isotype control antibody 3 . 1112G to rule out non-specific binding ( S1 Fig ) . To confirm the specificity of these mAbs for dsRNA , we then tested against a panel of 50 nt long nucleic acid oligos in capture ELISA ( dsDNA , ssDNA , ssRNA , RNA:DNA hybrid ) . Both 3G1 and 2G4 reacted to the dsRNA molecules but did not show significant reactivity to single-stranded RNA ( ssRNA ) , double-stranded or single-stranded DNA ( dsDNA or ssDNA ) or a double-stranded RNA:DNA hybrid ( RNA:DNA ) ( Fig . 9B ) . Together these data suggested that mAbs 3G1 and 2G4 bound specifically to dsRNA . The different binding profiles in capture ELISA suggest that the mAbs may have different affinities for small dsRNA molecules , with 2G4 able to bind equally well to 40 and 50 bp molecules whilst 3G1 preferentially bound longer ( 50 bp ) molecules . To assess the suitability of 3G1 and 2G4 antibodies to detection of a wide range of virus infections , we performed IFA staining of cells infected with representatives of six arboviral families . Dual staining with each mAb 3G1 and 2G4 and antibodies specific for corresponding viral antigens was employed to examine cells infected with the alphavirus Ross River virus ( RRV ) , flavivirus WNVKUNV , and the two negative-sense RNA viruses bovine ephemeral fever virus ( BEFV ) and Akabane virus ( AKAV ) . The positive-sense RNA viruses RRV and WNVKUNV , were both detected in IFA by 3G1 and 2G4 ( Fig . 10A ) with staining localised in the perinuclear region of infected cells for the flavivirus , while staining in alphavirus-infected cells appeared throughout the cytoplasm consistent with the expected location of replication complexes for the corresponding viruses . Both negative-sense RNA viruses AKAV and BEFV could not be detected with mAbs 3G1 and 2G4 in IFA ( Fig . 10B ) . Staining in IFA for the , mesonivirus Casuarina virus ( CASV ) , a novel positive-sense RNA virus , displayed a similar perinuclear staining pattern to that observed for flaviviruses ( Fig . 11A ) [11] . The dsRNA-encoding reovirus bluetongue virus ( BTV ) was detected in IFA by mAbs 3G1 and 2G4 , with staining observed in the perinuclear region of infected cells and coinciding with the locality of the anti-BTV antibody r847 ( Fig . 11B ) . The speckled pattern observed for the nuclear stain and destruction of the monolayer in infected cells was consistent with the induction of apoptosis in mammalian cells by BTV [37 , 38] . The results demonstrate that anti-dsRNA antibodies 3G1 and 2G4 can be used to detect infection with a wide range of positive-sense single-stranded and double-stranded RNA viruses but not with negative-sense RNA viruses Interestingly , in ELISA , mAbs 3G1 and 2G4 showed low reactivity to DENV-infected cells in comparison to WNVKUNV-infected cells ( Table 1 , S2A Fig ) . While WNVKUNV-infected cells showed a marked increase in 3G1 and 2G4 binding in comparison to the mock-infected cells at days 3 , 4 and 5 post-infection , there was substantially lower reactivity of the mAbs to cells infected with DENV-2 . To further investigate this discrepancy , IFA was performed on cells inoculated with DENV-2 and WNVKUNV under the same conditions as the ELISA , apart from the fixation process which was performed with 100% acetone ( c . f . 20% acetone for ELISA ) . In IFA , mAbs 3G1 and 2G4 displayed similar reactivity to both WNVKUNV and DENV-2-infected cells , suggesting the presence of similar levels of dsRNA . ( S2B Fig ) . This was further supported by Taqman RT-PCR , whereby high levels of DENV RNA were detected in infected cells at days 3 , 4 and 5 ( as concluded from the low Ct scores , S2C Fig ) . Together , these data indicate that the low reactivity of mAbs 3G1 and 2G4 to DENV-infected cells using ELISA is not due to insufficient RNA . To investigate the sensitivity of virus-detection by MAVRIC , in comparison to traditional methods using virus specific mAbs , fixed cell ELISA was performed on C6/36 cells infected with virus at various dilutions after 24 hours , 48 hours and 5 days post-infection . There was no significant difference in the detection of WNVKUNV by mAb 3G1 in comparison to the flavivirus E-protein specific antibody 4G2 at 48 hours and 5 days post-infection . However , at 24 hours post-infection , detection of WNVKUNV infection of at least one log higher was achieved by staining for the E-protein with mAb 4G2 in comparison to using 3G1 to detect dsRNA ( Fig . 12A ) . MAb 2G4 showed similar detection ability to the anti-E mAb only at 5 days post-infection ( Fig . 12A ) . In contrast , both antibodies 3G1 and 2G4 showed similar levels of detection of the alphavirus RRV at 48 hours post-infection and 5days post-infection when compared with the RRV E2 protein-specific antibody G8 ( Fig . 12B ) . Due to their potential application to virus detection and discovery , mAbs 3G1 and 2G4 were incorporated in an ELISA-based assay for streamlined screening of homogenates of mosquito pools collected around Australia . To assess this system we tested the reactivity of the mAbs against acetone-fixed C6/36 cells which had been inoculated with 40 mosquito pool homogenates which had previously been tested for vertebrate-infecting arboviruses by ELISA with viral antigen-specific mAbs or RT-PCR with genus specific primers . Using these methods , 32 of 40 samples were previously found to be positive for a known arbovirus ( Table 2 ) . Using mAbs 3G1 and 2G4 in fixed cell ELISA , 35 out of 40 samples were identified as positive for virus infection in at least one of four inoculated wells ( Table 2 ) . RT-PCR was also performed on the supernatant collected from homogenate-inoculated cells prior to fixation using generic primers to detect and confirm virus identification . Of the 35 pools identified as positive by the 3G1 and 2G4 ELISA , four were unable to be confirmed by RT-PCR . Pool 1026 which was previously identified as containing Kokobera virus ( KOKV ) was negative by 3G1 and 2G4 fixed-cell ELISA as well as RT-PCR suggesting that there was no virus replicating in these inoculated cells . Pools 3740 and 4189 , which were negative by fixed-cell ELISA were positive by RT-PCR for WNVKUNV and ALFV respectively , suggesting that these homogenates contained low level of virus below the threshold of detection by fixed-cell ELISA . These pools may become positive in fixed-cell ELISA with further passaging . However , since mAbs 3G1 and 2G4 require the replication of viable virus , it cannot be definitively concluded that viable virus was present in these samples by RT-PCR since detection of residual RNA from the inoculum cannot be ruled out . In addition , of 8 pools which were negative by conventional methods , 6 were identified as positive for virus infection by mAbs 3G1 and 2G4 , observation of CPE for 4 of these pools supported the presence of an unknown virus . These antibodies have also been instrumental in the detection of unknown viruses from mosquito pools . Using the same fixed-cell ELISA based approach , mosquito homogenates which were determined to be negative for known arboviruses by standard methods , including detection of viral antigen with virus-specific mAbs or viral RNA by RT-PCR , were screened using mAbs 3G1 and 2G4 ( Table 3 ) . Based on the reactivity of mAbs 3G1 and 2G4 in fixed-cell ELISA , pools containing a number of viruses were selected for further analysis . Viral RNA sequence was amplified from extracted RNA by RT-PCR using generic virus primers or random primers . Sequence was then obtained by Sanger sequencing or next generation sequencing . Preliminary analysis of these sequences by BLASTX indicated that three of these samples contained previously unidentified viruses ( Table 3 ) . A fourth sample ( pool no . 3 , Table 3 ) was determined by deep sequencing to contain the reovirus Liao Ning virus ( LNV ) which was only recently detected in Australia [10] .
We have produced two new monoclonal antibodies which recognise viral dsRNA present in the cytoplasm of cells infected with most positive-sense RNA and double—stranded RNA viruses . We have also demonstrated the potential of these reagents for the rapid detection and discovery of novel viruses from diverse viral families in biological samples . Double-stranded viral RNA is produced in cells infected with most positive-sense RNA viruses as an intermediate of genomic RNA replication . This process has been well characterised for the flavivirus WNVKUNV which was used as the reference virus in our studies . During flavivirus replication , the genomic RNA is used as a template for synthesis of a negative-sense complementary strand leading to the formation of dsRNA products referred to as the replicative form ( RF ) , from which new positive-sense RNA is generated [39 , 40] . The flavivirus non-structural ( NS ) proteins ( e . g . NS3 , NS2A and NS5 ) form a replication complex ( RC ) with the RF in membranous compartments derived from the endoplasmic reticulum known as vesicle packets ( VPs ) . Within the VPs , the RC and RF is anchored to the membrane via interactions with other membrane-associated non-structural proteins including NS4 and NS1[21] . Using dual staining in IFA we have demonstrated co-localisation of mAb 3G1 and 2G4 staining with several NS proteins involved in the RC ( NS1 , NS3 and NS5 ) , confirming their specific recognition of viral dsRNA produced during flavivirus replication . Strong , specific staining by 3G1 and 2G4 of cells infected with the ( + ) ssRNA viruses CASV ( Mesoniviridae ) , RRV ( Togaviridae ) and the dsRNA virus BTV ( Reoviridae ) , was also consistent with the expected location of viral dsRNA in these cells [7 , 41–43] . In contrast , no dsRNA was detected by these mAbs in the cells infected with the negative-sense RNA ( ( - ) ssRNA ) viruses BEFV ( Rhabdoviridae ) and AKAV ( Bunyaviridae ) . These results are consistent with the findings of others who observed that an anti-dsRNA mAb ( J2 ) did not bind cells infected with the negative-strand RNA viruses influenza A virus ( Orthomyxoviridae ) and LaCrosse virus ( Bunyaviridae ) [35] . This can be explained by the genome replication strategy of these viruses . Both rhabdoviruses and bunyaviruses encapsidate their genomic and anti-genomic RNA by complexing them with viral nucleoprotein during synthesis [44–47] . Presumably this prevents the formation of long stretches of dsRNA by obstructing complementary base pairing of the genomic and anti-genomic RNA . This enables the virus to avoid induction of cellular antiviral responses such as RNAi and interferon signalling , of which dsRNA is a potent activator [48 , 49] . The ability of these antibodies to bind native viral dsRNA as well as the synthetic dsRNA analogue Poly ( I:C ) which contains only two nucleotides ( inosine and cytidine ) , suggests a sequence-independent binding mechanism , similar to that reported for another anti-dsRNA mAb ( J2 ) [20 , 50] . Sequence-independent binding is a common mechanism for dsRNA recognition by dsRNA binding proteins ( dsRBPs ) . Current models for non-sequence specific interaction of these proteins with dsRNA suggests that their dsRNA binding domains ( dsRBDs ) interact with the sugar-phosphate backbone of the RNA molecule at successive minor and major grooves of the double helix [51 , 52] . Thus , the duplex form is suggested to be the major determinant for substrate recognition and differentiation from other nucleic acid duplexes [52 , 53] . DsRNA molecules take on the A-form duplex in which the minor and major grooves of the helix are of similar widths [52] . In comparison , dsDNA is primarily found in the B-form which is characterised by wider major grooves and narrower minor grooves . Finally , RNA-DNA duplexes are believed to take on an intermediate form in which the minor groove of the helix is of comparable size to the A-form but the major groove is significantly larger [52 , 53] . Given these previous insights , the specificity of both 3G1 and 2G4 to dsRNA and not to any other nucleic acid species tested , suggests that they recognise dsRNA via the A-form sugar-phosphate backbone in a manner similar to cellular dsRBPs . The sequence-independent recognition of dsRNA by these mAbs highlights their potential as useful tools for discovery and identification of unknown and novel viruses from biological samples . While most vertebrate-infecting flaviviruses tested in fixed-cell ELISA were detected using mAbs 3G1 and 2G4 , we were only able to demonstrate limited detection of DENV in cell culture by this method ( S2A Fig ) . Interestingly , when DENV-infected cells were tested by IFA performed in parallel , mAbs 3G1 and 2G4 showed reactivity to DENV-infected cells similar to that observed for cells infected with WNVKUNV ( S2B Fig ) . Quantitative PCR analysis of viral RNA present in infected lysates suggested that this discrepancy was not due to lower levels of viral replication in DENV-infected cells , and was unlikely to be due to an issue with sensitivity on the part of the MAVRIC ELISA ( S2C Fig ) . We hypothesise that the lower concentration of acetone used in fixation for fixed-cell ELISA may not be sufficient for exposure of dsRNA to mAbs in DENV-infected cells . This warrants further investigation . Finally , using a streamlined ELISA-based system we have demonstrated the potential of these mAbs for broad-spectrum surveillance of mosquito populations for infection with known circulating arboviruses . This system has been instrumental in the detection of four novel virus isolates from mosquito homogenates using a streamlined ELISA-based system , including LNV which until recently was not known to be present in Australia [10] . We have demonstrated that these antibodies detect flavivirus and alphavirus infection at 5 days post-infection with similar sensitivity to viral protein-specific antibodies commonly used for arbovirus surveillance . In addition , antibody 3G1 detected similar levels of RRV- and WNV-infection as the corresponding virus-specific mAbs at 48 hours post-infection . In contrast , 2G4 was able to detect RRV-infection to similar levels at 48 hours post-infection but only detected similar levels of WNV at 5 days post-infection . This suggests that mAb 3G1 is more sensitive than 2G4 , which may be due to the differing affinities of the two antibodies . Blind analysis of previously processed field samples using fixed-cell ELISA with mAbs 3G1 and 2G4 demonstrated the high sensitivity of this method as well as the ability to detect viruses which were not detected by conventional means . Of the 32 samples that were positive for known viruses by conventional methods , the fixed-cell ELISA with mAbs 3G1 and 2G4 detected all but 3 . One of these samples ( pool 1026 ) was also negative by RT-PCR suggesting that the virus in this sample was no longer viable . Two samples were negative by fixed-cell ELISA using mAbs 3G1 and 2G4 , but positive by RT-PCR ( pool 3047 , WNVKUNV; pool 4189 , ALFV ) , suggesting that if viable virus was present in these samples , that it was below the threshold of detection of this assay , leading to a false negative rate of 2/32 ( 6 . 25% ) . However , it must be highlighted that the detection of residual RNA from the inoculum by RT-PCR is possible and thus it was not confirmed whether viable virus was present in these samples . An additional six samples that were not positive by virus-specific ELISA or RT-PCR were positive by the 3G1 and 2G4 fixed-cell ELISA , subsequent analysis of these samples found that four pools caused visible CPE indicating the presence of replicating virus in these samples . Two of these samples which were negative for known viruses ( pools 4298 and 4271 ) , were positive by 3G1 and 2G4 ELISA but did not present with overt CPE . This highlights the power of this assay for the detection of novel viruses which may otherwise go unnoticed by conventional methods [6] . The two antibodies described here , now referred to as MAVRIC ( monoclonal antibodies against viral RNA intermediates in cells ) , provide a novel approach to broad spectrum virus surveillance and discovery . We have demonstrated the ability of these mAbs to detect infection with viruses from a diverse range of families both in ELISA and IFA with similar sensitivity to that of fixed-cell ELISA using virus-specific monoclonal antibodies . We have recently used these mAbs in our lab to discover and characterise a number of novel flaviviruses , mesoniviruses and reoviruses ( Table 3 ) ( 9 ) . This system combines the cost-effective platform of fixed-cell ELISA , with broad-spectrum detection of viruses based on the presence of dsRNA , independent of sequence . Thus , these antibodies provide a valuable tool for cost-effective , high-throughput screening of both known and unknown viruses in biological samples . Although the focus of our study was on the detection of arthropod-borne viruses in mosquito cells , we have also shown that both antibodies are able to recognise viral dsRNA species in Vero and DF-1 cells , suggesting this approach would also be applicable to the detection and discovery of various mammalian and avian viruses . | This paper describes a simple and cost-effective system for screening biological samples for virus-infection . The authors demonstrate the application of two antibodies to detect double-stranded RNA ( dsRNA ) which is a common molecule produced in infection by a number of different viruses . The use of antibodies which react with double-stranded RNA independently of sequence allows for detection of a diverse range of viruses and has been instrumental in the detection of known arboviruses from three different families and the discovery of a number of previously unknown viruses from Australian mosquito populations . This system provides a rapid and economical approach to virus surveillance and discovery . This is the first report of anti-dsRNA antibodies used in a streamlined system for virus detection and discovery in field-caught samples . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [] | 2015 | Viral RNA Intermediates as Targets for Detection and Discovery of Novel and Emerging Mosquito-Borne Viruses |
In plants , specific recognition of pathogen effector proteins by nucleotide-binding leucine-rich repeat ( NLR ) receptors leads to activation of immune responses . RPP1 , an NLR from Arabidopsis thaliana , recognizes the effector ATR1 , from the oomycete pathogen Hyaloperonospora arabidopsidis , by direct association via C-terminal leucine-rich repeats ( LRRs ) . Two RPP1 alleles , RPP1-NdA and RPP1-WsB , have narrow and broad recognition spectra , respectively , with RPP1-NdA recognizing a subset of the ATR1 variants recognized by RPP1-WsB . In this work , we further characterized direct effector recognition through random mutagenesis of an unrecognized ATR1 allele , ATR1-Cala2 , screening for gain-of-recognition phenotypes in a tobacco hypersensitive response assay . We identified ATR1 mutants that a ) confirm surface-exposed residues contribute to recognition by RPP1 , and b ) are recognized by and activate the narrow-spectrum allele RPP1-NdA , but not RPP1-WsB , in co-immunoprecipitation and bacterial growth inhibition assays . Thus , RPP1 alleles have distinct recognition specificities , rather than simply different sensitivity to activation . Using chimeric RPP1 constructs , we showed that RPP1-NdA LRRs were sufficient for allele-specific recognition ( association with ATR1 ) , but insufficient for receptor activation in the form of HR . Additional inclusion of the RPP1-NdA ARC2 subdomain , from the central NB-ARC domain , was required for a full range of activation specificity . Thus , cooperation between recognition and activation domains seems to be essential for NLR function .
A critical step in the lifestyles of plant pathogens is the secretion of effectors—pathogen-encoded proteins that are translocated into the plant cell , where they manipulate the host and promote pathogen growth [1] . Many effectors function to modulate basal immunity , but their presence in the plant cell may betray the pathogen and activate a second layer of effector-triggered immunity ( ETI ) if recognized by intracellular host immune receptors [2–4] . Most effector recognition occurs via the NLR family of immune receptors ( for nucleotide-binding and leucine-rich repeat protein ) . NLR activation results in an elevated immune response , characterized by generation of reactive oxygen species , activation of defense-associated genes , and a localized cell death known as the hypersensitive response ( HR ) [5] . Multiple modes of effector-triggered NLR activation have been described . Well-studied plant NLRs , such as RPS2 and RPM1 from Arabidopsis , recognize effectors indirectly [6–8] . These NLRs are activated not by association with effectors themselves , but instead by recognizing their biochemical effects in the plant cell , leading to models of NLR activation in which the receptors recognize perturbation of “guarded” host proteins . Guarded proteins can be true virulence targets or simply decoys inviting modification by the pathogen [9 , 10] . In contrast , NLRs including RPP1 , L6 , Pi-ta , and others interact directly with recognized effector alleles , suggesting a second model in which direct effector-NLR interaction is required for immune activation [11–15] . The modular domain architecture of NLRs allows for the integration of effector recognition and signaling activation by a switch-like mechanism [16] . An N-terminal domain , usually either a coiled-coil or Toll/interleukin-1-receptor ( TIR ) domain , mediates downstream immune signaling [17 , 18] and is followed by a nucleotide binding ( NB ) domain and two helical ARC subdomains ( for Apaf-1 , R protein , CED-4 ) . This composite NB-ARC domain functions as a switch through exchange of an internally bound ADP for ATP [19 , 20] , but may also participate in cis-regulatory interactions in order to maintain an ADP-bound “off” state [21–23] . These intramolecular interactions typically occur with a series of C-terminal leucine-rich repeats ( LRRs ) , which are critical for auto-inhibition in the absence of effector [24 , 25] . A second role for the LRRs is in effector recognition , where the protein-protein interaction capacity of LRRs can mediate effector binding [12 , 13] . Consistent with this role , NLR recognition specificity can be expanded through LRR variation [26 , 27] . Positively selected amino acids in Arabidopsis NLR proteins genome-wide are also disproportionately located in the LRRs , consistent with LRR co-evolution with effector proteins under selective pressure to evade recognition [28] . Recognition of the oomycete effector ATR1 by the Arabidopsis NLR RPP1 is consistent with the direct interaction model of receptor activation , and serves as a model system for studying the molecular basis of NLR function [29] . ATR1 is one of approximately 140 effectors expressed and secreted by the naturally-occurring Arabidopsis pathogen Hyaloperonospora arabidopsidis ( Hpa ) [30] , and is recognized specifically by the NLR protein RPP1 , leading to Hpa resistance [31] . Diverse ATR1 alleles from Hpa strains encode effectors that are differentially recognized by RPP1 [12 , 32] , and thus ATR1 can condition strain-dependent resistance on a given Arabidopsis ecotype [33] . Variation in the RPP1 receptor also contributes to the spectrum of resistance phenotypes; for example , RPP1-NdA ( from the Niederzenz ecotype ) and RPP1-WsB ( from the Wassilewskija ecotype ) vary in recognition specificity , with RPP1-NdA recognizing a smaller subset of the ATR1 alleles recognized by RPP1-WsB [12] . As Hpa is an obligate biotroph , surrogate systems are used to study the molecular basis for ATR1 recognition by RPP1 . Alleles of ATR1 and RPP1 can be co-expressed in Nicotiana tabacum , resulting in a visible HR only for combinations in which RPP1 is able to recognize the ATR1 variant [12] . Biochemical and genetic lines of evidence from this system support a direct interaction model of ATR1 recognition . Co-immunoprecipitation of ATR1 alleles with RPP1-WsB correlates with HR activation capability [12] , suggesting that direct interaction with the effector leads to receptor activation and signaling for ETI . Furthermore , the LRR domain of RPP1-WsB is sufficient for association with ATR1 , indicating a role for the LRRs in effector recognition . ATR1 has no known virulence function , and adopts a WY-domain fold common to oomycete effector proteins [34] , specifically with an N-terminal three-helix bundle and two tandem WY-domains comprising the “head” and “body” of a seahorse-like structure , respectively [35] . Single amino acid substitutions on both the head and body region can confer gain-of-recognition phenotypes to an unrecognized ATR1 allele , ATR1-Cala2 [35] . The surface-exposure of these substituted residues is consistent with these substitutions altering protein-protein interaction strength between ATR1 and RPP1 [12 , 35] . In this work , we address several outstanding questions regarding NLR receptor activation using the ATR1-RPP1 system . We performed a random mutagenesis screen of unrecognized ATR1-Cala2 to generate combinations of ATR1 mutations that activate each RPP1 allele . We show that different ATR1 mutants specifically associate with and activate resistance against either RPP1-NdA or RPP1-WsB , defining distinct recognition specificities for the two RPP1 alleles . We then used these ATR1 mutants to probe recognition and activation domains of both RPP1 alleles . Chimeric receptors revealed that while the LRRs were sufficient for recognition of ATR1 through molecular association , they were insufficient to recapitulate a receptor’s full range of specificity . Instead , inclusion of the ARC2 subdomain is further required for effective receptor activation .
Previous work in our lab employed natural variation across ATR1 alleles to identify effector surfaces involved in recognition by RPP1 . Amino acids conserved in recognized ATR1 variants were substituted into the distantly related and unrecognized allele , ATR1-Cala2 . Four substitutions were each sufficient to give gain-of-recognition HR phenotypes by RPP1-WsB , but not RPP1-NdA , and combining all four substitutions led to a robust response with similar timing and intensity to the naturally recognized allele , ATR1-Emoy2 [35] . Here , we term this combined mutant ATR1-Cala2 WsB-GOF ( Gain-of-Function ) . In combination with the crystal structure of ATR1 , these results indicated that surface-exposed residues on a central WY-domain were involved in recognition by RPP1-WsB [35] . While single amino acid substitutions can confer RPP1-WsB recognition , no natural alleles or mutants of ATR1 exclusively activate RPP1-NdA . To test whether other ATR1 surfaces could confer gain-of-recognition phenotypes , and whether allele-specific mutants could be obtained , we performed a random mutagenesis screen of ATR1-Cala2 for gain-of-recognition by either RPP1 allele ( Fig . 1 ) . We transiently co-expressed 2 , 240 clones of Agrobacterium tumefaciens expressing the mutagenized effector domain ( Δ51 ) of ATR1-Cala2 with RPP1-NdA and RPP1-WsB in Nicotiana tabacum and screened for visible cell death HR , indicating NLR receptor activation . Only two mutants were recovered from the screen: either a valine substitution at position 88 ( E88V ) or a set of three combined substitutions at positions 139 , 140 , and 142 ( S139T/Y140H/G142R ) was each sufficient to confer weak recognition by RPP1-NdA , but not RPP1-WsB at 48hpi , and both mutants developed stronger activation of RPP1-NdA by 72hpi ( Fig . 2A ) . Allele-specific responses could be combined , as combining E88V with the previously described WsB-GOF substitutions led to recognition by both RPP1 alleles ( S1 Fig . ) . HR strength varies by leaf age , and older , less sensitive leaves did not show similar gain-of-recognition phenotypes for either mutant . However , combining all four substitutions into a single construct , termed ATR1-Cala2 NdA-GOF , allowed for more robust activation of RPP1-NdA than either individual substitution ( Fig . 2A right , alternate leaf ) , albeit still weaker than activation by ATR1-Emoy2 or ATR1-Cala2 WsB-GOF of their respective RPP1 alleles ( Fig . 2A left , 48 hpi ) . All ATR1 mutants expressed to a similar level to the unrecognized WT ATR1-Cala2 allele ( Fig . 2B ) . Thus , mutant variants of ATR1-Cala2 activate RPP1-NdA but not RPP1-WsB , defining unique recognition specificities for each RPP1 allele . We next mapped the NdA-GOF mutations onto a homology model of the ATR1-Cala2 structure , using the solved ATR1-Emoy2 structure as a template ( Fig . 2C , see S2 Fig . for amino acid alignment ) . All four substitutions are predicted to be completely or partially surface exposed ( >25 Å2 exposed ) , consistent with a predicted role in direct interaction with RPP1 . All four substitutions also fall within the previously defined minimal region for RPP1 recognition [35] , further indicating that this helical region specifies recognition by RPP1-NdA . E88V occurs on an N-terminal three-helix bundle . S139T/Y140H/G142R occur on the first of two tandem WY-domain repeats , at the N-terminal portion of the α1 helix as defined across other RXLR-type oomycete effectors [34 , 36] . Thus mutations on the specific surfaces on the “head” and “body” of the seahorse-like ATR1 structure can lead to specific activation of RPP1-NdA , but not RPP1-WsB , as summarized in Fig . 2D . Three of the four constituent substitutions in the ATR1-Cala2 NdA-GOF mutant altered the predicted ATR1 surface charge . These three mutations either substitute a positively charged side chain for a neutral side chain ( Y140H , G142R ) in the first WY-domain or a neutral side chain for a negative side chain ( E88V ) in the N-terminal three-helix bundle . We hypothesized that any alteration in surface charge on these ATR1 surfaces would affect recognition by RPP1 . This is further supported by the fact that one of the component NdA-GOF substitutions , Y140H , occurs at the same residue as a previously identified WsB-GOF substitution , Y140D [35] . We generated alanine , lysine , and arginine substitutions at positions 88 and 140 in an ATR1-Cala2 background . Only two of these substitutions , E88R and E88A , conferred weak gain-of-recognition by RPP1-NdA but not RPP1-WsB ( Fig . 3A ) . This HR reaction was , however , weaker than the originally identified mutation , E88V , despite similar expression levels of all mutants ( Fig . 3B ) . Thus , simple changes in surface charge do not provide a consistent pattern of recognition phenotypes against RPP1-NdA or RPP1-WsB . Previously , the ability of ATR1 alleles to co-immunoprecipitate RPP1-WsB correlated with activation of HR upon transient co-expression [12] . We tested whether the novel gain-of-recognition HR phenotypes of ATR1-Cala2 NdA-GOF and WsB-GOF also correlated with allele-specific RPP1 association . ATR1-Cala2 WsB-GOF associated with RPP1-WsB at a level similar to the recognized allele , ATR1-Emoy2 , and did not associate with RPP1-NdA ( Fig . 4 ) . ATR1-Cala2 NdA-GOF did not associate with RPP1-WsB , but associated with RPP1-NdA , although more weakly than ATR1-Emoy2 ( Fig . 4 ) , consistent with the weaker HR phenotype we observed for this mutant ( Fig . 2A ) . Overall , the association of ATR1 mutants with RPP1 correlates with HR phenotypes in tobacco , consistent with direct interaction of these mutants with corresponding RPP1 proteins . Although Hpa is an obligate biotroph and has not been successfully cultured or genetically manipulated [32] , surrogate approaches allow delivery of Hpa effectors into the Arabidopsis host by bacterial type three delivery to assay for induced host defense responses [37 , 38] . We tested whether HR and association phenotypes for ATR1 GOF mutants observed in transient assays correlated with resistance phenotypes in Arabidopsis upon bacterial delivery . ATR1 alleles and mutants were fused with the secretion signal of AvrRps4 , mediating delivery by the endogenous type-three secretion system ( TTSS ) of strain DC3000 of the virulent bacterium , Pseudomonas syringae pv . tomato . Strains delivering alleles and mutants of ATR1 were inoculated into the recombinant inbred Arabidopsis line HRI3860 , which lacks functional RPP1 [32] , as well as transgenic HRI3860 lines expressing RPP1-NdA or RPP1-WsB . While all strains grew to similar levels by 3 days post-inoculation ( dpi ) on HRI3860 plants , delivery of ATR1-Emoy2 strongly inhibited bacterial growth in transgenic lines expressing either RPP1 allele ( Fig . 5A ) . Strains with either an empty vector or delivering ATR1-Cala2 were uninhibited in growth on either transgenic line . ATR1-Cala2 WsB-GOF delivery strongly inhibited bacterial growth on the RPP1-WsB expressing line , while ATR1-Cala2 NdA-GOF weakly inhibited growth on the RPP1-NdA expressing line ( Fig . 5A ) . Disease symptoms at 3 dpi correlated with growth inhibition . While ATR1-Cala2 delivering strains remained visibly virulent on both transgenic lines , producing chlorosis and necrotic lesions , delivery of ATR1 GOF alleles led to RPP1-NdA or RPP1-WsB-specific avirulence , leading to a healthy phenotype similar to that observed after delivery of the fully recognized allele , ATR1-Emoy2 ( S3A Fig . ) . We also tested whether gain-of-recognition alleles could elicit HR phenotypes in the same RPP1-expressing transgenic lines . Delivery of different ATR1 alleles by Pseudomonas fluorescens ( Pf0 ) engineered to express a TTSS was previously shown to yield allele-specific HR in Arabidopsis , correlating with HR phenotypes in the transient tobacco assay [33 , 39] . The ATR1-Cala2 WsB-GOF mutant elicited a visible HR on a RPP1-WsB transgenic line , but we were unable to detect activation of strong HR by Pf0 delivering ATR1-Cala2 NdA-GOF to an RPP1-NdA transgenic line ( S3B Fig . ) . This weaker HR recognition phenotype of the NdA-GOF mutant is comparable to growth inhibition , Co-IP , and transient HR phenotypes ( Fig . 2A , Fig . 4 , Fig . 5A ) . Despite the weak recognition and resistance phenotypes activated by ATR1-Cala2 NdA-GOF relative to those of ATR1-Cala2 WsB-GOF , the exclusivity of the mutants for activating either RPP1-NdA or RPP1-WsB allowed us to explore regions governing recognition and activation for each RPP1 protein . We hypothesized that , since NdA-GOF and WsB-GOF mutations occurred on different ATR1 surfaces ( Fig . 2C ) , unique receptor regions might be responsible for recognition of each mutant . We generated several chimeric RPP1 constructs to test the contribution of different NLR domains to activation ( Fig . 6A ) . As the leucine-rich repeats ( LRRs ) of RPP1-WsB are sufficient for association with ATR1-Emoy2 [12] , we first generated reciprocal chimeric constructs between RPP1-NdA and RPP1-WsB in which all predicted C-terminal LRRs were swapped ( see S4 Fig . for chimeric exchange points on pairwise sequence alignment ) . Neither WsB LRRs in an NdA context ( NdA605WsB ) nor NdA LRRs in a WsB context ( WsB598NdA ) led to autoactivity or activation by ATR1-Cala2 , and both retained the conserved ability to recognize ATR1-Emoy2 ( Fig . 6B ) . We next tested whether the LRRs from one RPP1 allele were sufficient to recognize the cognate ATR1-Cala2 NdA-GOF or WsB-GOF mutant . As expected , NdA605WsB , a chimera with RPP1-WsB LRRs , was fully activated by ATR1-Cala2 WsB-GOF ( Fig . 6C ) . To determine a potential cutoff point for recognition along the series of C-terminal LRRs , we constructed a series of LRR chimeras in which the chimeric exchange point was made every 4–5 repeats further C-terminal , based on homology-modeled LRRs [29] . Most chimeric exchanges were completely inactivated: NdA825WsB and NdA916WsB were detected by Western blot , but , unlike WT RPP1-NdA and RPP1-WsB , did not recognize ATR1-Emoy2 ( S5A Fig . ) . A third chimeric LRR construct , NdA1026WsB , was not expressed ( S5C Fig . ) . One chimera , however , NdA708WsB , expressed ( S5C Fig . ) , recognized ATR1-Emoy2 , but showed diminished HR in response to ATR1-Cala2 WsB-GOF relative to NdA605WsB ( Fig . 6C ) , indicating that the first 4 LRRs of RPP1-WsB are required for full recognition of this mutant . Surprisingly , a reciprocal chimera with RPP1-NdA LRRs , WsB598NdA , was only weakly activated by ATR1-Cala2 NdA-GOF ( Fig . 6D ) , only visible as mild HR on the backside of the leaf ( S5B Fig . ) . This indicated that the RPP1-NdA LRRs are insufficient for recapitulating the full range of RPP1-NdA specificity . We generated constructs with chimeric fusions further N-terminal relative to the LRRs , at the TIR-NBS , NBS-ARC1 , and ARC1-ARC2 domain junctions . While a chimera exchanging the TIR domain ( WsB266NdA ) expressed but was inactive even in recognizing the conserved recognized ATR1 allele , ATR1-Emoy2 ( S5A Fig . ) , all other constructs recognized ATR1-Emoy2 with similar timing and intensity ( Fig . 6D , S4B Fig . ) . Chimeras with RPP1-NdA C-termini comprising either the ARC2-LRR ( WsB473NdA ) or ARC1-ARC2-LRR ( WsB417NdA ) domains were able to activate in response to ATR1-Cala2 NdA-GOF ( Fig . 6D , S5B Fig . ) , with the strongest response generated by RPP1-NdA ARC2-LRRs in an RPP1-WsB context ( WsB473NdA ) . Thus , while the RPP1-NdA LRRs are insufficient to allow activation by ATR1-Cala2 NdA-GOF , inclusion of an RPP1-NdA ARC2 domain in chimeric constructs allows for receptor activation and contributes to RPP1-NdA specificity . The RPP1-NdA ARC2 domain could expand specificity in two possible ways—either the ARC2 domain is involved in recognition of the ATR1 mutant through allele-specific contacts , or it instead plays a role in facilitating activation upon ATR1 recognition by the LRRs . We carried out experiments to distinguish between these binding versus activation roles . A binding role would be supported by sufficiency of the RPP1-NdA ARC2 subdomain for both association and activation by ATR1-Cala2 NdA-GOF . To test an association role , we expressed only the NB-ARC or ARC1-ARC2 domains ( amino acids 296–605 or 424–605 respectively ) and probed for co-immunoprecipitation with ATR1-Cala2 NdA-GOF or ATR1-Emoy2 . Co-immunoprecipitation of either domain with ATR1 was not observed ( S6 Fig . ) , but the RPP1-NdA LRRs were also unable to interact with ATR1 in these experiments ( S7 Fig . ) . We thus directly tested ARC2 sufficiency for activation by generating a double chimeric construct , WsB473NdA605WsB , with an RPP1-NdA ARC2 subdomain in an RPP1-WsB context ( see cutoffs in S4 Fig . ) . This chimera did not activate in response to ATR1-Cala2 NdA-GOF ( Fig . 7A ) . Thus the RPP1-NdA ARC2 domain is insufficient for activation by ATR1-Cala2 NdA-GOF , a finding inconsistent with a model where direct ARC2 contacts fully condition receptor activation . An alternative , activation state model for expansion of specificity by the ARC2 domain is that the RPP1-NdA LRRs determine allele-specific recognition of ATR1 , but a corresponding RPP1-NdA ARC2 domain is required for its full activation capacity . One prediction of this model is that RPP1-NdA LRRs should be sufficient to associate with the ATR1-Cala2 NdA-GOF mutant . Unlike with RPP1-WsB LRRs , we were unable to co-immunoprecipitate RPP1-NdA LRRs ( amino acids 606–1164 ) even with the fully recognized allele , ATR1-Emoy2 ( S7 Fig . ) . We were , however , able to investigate an LRR binding role by co-immunoprecipitating full-length chimeric constructs . While WT ATR1-Cala2 did not co-immunoprecipitate any chimeric RPP1 constructs , the ATR1-Cala2 NdA-GOF mutant associated with RPP1 chimeras containing an RPP1-NdA LRR , despite different activation phenotypes in the HR assay ( Fig . 7B ) . Although this interaction of WsB598NdA with ATR1-Cala2 NdA-GOF was weaker than with ATR1-Emoy2 , the level was similar to that of interaction with WT RPP1-NdA , suggesting that different receptor sensitivities , and not association strengths , condition the intensity of allele-specific HR . Finally , the activation state model predicts that the threshold for specificity by a strongly recognized ATR1 allele will be lower for the desensitized WsB598NdA chimera . We tested transient expression of wild-type and chimeric RPP1 constructs against a gradient inoculum of a second Agrobacterium strain expressing ATR1-Emoy2 . From typical to low inoculum ( OD = 0 . 45 to OD = 0 . 03 ) , ATR1-Emoy2 activates all constructs evenly ( Fig . 7C ) . However , lowering the inoculum to OD = 0 . 02 led to specific activation of an RPP1-NdA ARC2-LRR containing construct , WsB473NdA , but not an LRR-only construct , WsB598NdA , consistent with a sensitizing effect of an allelic RPP1-NdA ARC2 domain . In summary , RPP1-NdA LRRs are sufficient for association with , but insufficient for full-strength activation by , an NdA-specific ATR1 mutant . Thus RPP1-NdA LRRs condition recognition , but efficient receptor activation further requires an RPP1-NdA ARC2 activation domain .
A set of mutations arising from our random mutagenesis screen of unrecognized ATR1-Cala2 conferred exclusive recognition by RPP1-NdA , but not RPP1-WsB ( Fig . 2A ) . These NdA-GOF mutations occurred on both the “head” region of ATR1 ( N-terminal three-helix bundle ) and on the “body” ( in the first of two tandem helical WY domains ) ( Fig . 2B ) . This structural location is consistent with previously described mutations of a differentially recognized allele , ATR1-Maks9 , which is recognized by RPP1-WsB but not RPP1-NdA . Both a “head” substitution , E92K , and a “body” substitution , D191G , conferred RPP1-NdA recognition to ATR1-Maks9 [12] . Site-directed substitution of positively charged and neutral sidechains on these surfaces in ATR1-Cala2 did not lead to consistent gain-of-recognition phenotypes ( Fig . 3 ) , suggesting that more intricately defined surface interactions mediate specificity for each allelic pair . Nonetheless , the location of ATR1-Cala2 NdA-GOF substitutions is consistent with multiple surfaces contacting RPP1-NdA . The GOF residues described here and previously [12 , 35] also have a higher degree of surface exposure than the overall molecule ( average of 81 Å2 relative to 65 Å2 for ATR1 overall ) , further consistent with surface contacts with RPP1 . Molecule-level resolution data on these interacting surfaces , for example from crosslinking or crystallography experiments , will likely inform the basis of the recognition phenotypes described here . The ATR1-Cala2 NdA-GOF mutant provides further evidence that direct activation of NLRs can show a gradient of response strength that depends on a variety of effector-NLR contacts . First , the four ATR1-Cala2 NdA-GOF substitutions are additive in strength of transient HR ( Fig . 2A ) , consistent with multiple contact points with RPP1-NdA that quantitatively increase binding strength . The ATR1 Cala2 NdA-GOF activated responses were also weaker compared to that activated by the fully recognized allele , ATR1-Emoy2 , in HR , co-immunoprecipitation , and bacterial growth inhibition assays ( Fig . 2A , Fig . 4 , Fig . 5A ) . Second , three of four residues substituted in ATR1-Cala2 NdA-GOF are also conserved in the recognized allele , ATR1-Emoy2 ( S4 Fig . ) . This suggests that RPP1-NdA recognition of the ATR1-Emoy2 allele occurs via interaction with different amino acid residues . Together , these data indicate that RPP1-activated resistance is quantitative in strength and that different allelic combinations can depend on distinct ATR1-RPP1 contact points . This flexibility of the interaction likely contributes to high levels of amino acid variation in both the recognition domain of ATR1 and the LRR region of RPP1 [29] . Quantitative NLR activation strength may also underlie the partial resistance phenotypes observed for many Hpa-Arabidopsis interactions [33] . Formally , co-immunoprecipitation could indicate complex formation through an intermediate host factor . However , the nature of all gain-of-function ATR1 mutations described—surface substitutions activating specific RPP1 alleles in an additive fashion—is more consistent with recognition conditioned by direct association . Biochemical characterization of interaction strength between RPP1 and various ATR1 mutants may correlate affinity with subtle HR phenotypes described here , but these experiments await reliable methods for RPP1 protein expression and purification , which have been recalcitrant to date . The ATR1-Cala2 NdA-GOF mutant also provides evidence for a high degree of specificity in the recognition spectra of RPP1 alleles . RPP1-NdA only recognizes a subset of naturally occurring ATR1 alleles recognized by RPP1-WsB [12 , 33] , and thus prior to this study we could not distinguish between two competing hypotheses—either both alleles have distinct but overlapping recognition spectra , or RPP1-WsB is simply more sensitive to activation by a wider array of ATR1 variants . Support for the former , specificity-based model comes from our random mutagenesis screen . First , we did not recover any ATR1 mutants that activated RPP1-WsB; a more sensitive receptor might be expected to recognize a larger range of random mutants . Second , the fact that the mutant described here , ATR1-Cala2 NdA-GOF , is recognized by and associates with RPP1-NdA but not RPP1-WsB disproves the hypothesis that the expanded RPP1-WsB recognition spectrum is due to increased sensitivity . Rather , it is consistent with the closely related receptors having sophisticated , individual recognition abilities , in addition to their shared ability to recognize certain ATR1 variants . “Arms race” co-evolution of effector and receptor [40 , 41] likely leads to the unique specificities of the two RPP1 alleles described here . Pathogen-driven receptor diversity may also lead to autoimmune consequences for the host , as RPP1 variants from other Arabidopsis ecotypes can condition hybrid incompatibility through genetic interaction with other loci [42] . Allele-specific activation by the ATR1-Cala2 NdA-GOF and WsB-GOF mutants allowed us to explore recognition regions in the respective recognizing RPP1 alleles . A chimeric exchange placing RPP1-WsB LRRs in a RPP1-NdA context led to full ATR1-Cala2 WsB-GOF recognition ( Fig . 6C ) , consistent with a recognition role for the LRRs . A smaller exchange excluding the first 4 LRRs gave a highly reduced recognition phenotype ( Fig . 6C , NdA605WsB vs . NdA708WsB ) , while still fully recognizing ATR1-Emoy2 . Polymorphic residues between RPP1-NdA and RPP1-WsB in these 4 LRRs likely mediate specificity for recognition of ATR1-Cala2 WsB-GOF ( S8 Fig . , left ) , including several residues on a predicted concave β-sheet surface associated with ligand binding in other LRRs [43] . Surprisingly , chimeric exchanges indicated a role of the RPP1-NdA ARC2 helical domain in activation . LRRs from RPP1-NdA did not allow for complete activation by ATR1-Cala2 NdA-GOF ( Fig . 6D , WsB598NdA ) , but a further N-terminal chimeric exchange including the RPP1-NdA ARC2 domain ( WsB473NdA ) greatly strengthened the response . Further experiments testing chimeric , full-length RPP1 constructs against this ATR1 mutant indicated that the ARC2 domain expands specificity by facilitating receptor activation rather than by associating with the effector . The RPP1-NdA ARC2 domain in an RPP1-WsB context was insufficient for receptor activation by ATR1-Cala2 NdA-GOF , and an RPP1-NdA ARC2 domain was not required for its allele-specific ATR1 association ( Fig . 7A , B ) . In addition , a chimera with RPP1-NdA ARC2 was able to activate in response to a lower inoculum of the fully recognized ATR1-Emoy2 allele ( Fig . 7C ) . Thus the ARC2 domain of RPP1-NdA functions to expand its specificity by decreasing the threshold for activation of the receptor . We speculate that polymorphisms between RPP1-NdA and RPP1-WsB in the ARC2 domain condition intramolecular allelic compatibility between domains of the receptor , possibly through interactions with the LRRs . Several ARC2 polymorphisms occur on the predicted surface of a homology modeled ARC2 , and could be candidates for LRR interaction ( S8 Fig . , right ) . The structural basis of plant NLR receptor activation remains unclear , but there is increasing evidence for a role of the NB-ARC domain not just as a nucleotide hydrolysis-based switch downstream of effector perception [19] , but as an active contributor to effector-triggered activation in combination with C-terminal LRRs . For example , recent data from chimeric exchanges in animal intracellular NLRs , the mouse NAIPs ( NLR family , apoptosis inhibitory protein ) , corroborate a role for central helical domains in specificity . NAIPs oligomerize and recruit Caspase-1 upon recognition of specific ligands [44] , and a series of chimeric exchanges between NAIP2 and NAIP5 indicated that specificity of oligomerization was determined by the second and third ARC-like helical domains rather than by the C-terminal LRRs [45] . Examples from plant NLRs also support a role for ARC domains in specificity . ARC helical domains of the flax receptor L can , in tandem with LRR polymorphisms , expand recognition of AvrL567 effector alleles [46] . Recently , ARC2 domain mutations were described in the wheat NLR Pm3 that expand its specificity against previously unrecognized wheat powdery mildew strains [47] . In contrast to inactivating or autoactivating mutations in other NLRs that map near the nucleotide binding pocket [24 , 48 , 49] , the Pm3 recognition-expanding mutations map to a region of the ARC2 subdomain predicted to be an exposed loop [47] , suggesting that the mutations affect intra- or intermolecular interactions . Intramolecular contacts between the ARC2 and LRR domains are thought to maintain an “off” state [23 , 25 , 46 , 50] , and specific ARC2 amino acid substitutions can affect ARC2-LRR binding affinity [22] . Here we provide data that the ARC2 subdomain is dispensable for effector association , but required for full-strength activation , in an allele-specific effector-NLR interaction . Thus compatibility between ARC2 and LRRs , even in closely related NLR variants such as RPP1-NdA and RPP1-WsB , may be required for a full range of specificity by allowing efficient receptor activation . We present a model of RPP1 function where cooperation between recognition by LRRs and activation by the ARC2 subdomain leads to a full-strength receptor response ( Fig . 8 ) . In an “off” state , RPP1-NdA is not activated by the unrecognized ATR1-Cala2 effector protein ( Fig . 8A ) . Substitutions on distributed surfaces of ATR1-Cala2 allow activation of RPP1-NdA but not RPP1-WsB ( Fig . 8B ) . Specificity is a multi-stage process: stepwise increases in recognition and activation strength in response to the ATR1 mutant can be achieved by substituting the recognition domain ( LRRs ) and activation domain ( ARC2 ) from RPP1-NdA . Complete activation strength is achieved with full intramolecular compatibility in the wild-type RPP1-NdA receptor . While our data provide new insight into the roles for specificity domains of ATR1 and RPP1 , it remains to be seen precisely how molecular contacts between the effector and receptor relieve plant NLR autoinhibition .
Escherichia coli strain DH5α was used for cloning and propagation of pEarleyGate and pEDV3 constructs , and was grown at 37°C in LB or LB agar supplemented with 25 μg /mL kanamycin or 10μg/mL gentamycin . Agrobacterium strain GV3101::pMP90 [51] was propagated at 28°C in LB supplemented with 50 μg/mL gentamycin . Pseudomonas strains were propagated at 28°C . Pseudomonas fluorescens ( Pf0 ) was grown on Pseudomonas Agar solid medium supplemented with 50μg/mL tetracyclin , 30 μg/mL chloramphenicol and 50 μg/mL gentamycin and Pseudomonas syringae pv . tomato DC3000 ( Pst DC3000 ) was grown on NYGA solid medium supplemented with 100 μg/mL rifampicin and 5 μg/mL gentamycin . FLAG-tagged ATR1 constructs were generated by attaching a primer encoded linker on the reverse primer , amplifying by PCR , and cloning into pENTR/D-TOPO ( Invitrogen ) . Error-prone PCR was carried out using the Diversify Random Mutagenesis Kit ( Clontech ) using condition 5 ( 640 μM manganese sulfate ) to amplify the Δ51 ATR1-Cala2-FLAG coding sequence . PCR reactions were recombined into pEarleyGate 202 [52] by LR recombination ( Invitrogen ) and populations of bacterial clones were pooled into libraries of random mutants . Prepared pEG202 ATR1-Cala2 from these libraries was transformed into Agrobacterium tumefaciens strain GV3101 by electroporation , and individual clones were selected for co-inoculation . Site-directed mutations were introduced into pENTR/D-TOPO ATR1-Cala2 through mutagenic primers using the QuikChange XL kit ( Stratagene ) . HR was assayed by co-inoculating Δ51 ATR1 constructs with RPP1-NdA or RPP1-WsB expressing Agrobacterium strains , containing genomic constructs in pEarleyGate 301 . Co-inoculation mixtures were infiltrated into 3–4 week old Nicotiana tabacum plants at OD = 0 . 45 per construct , as previously described [12] . Co-immunoprecipitation was performed using 24 hour post-infiltration N . benthamiana tissue , expressing FLAG-tagged ATR1 and 3xHA-tagged RPP1 . 1 g of tissue was homogenized , extracted , and processed as previously described [12] , with the following changes: 10 μL of Anti-FLAG affinity gel clone M2 ( Sigma ) was used for immunoprecipitation , while 1:1000 FLAG M2-Peroxidase ( Sigma ) antibody was used for immunoblotting . We modified the previously described EffectorDetectorVector pEDV3 [37] by exchanging the Sal1-EcoR1 fragment containing the HA tag with a linker containing Xho1-BamH1-Spe1-Flag-stop-EcoR1 . This vector , pEDV3F , was used as a backbone for BamH1/Spe1 insertion of the ATR1 gain-of-recognition mutants , which were PCR amplified from the corresponding pEarleygate clones . The resulting plasmids were sequence verified and conjugated into Pst DC3000 and Pf0 using triparental mating with E . coli strain PRK600 . Pst DC3000 disease assays were performed as previously described [33 , 39] . Briefly , the bacteria were resuspended in 10mM MgCl2 , adjusted to 1x105 cfu/mL , and inoculated into Arabidopsis leaves with a blunt syringe . Samples were taken at day 0 and day 3 , ground in 10mM MgCl2 with glass beads using a bead beater , diluted and plated on NYG agar containing rifampicin ( 50μg/ml ) , gentamycin ( 2 . 5μg/ml ) and cycloheximide ( 10μg/ml ) . For Arabidopsis hypersensitive response assays , Pf0 expressing the various ATR1 alleles were grown on Pseudomonas agar with glycerol and the appropriate antibiotics and resuspended in 10 mM MgCl2 to 1x109 cfu/ml . Half leaves of Arabidopsis were pierced and inoculated with bacterial suspension and symptoms scored at 24hpi . We constructed chimeric exchanges between RPP1-NdA and RPP1-WsB genomic constructs by overlap extension PCR . 5’ and 3’ chimeric fragments were generated by standard PCR and gel purified . Equimolar amounts of 5’ and 3’ fragments were used to self-prime an extension PCR reaction . Overlap extension reactions were carried out with 56°C annealing temperature for 15 cycles ( selecting for full length chimeric templates ) , followed by addition of end primers and 20 additional rounds with 64°C annealing temperature . Resulting chimeric PCR product was cloned into pENTR/D-TOPO through TOPO cloning ( Invitrogen ) , or ligated into a custom Gentamycin-resistant pENTR . Constructs were recombined into pEarleyGate 301 by LR recombination . Chimeras were tested in transient HR assays as described above . Primary amino acid sequences for ATR1-Cala2 , the RPP1-WsB NB-ARC , and LRR domains were modelled to ATR1-Emoy2 , Apaf-1 NB-ARC , and Toll-like receptor 3 structures using the Phyre2 server ( http://www . sbg . bio . ic . ac . uk/phyre2 ) [53] . Visualizations were generated using UCSF Chimera [54] . | Plants defend themselves against pathogens using specific multi-domain immune receptors , which are able to recognize secreted “effector” proteins from the pathogen , and thus activate an immune response . Variants of the Arabidopsis immune receptor RPP1 recognize different alleles of the oomycete effector ATR1 through direct association . RPP1 and ATR1 alleles from different ecotypes and strains show a spectrum of recognition phenotypes , reflecting coevolution by the plant and pathogen to evade and re-establish immunity . In this work , we identified mutations in an unrecognized ATR1 allele that lead to allele-specific recognition by RPP1 . Using chimeric constructs of the immune receptor , in which domains were swapped between two alleles , we were able to determine domains contributing to allele-specific activation . Our data point to the involvement of two domains in specific activation of immune receptors—one to associate with the effector , and one to sensitize the receptor and facilitate activation . We suggest that these domains must cooperate to efficiently and specifically recognize pathogen effectors . As NLRs confer pathogen resistance in many crop species , characterizing specificity domains involved in effector recognition will inform future efforts to breed or engineer disease resistant varieties . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2015 | Recognition and Activation Domains Contribute to Allele-Specific Responses of an Arabidopsis NLR Receptor to an Oomycete Effector Protein |
Usually , the occurrence of random cell behavior is appointed to small copy numbers of molecules involved in the stochastic process . Recently , we demonstrated for a variety of cell types that intracellular Ca2+ oscillations are sequences of random spikes despite the involvement of many molecules in spike generation . This randomness arises from the stochastic state transitions of individual Ca2+ release channels and does not average out due to the existence of steep concentration gradients . The system is hierarchical due to the structural levels channel - channel cluster - cell and a corresponding strength of coupling . Concentration gradients introduce microdomains which couple channels of a cluster strongly . But they couple clusters only weakly; too weak to establish deterministic behavior on cell level . Here , we present a multi-scale modelling concept for stochastic hierarchical systems . It simulates active molecules individually as Markov chains and their coupling by deterministic diffusion . Thus , we are able to follow the consequences of random single molecule state changes up to the signal on cell level . To demonstrate the potential of the method , we simulate a variety of experiments . Comparisons of simulated and experimental data of spontaneous oscillations in astrocytes emphasize the role of spatial concentration gradients in Ca2+ signalling . Analysis of extensive simulations indicates that frequency encoding described by the relation between average and standard deviation of interspike intervals is surprisingly robust . This robustness is a property of the random spiking mechanism and not a result of control .
Cellular behavior is the dynamics emerging out of molecular properties and molecular interactions . Hence , cells are indispensably subject to intrinsic noise due to the randomness of diffusion and molecule state transitions in gene expression [1] , [2] , signaling pathways and control mechanisms . It drives noise induced cell differentiation [3] , cell-to-cell variability of cloned cells [4] or second messenger dynamics [5] . While noise in gene expression can be attributed to small molecule numbers , we consider here noise in signalling pathways which occurs even in systems with large molecule numbers . Molecular interactions create nonlinear feedback like substrate depletion and allosteric regulation in enzyme kinetics or mutual activation of ion channels in membrane potential dynamics . They also couple active molecules inside cells spatially by diffusion of product and substrate or electric currents . If this coupling is strong enough , cells respond spatially homogeneous . Otherwise , we observe dynamic spatial structures formed by concentrations of molecules in specific states . These structures are often called microdomains [6]–[9] . The existence of these dynamic structures determines in some systems whether the cell obeys deterministic or stochastic mechanisms . The dynamic compartmentalization of the cell by concentration gradients may prevent the establishment of deterministic dynamics by the law of large numbers even if the total number of molecules in the cell would suggest it otherwise . Microdomains are too small to behave deterministically . Not even the whole ensemble of microdomains will behave deterministically , if they are only weakly coupled or if there are only a few of them . Consequently , noise is not averaged out on cell level . To determine whether we deal with a deterministic or stochastic system is important since these regimes may exhibit very different dependencies of behavior on system parameters [10] . For instance , repetitive spiking in intracellular signalling would be restricted to parameter values providing oscillatory dynamics with a deterministic mechanism [11] , [12] . It may occur with a stochastic system also for parameters which would lead to bistable or excitable dynamics in the deterministic limit , i . e . for larger or different parameter ranges [13] . In the non-oscillatory parameter ranges , the mechanism creating almost regular spike sequences can be coherence resonance [14]–[16] rather than the existence of a limit cycle in phase space of the local dynamics . Noisy systems with gradients usually show also a dependency of system characteristics on parameters of spatial coupling which spatially homogeneous systems do not exhibit . An example is the dependency of the spiking frequency on diffusion properties ( see below and [5] ) . In summary , the interaction between noise and gradients determines parameter dependencies and mechanisms . Recent experimental and theoretical studies on intracellular dynamics taught us that cells may indeed work in this regime and may exhibit repetitive spiking with non-oscillatory local dynamics . Functionally relevant gradients are also observed with intracellular cAMP [8] , [17]–[19] , pH [20] and in phosphorylation/dephosphorylation dynamics [21] , [22] suggesting that the lessons learned from dynamics may also apply to other systems . One of these lessons is that the randomness of single molecule state changes is carried up from the molecular level to cell level [23] , [24] . Cellular concentration spikes form random sequences of interspike intervals ( ISIs ) and that randomness arises from the randomness of single molecule state transitions [5] , [25] . Consequently , the fluctuations of cellular signals contain information on single molecule behavior . It is a task for modelling now to establish the relation between these fluctuations and single molecule properties to decode this information . Systems exhibiting the interaction between noise and gradients require modelling tools which can deal efficiently with the large concentration gradients and with the time scale range from molecular transitions to cell behavior . Here , we present such a modelling concept with the example of intracellular dynamics . It simulates all active molecules as stochastic Markov chains with all the individual state transitions and describes diffusion and some bulk reactions deterministically . Active molecules are those carrying the crucial feedbacks and nonlinearities . That allows for linearization of passive bulk reactions and the application of a multi-component Green's function to solve the partial differential equations in the cell analytically . We combine Green's functions with a local quasi-static approximation for the fast concentration changes and diffusion processes at the location of active molecules . That is possible due to the short diffusion time on the molecular length scale of a few nanometers . Since we use Green's functions for the long range concentration profiles we can restrict the calculation of concentration values to the location of active molecules . That renders this method extremely efficient even in 3 spatial dimensions . We will apply this concept to intracellular dynamics and compare simulated time dependent concentrations with single cell time series obtained from cultured astrocytes all measured under the same condition without any stimulation . is a ubiquitous second messenger in eukaryotic cells that transmits a variety of extracellular signals to intracellular targets . controls fertilization , cell differentiation , gene expression , learning and memory [26] . It triggers secretion in glands , muscle contractions in the heart and transmits apoptosis signals [27] , [28] . A main mechanism to increase the cytosolic concentration is release from intracellular stores , especially from the sarcoplasmic reticulum by ryanodine receptor channels ( RyRs ) or the endoplasmic reticulum ( ER ) by inositol 1 , 4 , 5-trisphosphate receptor channels ( ) . These channels open in a dependent fashion - a self amplifying effect known as induced release ( CICR ) [27] , [29] . If a single channel opens , is released into the cytosol , diffuses to adjacent channels and increases their open probability . Thus release may spread into the entire cell leading to a global cytosolic concentration spike . The inositol 1 , 4 , 5-trisphosphate ( ) pathway initiates release from the ER in many cell types ( including astrocytes [30] ) , since binding of to the primes them for activation by ( Figure 1 in Text S1 ) . The spatial arrangement of in channel clusters leads to a hierarchical system with the structural levels channel , channel cluster and cluster array , which is the cell level . pumps and buffers generate large gradients close to open channel clusters . Thus , channels within a cluster are strongly coupled and the coupling between clusters is only weak - the geometrical hierarchy entails a hierarchy of coupling strengths . Stochastic binding of and to the binding sites of leads to random opening of a single channel in a cluster [31] , [32] . This causes other channels of the same cluster to open also leading to a puff . An individual cluster is stochastic due to the small number of per cluster [33]–[35] . The opening of a single cluster can only be detected by adjacent clusters due to the strong gradients [23] , [24] , [27] , [36] , [37] . Since they are again only a few , it remains random whether they are opened by the initial puff . If a supercritical number of puffs arises , release spreads into the whole cell causing a global spike . Thus , due to the hierarchy of coupling strength , randomness is carried up from the channel level to the cell level . In order to model the hierarchical system , we have to consider the stochastic behavior of individual and the spatial heterogeneity of cells induced by clustering . That leads to a reaction diffusion system ( RDS ) with local stochastic source terms . For sufficient fast simulations , we decompose the system into local stochastic dynamics comprising channel state transitions and fast local concentration changes and a deterministic global dynamics for which we derive an analytical solution in form of a three component Green's function ( Text S1 ) . The solution is driven by stochastic channel behavior described by a hybrid deterministic-stochastic algorithm . We apply the model to a variety of experiments to demonstrate its potential .
Our modelling concept simulates active molecules individually by Markov chains , the concentration dynamics in the range of the molecule locally quasi-statically and the diffusional long range coupling by Green's functions . Simulations are orders of magnitude faster than numerical schemes based on spatial grids . Their efficiency derives from the methods which we apply . The use of hybrid deterministic-stochastic algorithms for the Markov chains allows for time steps much larger than traditional Gillespie algorithms . In between stochastic molecule state transitions , we integrate the concentration dynamics . The local quasi-static approximation reduces clusters to spatial -function sources which turns integrals into sums . It also substantially reduces the number of modes to be used in the Green's function . And finally Green's function enables us to restrict the calculation of concentration values to the locations of active molecules . dynamics and spatial channel clustering lead to the hierarchical system depicted in Figure 1 . channels are tetrameres [38] . A single channel opens and closes in dependence on binding and dissociation of and to the binding sites of its subunits ( see below ) . An open channel conducts a current from the ER into the cytosol which is due to the huge concentration difference of up to 4 orders of magnitude across the ER membrane . form clusters on the membrane of the ER consisting of 1 to 10 channels [33] , [35] . They physically interact within a cluster and are consequently separated by a few nanometers only [35] . The in a cluster are strongly coupled by the large local concentration close to open channels . Typical inter-cluster distances found experimentally are in the range of 1–7 [39] . Figure 1A shows a representative example of cluster arrangement used in simulations . Due to cytosolic buffers and SERCAs , the local concentrations close to an open channel cluster exhibit large gradients such that coupling between clusters is weak compared to intra-cluster coupling . This leads to the hierarchical organization of signals . Stochastic opening of a single channel ( blip ) is locally amplified by CICR leading to a puff ( Figure 1B and D ) . The concentration gradients keep the probability for activation of adjacent clusters small and only a fraction of puffs activates several neighboring clusters . Once a supercritical number of open clusters is reached , more of them open forming a global signal . In that way , the triggering random opening of a single is carried up to the macroscopic scale . The mechanism transforms the fast noise of channel state changes on a millisecond time scale into fluctuations of interspike intervals of tens of seconds as shown in Figure 1D . An early and widely used channel state model is the DeYoung-Keizer model [40] , [41] . It assumes independent subunit dynamics and allocates three binding sites to each subunit as shown in Figure 1C . One site for and one for that cooperatively activate the subunit . Another binding site with lower affinity for inhibits the subunit dominantly . These two different affinities lead to a biphasic dependence of the stationary open probability on the concentration ( see Figure 1 in Text S1 ) . Only the state out of the 8 possible subunit states corresponds to an active subunit ( Figure 1C ) , where the first index refers to the binding and is 1 , if is bound and 0 otherwise . Analogously , the second and third index describe binding to the activating and inhibiting site , respectively . A channel opens , if at least 3 subunits are in the active state . The 12 possible transitions between the 8 subunit states correspond to transitions in a state scheme forming a cube ( Figure 1C ) . Some of the transition probabilities depend on the local and concentrations ( Figure 1 in Text S1 ) . In simulations , the transitions are realized by a hybrid deterministic-stochastic algorithm [42] , which uses the local concentrations and the dissociation rates and binding rate constants given in Table 1 in Text S1 . Since within one cluster are close to each other , a cluster can be approximated by one spatial -source for the purpose of simulating the cluster current in the long range cellular dynamics . The current depends on the number of open channels , the time course of which comes out of the stochastic simulation of channel states . It is proportional to the concentration difference across the ER membrane at the location of the channel molecule . Hence , we actually need to solve the complete reaction-diffusion problem to determine it . But the concentration difference at the cluster is not well defined with a -source term . Therefore , we calculate the cluster current using a spatially extended cluster with radius as described in detail in Ref . [43] . The solution of that problem converges within fractions of a millisecond to its stationary state in the range of the channel molecule [43] . That part of the solution is all we need to calculate the current of the th cluster . Using the stationary concentration profiles we obtain: ( 1 ) with where denotes the channel flux constant . and are the diffusion coefficients of in the ER and the cytosol . The cluster radius depends on the number of open channels and the single channel radius . The advantage of the approximation is that it takes local ER depletion into account but only depends on the the spatially averaged concentrations and , which form the boundary conditions for the local quasi-static approximation ( see [43] for details ) . If channel distances within a cluster are of the order of magnitude of the diffusion length of free , the internal cluster geometry becomes relevant . In that case , several -functions can be used for one cluster . The approximation allows as well for determination of the local concentration at an open channel cluster resulting from its own current ( 1 ) as ( 2 ) the validity of which had been shown for the buffer concentrations used here [43] . Note that the total concentration at a cluster is the sum of the concentration ( 2 ) and the concentrations induced by currents of other open channel clusters . After closing , the concentration is determined by the cellular concentration dynamics ( see below ) 10 nm apart from the release site . The modelling strategy for the cellular dynamics is based on the separation of two length scales . On the microscopic scale of channel clusters , we use a detailed and stochastic channel model to determine local currents . On the macroscopic scale of the cell , we use a linearized spatial bi-domain model , and Green's function to integrate it . The microscopic scale determines the currents representing the sources of the macroscopic scale . We implement ideas proposed in [43] and use the currents of Eq . ( 1 ) as the amplitudes of the spatial -functions representing the cluster source terms in Eqs . ( 3 ) . A similar approach was taken by Solovey et al . [44] . We circumvent the concentration divergence at -function sources by using Eq . ( 2 ) for the value of the local concentration at open clusters . Vice versa , the macroscopic scale affects the concentration values entering the transition rates of the microscopic state schemes . The ER is a tubular network spreading throughout the cell [45] . Diffusion in such a geometry can be described by diffusion in unrestricted space with a decreased diffusion coefficient [46] . Subsequently , we can superimpose the ER and the cytosol leading to a bi-domain model . Due to the quasi-static approximation ( Eq . 1 ) , we do not need to determine the spatially resolved concentration in the ER . Lumenal and cytosolic domains are coupled by a homogeneous leak flux through the ER membrane , re-uptake of the ER by SERCA pumps and by the stochastic channel currents . Within the cytosol we take free , one mobile buffer and one immobile buffer with the total concentrations and into account leading to the reaction diffusion equations ( 3a ) ( 3b ) ( 3c ) where we used buffer conservation and linear pump and leak fluxes with the flux constants and . is the stochastic channel cluster current of the th cluster with strength defined by Equation ( 1 ) . Scaling concentrations , space and time with typical values reveals the number of independent parameters . It entails the definitions of Table 2 . We linearize Eqs . ( 3 ) , since we would like to use Green's function to solve them . Our parameter values are in the range of the applicability of the linearization to the buffer dynamics as described by Smith et al . [47] for the stationary profiles . We additionally have linearized the pump dynamics . The linearization does not exhibit saturation , which is relevant for calcium concentrations above , with being the dissociation constant of the pump ( Figure 2 in Text S1 ) . These concentrations occur close to open clusters . In that area , the dynamics are dominated by the diffusion term and the channel term , which reduces the relative error due to the linearization of pump and buffer rates substantially . However , if precise knowledge of concentration values close to open channels or clusters is required , the complete non-linear reaction diffusion equations must be solved like e . g . in [42] . The scaled linear reaction diffusion system ( Text S1 ) describes the spatially resolved concentration dynamics by: ( 4a ) ( 4b ) ( 4c ) where the leak flux depends on the average lumenal concentration , only . All the reaction rate constants depend on the resting state concentration , and due to the linearization: , and . For simplicity we subsumed also and under . The cytosolic concentrations are determined by the 3-component Green's function with clusters localized at ( see also Figure 3 in Text S1 ) ( 5 ) with the Bessel function of the first kind and the Legendre polynomial , where is the angle between the source location and the point given by ( 6 ) The are determined by the boundary conditions at the plasma membrane ( see Text S1 ) . The three-component response functions and include the time integration over the source history , i . e . the time dependent channel flux strength , and take the buffer reactions as well as the coupling with the ER into account: ( 7a ) ( 7b ) with the dimensionless cell radius and the normalization factors and given in the Text S1 . The coupling between the cytosol and the ER by and as well as the reaction rates of with the two buffers determine the time constants of the response functions ( 0 ) , which are implicitly given by the roots of the determinant of the coupling matrix ( 8 ) The method allows for spatially resolved concentration dynamics as shown in Figure 2 and in the Video S1 by an iso-concentration surface of 2 . An initially opening cluster increases the open probability of adjacent clusters and release is spreading through the cell until inhibition stops release . For the global dynamics , the average concentrations are obtained by spatial integration of the analytical solution ( 9 ) as ( 9 ) where denotes the cell radius . The first component of describes the cytosolic average concentration . With this , the lumenal average concentration in dimensionless units is determined by ( 10 ) which takes into account the leak , pump and channel fluxes , and is the volume ratio of the cytosol and the ER . denotes the equilibrium average lumenal concentration at . The difference between the average cytosolic and lumenal concentration − determines the cluster current according to Eq . ( 1 ) ( see Text S1 ) . The two main approximations of our method are the local quasi-static approximation and the linearization of the passive bulk processes . These assumptions do not allow for a precise study of the intra-cluster concentration dynamics . That can be done with finite element methods like in ref . [42] . The structure of the Green's function solution enables an elegant parallel algorithm that we call the Green's cell . It is orders of magnitude faster than finite element methods and able to simulate long lasting whole cell dynamics in feasible computing time . In the Green's cell algorithm the actual concentration of each cluster is calculated with the Green's function and local quasi-static approximation in dependence on the source history of all clusters by a single process . The concentrations are sent to the master process which determines the corresponding state transition and reaction time by the hybrid algorithm and also calculates the average concentrations . The transition times are re-distributed to the cluster processes where they are used to update the concentrations . For further details see Figure 4 in Text S1 . Our recent experimental investigation started from the assumption of a random spike generation by wave nucleation followed by a deterministic refractory time . This prediction yields in a linear dependence of the standard deviation on the average period which was also experimentally confirmed [5] . Previous studies report a possible feedback of on PKC activity in glutamate stimulated rat astrocytes [48]–[50] . This may lead to a positive feedback on the level by activation of PLC . The measured relation between standard deviation and average of interspike intervals for spontaneous spiking has a slope equal to 1 [5] , demonstrating that spike generation is poissonian and the spike generation probability is constant on the time scale of ISI . Clearly , there is no feedback on that time scale . To show that the experimental findings are indeed consistent with our ideas of spike generation , we use our modelling tool to study how molecular noise of single channels can be translated into global signals and whether it is sufficient to cause the observed randomness of spike sequences . Figure 3A shows an example of single cell measurements , where the upper panel exhibits the fluorescent signal related to the cytosolic concentration and the lower panel the individual ISIs . It demonstrates the stochasticity of spiking , since variations in ISIs are in the range of their average . Simulations of a cell with 47 clusters each containing a random number of between 4 and 16 exhibit a behavior very similar to experiments showing that single channel noise can lead to time varying ISIs , since there are not any other sources of noise in the simulations ( Figure 3B and C ) . The simulated oscillations exhibit in accordance with experimental observations different flavors ranging from rare and irregular spiking to fast and more periodic spiking . The standard deviation depends linearly on the average period [5] . Recently we have shown that this linear dependence is not a self-evident relation [51] . In particular , it was found that self-sustained oscillatory systems exhibit a different relation than the one observed in spiking experiments . The dependence of on obtained here in simulations is shown in Figure 3D and exhibits a linear dependence with a slope of 1 which was found in experiments for spontaneous oscillations [5] , [52] . The offset of the regression line on the -axis of about 20 s is the deterministic recovery time . The different − data points in Figure 3D result from different combinations of the and base level concentrations , which are both parameters in the model . In vivo the concentration is related to the stimulation level by activation of Phospholipase C and production . The base level is determined by the leak and the pump flux through the ER membrane . In simulations , we adjust the leak flux according to and the pump strength . If both concentrations are rather high in the range of no spiking occurs since channels are activated as soon as they are in the excitable state ( Figure 5 in Text S1 ) . We observe fast and regular spiking ( Figure 3C , E and F and Figure 5 in Text S1 ) for intermediate concentrations . The ISIs have a close to the deterministic refractory time , since a new spike is initiated as soon as the recovery time has elapsed . Regular spiking corresponds to cells with small in Figure 3D . A further decrease in one of the concentrations increases and , in a way depending on the other concentration ( Figure 3B , E and F ) . If both concentrations are small , global spiking vanishes and the signal consists of uncorrelated blips . In the previous analysis of the dependence of oscillations on the concentrations , we have already seen that the modelling tool can generate a large spectrum of signals ranging from stochastic spiking to almost periodic oscillations . Here , we show that the model can produce all known -induced forms of signals in dependence on physiologic parameters . Figure 4 exhibits different experimental signal forms and the corresponding simulation results for a cell with 32 clusters . The variety of signals is achieved by varying cell parameters leading to distinct cell responses as shown by the behavior of open channels ( black ) and number of inhibited subunits ( magenta ) as well as by the resulting average concentration in the cytosol ( red ) and in the ER ( blue ) . Fast and rather regular oscillations occur by the interplay of activation and inhibition leading to array enhanced coherence resonance as was hypothesized before [5] . This can be seen in the behavior of the channel state dynamics . The number of inhibited subunits ( magenta ) increases dramatically during a spike and finally inhibition terminates it ( Figure 4A ) . In the following the number of inhibited subunits relaxes slowly towards its resting level . Only very few channels open directly after a spike and these openings do not initiate a new spike , since the number of inhibited subunits is still to high ( higher than approximately 220 ) . That causes the deterministic time also observed experimentally [5] , [52] . But a spike occurs very soon after the number of inhibited subunits has fallen below a critical range since the open probability is rather high with these parameter values . That keeps the stochastic part of the ISI small and spike sequences regular . Moreover , the amplitude of the spike of open channels seems to be smaller , if the spike is initiated at times where the number of inhibited subunits is still high . We find longer and more irregular ISIs for decreased and base level concentrations , since the probability of a channel opening is decreased . As a consequence , the cell relaxes to a resting state between spikes with only a few inhibited subunits ( Figure 4B ) . The spike amplitudes of both the number of open channels and of the average concentration are slightly increased compared to the regular spiking . SERCA pumps also shape signals . Recent studies have shown that different phenotypes of cloned cells with regard to signalling occur due to small variations in SERCA expression levels and activity of RyR [4] . Here , we find that a decreased SERCA activity leads to a burst like behavior ( Figure 4C ) , since is removed slower from the cytosol and thus can activate channels which have recovered early from inhibition or channels which have not been activated before . For even smaller SERCA activity , cells exhibit long lasting plateau signals often with superimposed oscillations ( Figure 4D ) . In these cases , released stays within the cytosol and reactivates again and again . Cooperativeness induced by inhibition leads to superimposed oscillations on the high level . The panels of Fig . 4 provide also an idea of cell variability within one cell type and even within one experiment . A direct consequence of the diffusion mediated signal mechanism is the dependence on the strength of spatial coupling by diffusion . That coupling strength can be modulated by exogenous buffers , since they reduce the diffusion length of free . We took advantage of this property of buffers to demonstrate the spatial character of oscillations [5] . Note that we used concentrations of buffers much smaller than usually applied in order to suppress any kind of signal . We measured spiking for several minutes to obtain reference values for ISIs , loaded additional buffer and continued measuring ( see Figure 5A ) . The individual ISIs ( blue crosses ) are increased and exhibit a larger variability after buffer loading . To understand the experimental observation in more detail , we use simulations to analyze the response to additional buffer . Analogously to the experiment , we simulate a fixed cellular arrangement with different mobile buffer concentrations . Figure 5B shows a representative example , where the red and the blue parts correspond to 25 and 250 EGTA , respectively . Like in the experiment , larger buffer concentration leads to less and more irregular spiking . In the part with the higher buffer concentration , we observe isolated events which do not lead to global waves since coupling of clusters is too weak . These local events are rare in the reference measurements , since a triggering event initiates a global wave very likely . From population simulations , where individual cells differ in their spatial arrangement of clusters , initial buffer and base level concentrations , we obtain the − relation shown in Figure 5C , where cells are shifted by an increase of 10 in the EGTA concentration . Similar to experiment [52] , cells exhibit individual increases of and with a slope of the shift close to 1 comparable with the population slopes for the two measuring periods . BAPTA and EGTA are common buffers to suppress signals and we have used both in experiments [5] . Cells responded much more sensitive to BAPTA than to EGTA . BAPTA has much larger binding and dissociation rate constants than EGTA ( Table 1 ) . A disadvantage of the experiment is that the buffer is loaded into a cell by its esterificated form and the total amount that has entered is unknown and difficult to control . Here , we use modelling to illuminate the influence of the different buffer kinetics and concentrations of EGTA and BAPTA . Figure 5D shows the dependence of for fixed cell parameters on the buffer concentration in magenta for EGTA and in black for BAPTA , where squares denote simulations with a single channel current of 0 . 12 pA and the dots correspond to 1 . 2 pA . The larger current was achieved by an increased lumenal concentration . Cells only differ in the buffer type . We see that increasing BAPTA has a stronger effect than EGTA , which is mainly caused by the larger capture rate . Moreover , we observe a nonlinear dependence of on the buffer concentration . The nonlinearity explains the individual shifts of cells in the − plane shown Figure 5D . The comparison of the two different current strengths for BAPTA ( black ) indicates the role of spatial coupling . Higher currents lead to stronger coupling , and subsequently increasing buffer concentrations have a smaller effect on . From the buffer simulations , we can determine the − relation shown in Figure 5E . For the smaller currents , there is no qualitative difference between EGTA and BAPTA . Both exhibit a slope close to 1 as shown by the regression lines and an estimated deterministic time of 20 s . The simulations with higher cluster currents indicate a similar deterministic refractory period but the slope of the − relation decreases to approximately 0 . 6 . This might explain the experimentally found differences between cell types . Larger currents lead to stronger coupling on the macroscopic length scale and hence to smaller variations . To confirm these findings and to test the dependency of the slope on other parameters , we analyze spiking of cells for the two different single channel currents . In each simulation set the cells have identical properties and differ only with respect to the buffer content leading to the distinct and values in Figure 5E ( see also Section 6 in Text S1 ) . From these values we determine the population slopes . Figure 5F shows averaged over different spatial arrangements , concentrations ( stimulation levels ) and pump strengths ( see Figure 6 in Text S1 ) . Analogously , we investigated , and ( data not shown ) . The results are very similar to those with . For smaller single channel current we obtain always a slope close to 1 when varying all 4 cell properties and for the larger current a slope to 0 . 6 . Varying the buffer concentration , spatial arrangement of clusters , concentration or pump strength ( within certain limits ) does not change the − relation but only the position of the system on it .
Both the experiments and simulations show a simple linear relation between the standard deviation of ISI and the average ISI . The existence of this linear relation turned out to be surprisingly robust . It survives even an increase of the single channel current by an order of magnitude . This relation describes for each individual cell the response to stimulation changes . Cells shift the spike pattern from slow and irregular to fast and regular along the − relation when we increase stimulation . That supplements the current ideas on frequency encoding [54] , [56] . At the same time , the − relation describes the outcome of spiking experiments with a group of cells . In the experiments , we subjected a sample of cells to the same protocol , and we obtained as many different responses as there are cells in the sample [5] . That set of responses is not arbitrarily scattered across the − -plane but aligns along the − relation . All the variability among individual cells with respect to expression levels of pathway components , cell volume , ER volume , shape , ion concentration , etc . does not lead to severe deviations from this − relation . spiking is robust against variability of many pathway components in the sense that the − relation is robust . We learn from the simulations here , that it is rather the stochastic spike generation mechanism than control and regulation which provides for this robustness . If we call the − relation from a single cell obtained by parameter changes individual relation and that obtained from a sample of cells population relation , we can describe our findings as identity of individual and population relation . We could reproduce the variability within a population of cells in simulations by varying cluster array geometry , pump strength , stimulation or buffering conditions . Changing these parameter values simply shifted the system on the − relation and did not modify the relation . But changing the single channel current by one order of magnitude did change the slope of the − relation . That suggests a mathematical definition of robustness which accounts for the fact that cells should be able to execute certain functions ( e . g . to spike with a range of ISI ) , but not necessarily at the same strength of stimulation or normalized values of other parameters . We denote with and two variables describing the function ( e . g . and ) , and with , … , and , … , two sets of parameters ( e . g . stimulation strength , temperature , cell volume ) . The relation between and is robust with respect to value changes of parameters , if it has the structure . The parameters change only the value of while the control also the properties of , i . e . the properties of the pathway . We call this robustness of the function functional robustness ( in difference to the robustness of the value of ) . If we identify the stimulation strength with , all cells distinguished by the values of only can realize frequency encoding with the same − relation by varying . They can realize this function also by varying another -parameter or several of them: function and functional robustness are closely related . The statement on robustness can also be interpreted with respect to identity of pathways converging onto spiking . signals can be caused by many different stimuli . The pathways upstream from responding to the stimuli must differ with respect to their value of the , in order to be distinguishable by pathways downstream from . In summary , cells realize frequency encoding - the function of spiking - by mainly moving up and down the relation between standard deviation and average of ISI and to some degree by modulating the deterministic part of the ISI [52] . The − relation exists for a stochastic process only , since holds for deterministic systems . The − relation turned out to be functionally robust with respect to changes of values of one set of parameters . That set may describe cell variability within one cell type or pathway . Changing substantially another set of parameters modified the − relation . That set appears rather to specify the identity of pathways converging on spiking . Our model predicts that close proximity of clusters is a prerequisite for a spontaneous response to spread throughout a cell . Indeed , there are types of astrocytes in which responses spread within the cell and those , such as Bergmann glia where this is not observed . Interestingly local , subcellular spontaneous responses have been recorded which represent functional microdomains [57] . Complementary to the functional units , morphological units are described which are separated from each other by fine processes [58] . It is well conceivable that these thin processes separate endoplasmic reticulum between microdomains by more than 2 and according to our model this separation would prevent the spread of a local signal to other parts of the cell . In contrast , in cultured astrocytes , the endoplasmic reticulum is preferentially arranged around the cell center without apparent discontinuity [59] and these cells frequently exhibit spontaneous responses . In situ , spontaneous responses are reported for hippocampal astrocytes and these astrocytes are less polarized as compared to Bergmann glial cells and we would predict that they are less compartimentalized . Indeed , morphological studies indicate that hippocampal astrocytes have five to ten main processes from which smaller extensions branch off [60] . The synchronized activity obviously can spread within the volume of the main processes and soma of hippocampal astrocytes . Moreover , in contrast to culture , the endoplasmic reticulum in astrocytes in hippocampus tissue is preferentially located close to the plasma membrane [59] . These different morphological arrangements result in distinct patterns of responses and as a consequence in different gene expression patterns [53] . The rise of cell imaging during the last decades illustrated the spatial structure of cells and protein localization . Obviously , cells are not homogeneous and active molecules coupled by diffusional transport are very common . Concentration gradients are functionally relevant , if they create microdomains inside which a pathway is in a state different from its state at other locations in the cell . They have been shown to exist for ‘the other’ fast diffusing intracellular messenger cAMP and in phosphorylation/dephosphorylation dynamics . Hence , the need for spatially resolved cell models exists and we can apply the modelling concept , if all essential non-linearities are in the discrete active molecules or the boundary conditions and we can linearize remaining bulk reactions . The excellent validity of the linearization for the buffer reactions of dynamics has been shown by Smith et al . [47] . We expect a degradation reaction like the cAMP degradation by PDEs also to be linearizable in good approximation . If local concentrations at active molecules should be outside the range of validity of the linearization , that can be fixed by the choice of the local quasi-static approximation of the diffusion process there in many cases . The non-linearities of cAMP production by membrane-bound adenylyl cyclase can be formulated as boundary condition and Green's function must then be used iteratively with an update of the boundary condition in each time step . These remarks illustrate that there is flexibility in the choice of reactions to be linearized which crucially expands the applicability of the concept .
Astrocyte cell cultures were prepared from cortex of newborn NMRI mice [61] . Briefly , brain tissue was freed from blood vessels and meninges , trypsinised and gently triturated with a fire-polished pipette in the presence of 0 . 05% DNAase ( Worthington Biochem . Corp . , Freehold , NY , USA ) . Cells were washed twice and plated directly on poly-L-lysine ( PLL; 100 ; Sigma , Deisenhofen , Germany ) coated glass coverslips ( ) at densities of 3 to cells/coverslip , kept in -10-cm-dishes using Dulbecco's modified Eagle's medium ( DMEM ) supplemented with 10% fetal calf serum ( FCS ) , 2 mM L-glutamine , 100 units/ml penicillin , and 100 streptomycin . One day later , cultures were washed twice with Hank's balanced salt solution ( HBSS ) . Cells were maintained for at least 4 days and after reaching a subconfluent state , microglial cells and oligodendrocytes as well as their early precursors were dislodged by manual shaking and removed by washing with HBSS . The purity of the astrocytes was routinely determined by immunofluorescence using an antibody against glial fibrillary acidic protein ( GFAP , Sigma ) , a specific astrocytic marker . The cultures typically exhibited more than 90% cells positive for GFAP . Cultured cells plated on glass coverslips were measured between p4 and p6 . Cells were incubated with the indicator dye Fluo-4-acetoxymethyl-ester ( Fluo-4 AM , 5 , Molecular Probes , Eugene , USA ) for 30 min at room temperature in HEPES buffer ( 148 . 9 mM NaCl , 5 . 4 mM KCl , 1 mM , 10 mM , 10 mM HEPES , 5 mM D-glucose , pH 7 . 4 ) containing 0 . 01% Pluronic-127 ( Molecular Probes ) . Subsequently cells were washed and kept in HEPES buffer for 15–20 min prior to the measurements with the conventional imaging system at a frequency of 0 . 33 Hz . Cultures were fixed within the microscope chamber of an upright microscope ( Axioskop FS , Zeiss , Oberkochen , Germany ) equipped with a 20× water immersion objective ( UMPlanFl , numeric aperture: 0 . 5 , Olympus , Hamburg , Germany ) by a U-shaped platinum wire and superfused with HEPES buffer at 20 . Substances were applied by changing the perfusate . Cells were illuminated ( 495 nm ) from a monochromator ( T . I . L . L . Photonics ) and fluorescent images ( 515–545 nm ) collected every 3 s with a 12 bit camera ( SensiCam ) on an upright microscope . At this state , no intercellular waves were observed . Single cell time series were extracted from these images with ImagingCellsEasily software . | The number of proteins organizing cellular processes is huge . The challenge for systems biology is to connect the properties of all these proteins to cellular behavior . Do individual state changes of molecules matter for cell behavior despite these large numbers ? Recently , we have experimentally shown for four cell types that intracellular Ca2+ signalling is driven by single channel dynamics . Molecular fluctuations are used constructively for a stochastic spike generation mechanism . The hierarchical structure of Ca2+ signalling prevents averaging of fluctuations and , consequently , the sequence of global spikes still reflects this molecular noise . Here we present a stochastic 3-D multiscale modelling tool living up to these findings by following the consequences of individual channel state changes up to cell level . We simulate the variety of cell responses in different experiments . The stochastic spike generation mechanism is surprisingly robust , providing new insights into the relation of function and robustness . The modelling concept can be applied to a large class of reaction-diffusion processes including other pathways like cAMP . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"physics/interdisciplinary",
"physics",
"cell",
"biology/cell",
"signaling",
"computational",
"biology/signaling",
"networks",
"computational",
"biology/molecular",
"dynamics",
"computational",
"biology",
"computational",
"biology/systems",
"biology"
] | 2010 | Calcium Signals Driven by Single Channel Noise |
The definitive method for diagnosis of porcine cysticercosis is the detection of cysticerci at necropsy . Cysts are typically located in the striated muscle and brain . Until recently Taenia solium cysticerci have not been definitively identified in other tissue locations , despite several comprehensive investigations having been undertaken which included investigation of body organs other than muscle and brain . Recently a study conducted in Zambia reported 27% infection with T . solium in the liver of pigs with naturally acquired porcine cysticercosis , as well as some T . solium infection in the lungs and spleen of some animals . We investigated the cause of lesions in sites other than the muscle or brain in a total of 157 pigs from T . solium endemic regions of Uganda and Nepal which were subjected to extensive investigations at necropsy . Lesions which had the potential to be caused by T . solium were characterised by macroscopic and microscopic examination , histology as well as DNA characterisation by PCR-RFLP and sequencing . Lesions were confirmed as being caused by Taenia hydatigena ( both viable and non-viable ) , by T . asiatica and Echinococcus granulosus ( in Nepal ) and nematode infections . No T . solium-related lesions or cysticerci were identified in any tissue other than muscle and brain . It is recommended that future evaluations of porcine cysticercosis in aberrant tissue locations include DNA analyses that take appropriate care to avoid the possibility of contamination of tissue specimens with DNA from a different tissue location or a different animal . The use of appropriate control samples to confirm the absence of cross-sample contamination is also recommended .
Taenia solium is a cestode parasite which is an important cause of human morbidity and mortality , particularly in many developing countries where sanitary conditions and pig rearing practices favour the parasite’s transmission . In endemic countries , T . solium has been found to be associated with 29% of cases of epilepsy [1] . Cysticercosis caused by T . solium is recognised by the World Health Organization as a Neglected Tropical Disease [2] . T . solium is a zoonotic parasite with pigs acting almost exclusively as the intermediate host responsible for transmission . Completion of the parasite’s life cycle has been prevented in many developed countries through improvements in sanitation and hygienic pig rearing practices , which prevent the animals from being exposed to human faeces . Similar improvements in developing countries where T . solium is currently endemic would be expected to reduce the incidence of human cysticercosis . However , large areas of Africa , Asia and Latin America where T . solium is endemic remain economically disadvantaged , presenting a substantial hindrance to implementation of control programs . New tools have become available for prevention or treatment of porcine cysticercosis [3] , and the existence of these tools has encouraged an increasing number of efforts to evaluate different control strategies . A critical aspect of implementing a control program for cysticercosis is the evaluation of the program’s effectiveness [4] . The most common tools used for monitoring T . solium control have been the evaluation of changes in the prevalence of human taeniasis and/or porcine cysticercosis . Monitoring changes in taeniasis is hampered by the tapeworm often occurring at low prevalence in the population , even in many highly endemic areas , and the common co-endemicity of other species of Taenia which cause false positive reactions in some tests for taeniasis [4] . Evaluation of changes in porcine cysticercosis has been the most commonly used method for monitoring the impact of T . solium control measures . The relatively high prevalence of porcine cysticercosis in endemic areas , the short life span of pigs and their critical , direct role in transmission of the parasite , all favour evaluation of changes in the incidence of porcine cysticercosis as the most practical method for evaluating T . solium control efforts at project level . Serological methods for evaluation of the prevalence of porcine cysticercosis have been found to be highly non-specific [4] . For this reason , direct measures of infection are currently the only reliable method for diagnosis of porcine cysticercosis . Accurate diagnosis requires detailed post-mortem examinations involving the slicing of affected tissues to count cysts . In pigs , cysticerci occur most commonly in striated muscle tissues and brain . Mature , viable T . solium cysticerci are readily identified macroscopically and viability can be confirmed simply by demonstration of evagination of excised cysticerci [5] . In natural and experimental infections , some or all cysts may die in the tissues . Necrotic lesions may contain a detectable , although non-viable cysticercus , the presence of which confirms T . solium cysticercosis . Non-viable lesions caused by T . solium which contain no sign of the parasite may also occur and some of these can be diagnosed using specific detection methods for parasite DNA . Most studies of the tissue localization of T . solium in pigs have found the cysts to be restricted to the striated muscle tissue and nervous tissue . In rare circumstances of intense infection , cysts have been described also in the liver , lung and spleen [6] . Recently , Chembensofu et al . [7] found a high prevalence of T . solium cysticerci in naturally infected pigs in Zambia in tissues other than the striated muscle or nervous tissues . Chembensofu et al . found cysts in unusual tissue locations at a high frequency , even though other studies examined relatively large numbers of infected pigs and included very heavily infected animals and no cysts were found in unusual locations [8 , 9] . Chembensofu et al . confirmed T . solium diagnosis by DNA analyses using PCR-restriction fragment length polymorphism ( PCR-RFLP ) . A field trial of the use of TSOL18 vaccination and medication for control of T . solium transmission by pigs was undertaken recently in the Banke district of Nepal [10] . Among 69 pigs examined at the conclusion of that trial , as well as cysts in striated muscle and nervous tissue , lesions were detected also in the liver and lungs of numerous animals . Similarly , we investigated the occurrence of cystic lesions in free-ranging pigs from T . solium endemic areas in Bukedea and Kumi Districts of Uganda and identified a number of animals containing suspect lesions in the liver and lungs that could possibly have been caused by T . solium . Here we detail the results of investigations that were undertaken which sought to identify the cause of these lesions in the Nepalese and Ugandan pigs , paying particular attention to the evaluation of control samples so as to provide evidence for the absence of false positive results in PCR-RFLP due to sample contamination .
Sixty-nine pigs of slaughter age and weight underwent extensive post mortem analyses in Nepal while 98 pigs were necropsied in Uganda . Post mortem methods are detailed by Poudel et al . [10] . The viscera were removed and the tongue , masticatory muscles , brain , heart , liver , lungs , both kidneys and the full diaphragm were retained in numbered containers . The muscles from each side of the carcass were dissected from the bones . All the retained organs , and muscles of the right-hand side of the carcass , were sliced by hand at intervals of approximately 3 mm and examined meticulously for the presence of T . solium cysticerci or other lesions . When no cysticerci were detected in the tongue , masticatory muscles , diaphragm , brain or muscles from the right-hand side of the carcass , the muscles of the left-hand side of the carcass were also sliced . Cysticerci in the striated muscles and brain were recorded and characterised as viable or non-viable [10] . For lesions identified in organs other than the striated muscle or brain , particularly in the liver and lung , representative samples were taken for DNA analyses , histological analyses and examination under a dissection microscope ( Olympus model SF10 ) . Specimens for histology were placed in 10% buffered formalin . Samples for DNA analyses were taken with care to minimize the potential for cross contamination of the samples with DNA from any other source . Samples were excised with a new , sterile scalpel blade and placed into a >10x excess volume of RNAlater tissue storage reagent ( Sigma-Aldrich ) . Control tissue samples were taken from the same organ using a different , new scalpel blade , from sites adjacent to the place where the lesion had been sampled but which contained normal tissue only . The same procedure was used for lesions and control tissue samples . They were washed with DNA extraction buffer ( 50 mM Tris pH 8 . 0 , 50 mM EDTA , 100 mM NaCl , 0 . 5% SDS ) to remove RNAlater storage buffer . DNA was extracted by digestion with Proteinase K ( 0 . 2 mg/ml , Promega ) in DNA extraction buffer at 56°C for at least 16 hours or until proteinaceous material was completely dissolved . The DNA was purified by extraction with an equal volume mixture of phenol/chloroform and centrifugation at 18 , 000g for 15 min . The phenol/chloroform extraction was repeated in a new microfuge tube and residual phenol was removed by chloroform extraction and centrifugation . The DNA was precipitated by the addition of two volumes of absolute ethanol , incubation at 4°C for at least 1 h and centrifugation at 18 , 000g . DNA pellets were washed with 70% ethanol , centrifuged , dried after removal of ethanol and dissolved in sterile deionized water . The concentration of DNA in each sample was measured using a spectrophotometer ( NanoDrop ND-1000 , Thermo Scientific ) and stored at -20°C . For use in control PCR reactions , identical methods were used to isolate DNA from specimens of a variety of parasite species stored at -80°C . DNA isolated from pig tissue lesions was amplified using primers described by Poon et al [11] as pan-nematode , having been designed so as to amplify DNA from a variety of nematode parasites . These primers targeted a 166 bp fragment of the COX1 gene: pan_nematode_cox1_692F 5′-TGTCTTTACCWGTTTTRGCTGG-3′ and pan_nematode_cox1_835R 5′-CCGAAAGCAGGYAAAATHARAA-3′ . PCR conditions were the same as those described for the cestode PCR detailed below , except an annealing temperature of 50°C was used , with Ascaris suum and Ascaris lumbricoides DNA used as positive controls . The mitochondrial 12S ribosomal RNA gene was amplified by PCR using the following primers originally described by Geysen et al . [12]: TaenF , 5’ GTTTGCCACCTCGATGTTGACT 3’ and ITMTnR , 5’ CTCAATAATAATCGAGGGTGACGG 3’ . These primers were selected for PCR amplification due to the high conservation of DNA sequence at this locus , allowing species identification by targeting the 890 bp of the mitochondrial 12S ribosomal RNA gene . PCR amplification was based on the methods also described by Rodriguez-Hidalgo et al . [13] and Somers et al . [14] and was performed in a final volume of 25 μl with the following modifications: 1 μl template DNA ( 2 ng/ μl ) , 5 μl 5X SuperFi Buffer ( Invitrogen , including 7 . 5 mM MgCl2 ) , 0 . 5 μl ( 10 mM ) dNTP mix , 1 . 25 μl ( 10 μM ) of each forward and reverse primer , 0 . 25 μl ( 2 U/μl ) Platinum SuperFi DNA Polymerase ( Invitrogen ) , and 15 . 75 μl sterile de-ionized water . The PCR reactions were performed in a Bio-Rad T100 thermal cycler using the following conditions: initial denaturation ( 98°C , 30 s ) , followed by 35 cycles of denaturation ( 98°C , 10 s ) , annealing ( 63°C , 10 s ) and extension ( 72°C , 90 s ) , and a final extension step ( 72°C , 5 min ) . To screen tissue samples and differentiate between T . solium , Taenia hydatigena , Echinococcus granulosus and Taenia asiatica infection ( the most likely parasites to cause similar lesions to T . solium ) , PCR amplified products were digested using restriction enzymes to obtain fragments of the mitochondrial 12S ribosomal RNA gene . These fragments were separated by agarose gel electrophoresis to reveal distinct restriction fragment patterns for each taeniid species . The location of predicted restriction sites for the 12S ribosomal RNA gene for each of the taeniid species are shown in Fig 1 . Double digests of PCR products contained DdeI and HinfI in CutSmart buffer ( New England Biolabs ) in the same reaction tube , according to the manufacturer’s instructions and were prepared as also described by Somers et al . [14] . Single digests of PCR amplicons contained HpaI ( New England Biolabs ) in CutSmart buffer and were prepared as also described by Devleesschauwer et al . [15] . Restriction digested PCR products were separated by agarose gel electrophoresis ( 3% agarose in 45mM Tris-Borate , 1mM EDTA , pH 8 . 3 ) . DNA restriction fragments separated by electrophoresis were stained with SYBR Green ( Invitrogen ) and visualized using a Safe Imager transilluminator ( Invitrogen ) . The PCR products were separated by electrophoresis in 1 . 2% agarose ( 50mM Tris , 20mM sodium acetate , 2mM EDTA , pH 8 . 3 ) , excised from the gel , purified using the Minelute purification kit ( Qiagen ) according to the manufacturer’s recommendations and quantified using a NanoDrop spectrophotometer . DNA sequencing was performed by Micromon ( Melbourne , Australia ) using the same primers as were used in the PCR reactions and a BigDye Terminator Cycle Sequencing kit ( Applied Biosystems ) . PCR products were sequenced in both directions , assembled and analysed using Geneious 11 . 1 ( Biomatters , Auckland , New Zealand ) . BLAST was used to compare the PCR-derived sequences with reference sequences of taeniid mitochondrial genomes in the GenBank database . Lesion specimens were recovered from pig tissues , trimmed to remove excess tissue and stored in 10% buffered formal saline . Paraffin embedded 5μm tissue sections were prepared and stained with haematoxylin and eosin .
Lesions that could potentially be caused by T . solium were identified in the livers , lungs and kidneys , as well as in the tissue locations more commonly associated with T . solium ( striated muscle and brain ) . Lesions were detected in the liver in 33% of the pigs from Uganda and 26% of the pigs from Nepal . Lesions were identified frequently in the lungs of animals in Nepal ( 35% ) but less commonly in the animals from Uganda . Individual animals from both sites were also identified with lesions in the kidney and spleen . The lesions varied greatly in their macroscopic appearance ( Figs 2 and 3 ) and included small ( 1-2mm ) solid spots on the surface of the liver , with similar spots in the parenchyma of the liver or in the lung tissue , and larger , sometimes vesicular-looking lesions , in the liver and lung . Sometimes single lesions were present while at times there were several similar lesions or , occasionally , large numbers of similar lesions . The cause of the majority of lesions in the liver , lungs , spleen and kidney could not be determined by their macroscopic characteristics ( Fig 2A , 2B and 2C ) . Lesions identified in the striated muscle were generally typical of viable or non-viable cysticerci of T . solium ( Fig 2D and 2E respectively ) . Lesions that contained what was clearly identifiable as a cysticercus were also identified in the liver of some animals ( Fig 2F , Fig 3A ) . In Nepal , but not Uganda , cysts were identified in both the liver and lungs that were clearly caused by E . granulosus , including cysts which contained germinal membrane , brood capsules and protoscolesces . In both Nepal and Uganda , mature T . hydatigena cysts were found in the liver , and were identified macroscopically . On histological examination , lesions were often found to be foci of lymphoid hyperplasia due to antigenic stimulation , the cause of which was not apparent . Some had the appearance of migratory nematode tracts , with central necrosis surrounded by eosinophilic inflammation , although no definitive cause could be found . Some lesions in the lungs were found to have a filamentous inclusion within a solid mass ( Fig 4A and 4B ) . On histological examination , a degenerate nematode was evident in cross-section with visible cuticle , musculature and uteri with embryos ( Fig 4C ) and/or containing embryonated nematode eggs approximately 50μm in size , consistent with Metastrongylus spp . DNA was purified from representative examples of lesions found in pigs , particularly those found in tissues other than the striated muscle or brain , the origin of which could not be determined macroscopically . The samples were processed for PCR amplification using pan-nematode oligonucleotide primers however no amplification products were found , whereas amplification of control DNA from Ascaris spp . amplified an expected product of 166bp . PCR amplification of the mitochondrial 12S ribosomal RNA gene using the TaenF and ITMTnR primers on control DNA samples resulted in PCR products of 890–900 bp for T . solium , T . hydatigena , T . asiatica , T . saginata and E . granulosus . For those DNA samples from tissue lesions which did amplify a product with the TaenF and ITMTnR primers , the products were within a comparable size range . Delineation of the 12S ribosomal RNA amplified products to species level was undertaken using RFLP ( Fig 5 ) and DNA sequencing . Many of the Nepalese cystic lesions processed in RFLP following DdeI-HinfI digestion of the 12S amplicon produced a similar restriction digestion profile , consisting of four major DNA fragments ( 330 bp , 250 bp , 125 bp , 20–35 bp ) . This restriction profile was identical to the pattern obtained using control DNA purified from E . granulosus ( Fig 5 , lane 25 ) and was consistent with the predicted location of restriction sites ( Fig 1 ) . DNA sequencing of the 12S amplification product from putative E . granulosus cysts , identified in the Nepalese pig samples by PCR-RFLP , confirmed the parasite species as being E . granulosus . DdeI-HinfI digestion of the PCR-RFLP DNA products obtained using DNA from lesions in the livers of two of the Nepalese pigs ( Fig 3C and 3D; Fig 5 , lanes 2 and 21 ) differed from the result obtained for the majority of lesions which were confirmed as being E . granulosus ( Fig 5 , lanes 3–20 ) . These two PCR-RFLP patterns also differed from those obtained for the T . solium , T . hydatigena and T . saginata controls ( Fig 5 , lanes 24 , 26 , 27 , respectively ) , but were identical to the pattern seen using control DNA from T . asiatica ( Fig 6 ) , indicating that these liver lesions were caused by T . asiatica . DNA sequencing of the 12S PCR product and BLAST comparisons confirmed that these two pigs had lesions caused by T . asiatica . DNA was purified from representative examples of lesions found in pig tissues other than the striated muscle or brain , the cause of which could not be determined macroscopically . The samples were processed for PCR amplification using pan-nematode oligonucleotide primers , however no amplification products were found , whereas amplification of control DNA from Ascaris spp amplified an expected product of 166 bp . For those DNA samples from tissue lesions which did produce a PCR amplification product of the mitochondrial 12S ribosomal RNA gene using the TaenF and ITMTnR primers , the products were within a comparable size range to those seen using control cestode DNA samples detailed above . RFLP analysis of the PCR product digested with DdeI-HinfI , followed by DNA sequencing in some cases , allowed the determination of the parasite species responsible for those lesions which did generate a 12S ribosomal RNA gene product in PCR . Samples from the livers of two pigs ( Fig 3A and 3B ) were unable to be clearly differentiated from being either T . solium or T . hydatigena following digestion with DdeI-HinfI in PCR-RFLP , since their restriction patterns ( Fig 7 , lanes 2 and 3 ) , consisting of three DNA fragments ( 550 bp , 290 bp and 20–35 bp ) , were similar to both the T . solium and T . hydatigena controls ( Fig 7 , lanes 4 and 5 respectively ) . The restriction pattern of the two liver lesions obtained using DdeI-HinfI digestion , also differed significantly from the profile obtained for control DNA from E . granulosus and T . saginata ( Fig 7 , lane 6 and 7 ) . Digestion of the 12S PCR products derived from these two lesions using HpaI ( Fig 7 , lanes 9 and 10 ) produced profiles consisting of two restriction fragments ( 670 bp , 220 bp ) of the 890 bp PCR product , identical to the T . hydatigena control ( Fig 7 , lane 12 ) . These patterns differed from those obtained with the T . solium control DNA ( Fig 7 , lanes 4 and 11 ) and from DNA obtained from a viable T . solium cyst from muscle ( Fig 2D; Fig 7 , lanes 1 and 8 ) . DNA sequencing of the 12S PCR products from the two pig liver lesions shown in Fig 3A and 3B and BLAST comparisons confirmed they were caused by T . hydatigena .
In pigs from Nepal and Uganda , lesions were detected in body organs which are not normally sites of infection with T . solium , but which could potentially have been caused by T . solium . Although viable cysticerci were found in the liver of some animals , none were found to be cysticerci of T . solium . A wide variety of other lesions were also identified in various tissues , particularly in the liver and lungs . Approximately a third of the animals were found to have lesions in the liver which could , potentially , have been caused by T . solium , and , in Nepal , a quarter of the animals had lesions in the lungs which could have been confused with non-viable lesions caused by T . solium . The characteristics of the lesions varied greatly , both in their number in individual animals and in their macroscopic characteristics ( Figs 2 and 3 ) . The cause of many of the lesions was unable to be determined by macroscopic or microscopic observation or by histological or DNA investigations . However , the cause of some lesions was able to be determined definitively ( Fig 2D–2F; Fig 3A–3F ) . Lesions were identified as having been caused by nematode infections , in both Uganda and Nepal by T . hydatigena , and in Nepal by T . asiatica and E . granulosus . No lesions in sites other than the striated muscles and brain were identified as having been caused by T . solium . Several studies have detailed the tissue localization of T . solium cysticerci in naturally infected pigs which found the cysts to be restricted to the striated muscles and brain , as we did in this investigation . Boa et al . [8] undertook an extensive study involving detailed dissection of all the striated muscle tissue , brain and many body organs of 24 pigs from the Mbulu District of Tanzania with naturally-acquired T . solium infections , including 10 animals harbouring more than 10 , 000 cysts and one animal with >80 , 000 cysts . No T . solium cysts were found in any of the animals in the spleen , kidneys , lungs or liver . Phiri et al . [16] examined the tissue distribution of T . solium cysts in 31 naturally infected pigs from the southern and eastern regions of Zambia including the thorough slicing of body organs and found no cysts in the spleen , kidneys , lungs or liver . Similarly , Singh et al . [9] undertook thorough slicing of body organs of 19 naturally infected pigs from the Punjab region of India and found no cysts in the spleen , kidneys , lungs or liver . Onah and Chiejina [17] examined 483 naturally infected pigs from Enugu State in Nigeria by extended meat inspection , although they did not undertake tissue slicing . Among the 483 infected animals examined , 403 animals were found to have what was referred to as generalized infections , most with ‘vast numbers’ of cysticerci . They comment that “despite the heavy infections encountered , no cysts were found in the kidneys , liver or lungs of any pigs” . Infections with T . solium cysticerci have been described in organs other than the striated muscles or brain in animals experimentally infected with T . solium from China [18–20] . Small numbers of T . solium cysts have been reported in a small number of naturally infected young piglets from Mexico [21] and two cysticerci were identified in the liver of mature pigs from Nepal [22] , however neither study included clear evidence to differentiate these as being T . solium rather than either T . hydatigena or T . asiatica . In his treatise on meat inspection , Ostertag [6] comments that while T . solium cysts are normally found in the muscles , they may at times be seen in the lymph nodes and subcutaneous fat , while in an even more extensive infection the parasites may be present also in the liver and the lungs . In addition to these poorly substantiated reports of the occasional occurrence of T . solium cysts occurring in pigs in tissue locations other than the striated muscles and brain , one report stands out as the only instance where cysts were commonly found in naturally infected pigs in other body organs . Chembensofu et al . [7] undertook a detailed study of the tissue distribution of T . solium cysts in 37 naturally infected pigs from Katete and Sinda Districts in the Eastern Province of Zambia . Ten animals ( 27% ) were diagnosed as having cysts in the liver , one with a cyst in the spleen and two with cysts in the lungs . At least one cyst from every organ found to be infected was confirmed as being T . solium by PCR-RFLP , which would appear to provide definitive evidence that they were indeed T . solium . Chembensofu et al . [7] did not provide supportive information such as figures to aid in the description of the characteristics of cysts found in unusual locations , or the PCR-RFLP results . No information was given about what controls were used in the PCR-RFLP analyses , especially controls that would indicate contamination was unlikely to have been possible at the time the specimens were collected . To the best of our knowledge , we used the same PCR-RFLP methods as those used by Chembensofu et al . [7] , although detailed methods were not provided directly by Chembensofu and colleagues , nor were they included in the cited references . When tissues of a pig that has a heavy infection with T . solium cysts are being sliced by hand , the instruments and cutting boards etc become extensively contaminated with whole cysticerci , bladder fluid , cyst walls and scolesces . Meticulous care would be required to ensure that there was no possibility that specimens collected from one animal could not be contaminated with DNA from a previously dissected animal or from cysts from other tissues of the same animal . In our study , particular care was taken in this regard . Specimens for DNA analyses also included tissue samples from the same organ adjacent to the area where a lesion was excised but containing no lesion . None of these control tissue samples amplified a PCR product indicative of the presence of DNA from a cestode parasite . We attempted to use “pan-nematode” primers [11] in PCR in order to screen DNA from lesions found in the liver or lungs of pigs that had the potential to be interpreted as being caused by T . solium , however none of the samples generated a PCR product . Clear evidence was obtained from lung lesions as being associated with nematode parasites ( Fig 4 ) , likely to be Metastrongylus spp . While the PCR primers developed by Poon et al . [11] are referred to by them and us as pan-nematode , the primers were not designed with the COX1 sequence from Metastrongylus spp . Our subsequent characterisation of the cox1 gene sequences corresponding to the primer locations in Metastrongylus spp found substantial differences in the downstream primer sequence in Metastrongylus , providing an explanation for the lack of amplification of a PCR product in those instances . We found that PCR-RFLP using primers derived from the mitochondrial 12S ribosomal RNA gene was a valuable tool for species identification of Taenia and Echinococcus and for pre-screening many samples prior to DNA sequencing . Our use of “pan-cestode” PCR primers [23] annealing to a conserved region of the 12S ribosomal RNA gene , allowed the amplification of a fragment of the gene from T . solium , T . hydatigena , T . asiatica , T . saginata and E . granulosus . For those samples where a PCR product was amplified , the methodology used definitively differentiated suspect lesions as being caused by a particular species of Taenia or Echinococcus . Ooi et al . [24] list eight helminth parasites that produce lesions in the liver of pigs other than T . solium . These include Schistosoma japonicum , Taenia hydatigena , T . asiatica , E . granulosus , Echinococcus multilocularis , Ascaris suum , Toxocara canis and Stephanurus dentatus . We confirmed the presence of viable and non-viable lesions caused by T . hydatigena ( eg . Fig 3A and 3B ) among the liver specimens examined here , and also two animals from Nepal with T . asiatica infection ( Fig 3C and 3D; Fig 6B ) . The two animals infected with T . asiatica were derived from different village locations in the Banke District , one from Khatikanpurwa and one from Mahapurwa . A previous study reported T . asiatica infecting pigs in Nepal [25] . Our investigations to determine whether the T . asiatica lesions had genetic features consistent with being T . asiatica or being hybrids of T . asiatica and T . saginata [26] were inconclusive using criteria according to Sato et al . [27] . Hepatic lesions were found in the livers of animals in both Uganda and Nepal which contained a viable cysticercus . In most cases the cysticercus had characteristics typical of T . hydatigena ( Fig 2F ) . However , one instance occurred when a cysticercus was found that was of a size more typical of T . solium ( Fig 3A ) . The scolex of this cysticercus contained two rows of hooks , the largest of which had a mean length of 170 μm , while the mean length of the small hooks was 113 μm . Speciation of this cysticercus based only on hook length was not possible since the hooks were within a size range for both T . hydatigena and T . solium [28] . PCR-RFLP and sequencing of DNA from this cyst confirmed it as being an immature T . hydatigena cysticercus . Among the 167 pigs which were examined in this study , six animals were identified which had only a single caseous or calcified lesion in the muscle tissue , the cause of which could not be definitively identified . As there are numerous potential causes of caseous or calcified lesions in muscle tissue other than T . solium , we applied the same diagnostic criteria as was applied by Sah et al . [22] and Poudel et al . [10] in not classifying these animals as T . solium infection . Animals found to have two or more caseous or calcified lesions in the muscle tissue were classified as being cases of T . solium infection since this is the most likely cause of multiple non-viable lesions in the muscle tissues of pigs reared under free-roaming conditions in countries that are endemic for T . solium . We did not find any tissue lesion , other than in the striated muscles and brain , in free-ranging pigs from Uganda and Nepal which could be proven to be caused by T . solium . These findings are consistent with those of all previous studies of the tissue localization of cysticerci in naturally infected pigs where meticulous dissections of mature animals has been undertaken , with the exception of the study recently published by Chembensofu et al . [7] . Their re-visiting of necropsy for detection of porcine cysticercosis in 38 infected pigs from Zambia found 27% of the animals to have T . solium cysticerci in the liver , as well as some animals having T . solium in the lungs and spleen . T . solium may occur sporadically in aberrant tissue locations , as has been mentioned anecdotally by others , for example by Osterag [6] . However , it is difficult to reconcile one study finding as many as 27% of pigs being infected with T . solium in the liver , when no animal was found to have T . solium in the liver among a larger total number of animals investigated by others [8 , 16; this study] , including pigs from the same Zambian Province [16] from which Chembensofu et al . [7] sourced some of their animals . Future studies involving comprehensive evaluation of porcine cysticercosis by necropsy , that may include investigation of organs such as the liver , lungs and spleen , should take appropriate measures before concluding that lesions found in aberrant locations were caused by T . solium . Any DNA analyses undertaken on lesions in aberrant tissue locations that are suspected to have been possibly caused by T . solium should be supported by evidence from appropriate control samples . Clearly enunciated steps should also be taken to minimize the possibility of contamination of necropsy specimens with T . solium DNA from a different animal or tissue location . | The only currently available method for accurate assessment of the prevalence of porcine cysticercosis is to undertake detailed examination of tissues at necropsy . Most investigations have found that in pigs , Taenia solium encysts in the striated muscles and nervous tissue . Recently an investigation of porcine cysticercosis carried out in Zambia described 27% of infected pigs as having T . solium cysts in the liver , as well some animals having cysts in other unusual tissues locations . We collected lesions from the tissues of pigs derived from T . solium endemic areas in Uganda and Nepal . Particular care was taken to avoid cross contamination of specimens with T . solium DNA from other tissue sites or from other animals . Control tissue samples were collected from the same organs as lesions were collected , but from sites where no lesion was present . The samples were assessed macroscopically , histologically and by PCR-RFLP analysis of DNA isolated from the lesions . Evidence was obtained for lesions in tissues other than the striated muscles and nervous tissue being caused by Taenia hydatigena , Taenia asiatica , Echinococcus granulosus and nematode parasites , however no evidence was found for the presence of any lesion in these tissues being caused by T . solium . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"medicine",
"and",
"health",
"sciences",
"liver",
"immune",
"physiology",
"pathology",
"and",
"laboratory",
"medicine",
"muscle",
"tissue",
"spleen",
"tropical",
"diseases",
"vertebrates",
"parasitic",
"diseases",
"animals",
"mammals",
"lungs",
"organisms",
"respiratory... | 2019 | Accurate diagnosis of lesions suspected of being caused by Taenia solium in body organs of pigs with naturally acquired porcine cysticercosis |
A standard view in neuroeconomics is that to make a choice , an agent first assigns subjective values to available options , and then compares them to select the best . In choice tasks , these cardinal values are typically inferred from the preference expressed by subjects between options presented in pairs . Alternatively , cardinal values can be directly elicited by asking subjects to place a cursor on an analog scale ( rating task ) or to exert a force on a power grip ( effort task ) . These tasks can vary in many respects: they can notably be more or less costly and consequential . Here , we compared the value functions elicited by choice , rating and effort tasks on options composed of two monetary amounts: one for the subject ( gain ) and one for a charity ( donation ) . Bayesian model selection showed that despite important differences between the three tasks , they all elicited a same value function , with similar weighting of gain and donation , but variable concavity . Moreover , value functions elicited by the different tasks could predict choices with equivalent accuracy . Our finding therefore suggests that comparable value functions can account for various motivated behaviors , beyond economic choice . Nevertheless , we report slight differences in the computational efficiency of parameter estimation that may guide the design of future studies .
Value ( or utility ) functions have been defined to account for preferences revealed in choice tasks [1] . One basic principle is that if an agent prefers A over B , then for this agent the value of A is higher than the value of B . Assuming basic axioms of expected utility theory , cardinal functions have been described , such that option values can be positioned on a numeric scale [2] . Cardinal values rely on the notion that choice probability depends on the distance between option values , as well as on their distance from a reference point [3] . Value functions can be parameterized when choice options are combinations of objective quantities , e . g . , the probability and magnitude of monetary payoff . The parameters can then be estimated through fitting procedures that maximize the likelihood of observed choices under the valuation model . Fitting choices involves specifying a function relating choice probability to option values , generally a softmax rule [4] . Thus , most studies have used choice data to infer functions that assign cardinal values to any possible option . Alternatively , a more direct approach has been used in the neuroeconomics literature , using behavioral tasks in which subjects assign cardinal values to available options , instead of inferring value functions from their choices . One possibility is to ask subjects to rate on analog scale the desirability ( or likeability ) of the outcomes associated to the different options [5] . Another possibility is to ask subjects to express the maximal cost ( e . g . price , effort or delay ) that they are willing to endure in order to obtain these outcomes [6 , 7] . The aim of the present study was to compare the value functions derived from these direct cardinal measures with the value functions derived from fitting choice data . We selected , in addition to a standard binary choice task where subjects state their preference between two options , a subjective rating task where subjects score the desirability of every possible outcome and an effort production task where the probability of obtaining the outcome depends on the force produced with a handgrip . Standard models of behavior in these tasks suggest that ratings and forces can be taken as direct measures of the subjective outcome values that drive choices ( see Methods ) . However , there are a priori reasons why the value functions elicited by the different tasks should differ in their form or in their parameters . In our perspective , the key difference between tasks is the nature of the cost . In choice tasks , the response entails an opportunity cost , corresponding to the value of the non-selected option [8] . The response is therefore based on the value difference between the two possible outcomes , which is often called decision value . As the motor response is generally similar for the two options , there is no need to consider action costs . In effort tasks , the response is associated with a specific cost due to energy expenditure , which may be signaled through muscular pain . The response therefore aims at maximizing the net value , i . e . the trade-off between outcome value and action cost [9] . In rating tasks , the variation in action cost across the possible positions on the scale is usually negligible , although the extremes may be longer to reach . Thus , the response should be a direct expression of outcome value . As decision values , net values and outcome values may be computed by different brain systems , they may follow different functions [10] . In addition , there is a cost that may be common to all behavioral tasks , which is social reprobation . Some responses may be more socially acceptable than others , particularly if moral considerations are involved [11] . This social cost may be more salient in rating tasks , which have no other consequences and can therefore be considered as ‘hypothetical’ decisions . By opposition , choice and effort tasks are typically consequential: they determine the outcome , either deterministically or probabilistically , and therefore involve ‘real’ decisions . Hypothetical and real decisions have been compared in a number of studies using various tasks [12–16] , with contrasted results and no proper model comparison . Yet it may seem intuitive that subjects in rating tasks are more likely to pretend having values they do not have , for reputation concerns , because there is no obvious costly consequence . To assess this potential difference between tasks we used options that combined money for the subject ( gain ) and money for a charity ( donation ) , with the aim of triggering moral dilemma . Also , each behavioral task may be susceptible to specific artifacts . For instance , the rating scale is somewhat arbitrary , and may yield distortions of value functions due to framing or anchoring phenomena [17] , particularly if subjects are not familiar with the range of values spanned in the set of options . Effort exertion , between zero and maximal force , may be less arbitrary but susceptible to fatigue , which may increase with the number of performed trials and influence effort cost , and hence the values expressed by participants [18] . In the present study , we compared the value functions elicited by the different tasks for a same set of composite outcomes , each combining gain and donation . We found that the same valuation model provide the best fit of behavior in the three tasks , with slight differences in parameter estimates .
Subjects ( n = 19 ) participated in three tasks aimed at measuring subjective values of bi-dimensional outcomes composed of one gain for themselves and one donation for a charity organization they selected prior to the experiment ( Fig 1 , top ) . In the rating task , participants rated how much they would like to obtain the composite outcome using a scale graduated from 0 to 10 . The feedback was probabilistic and they obtained the outcome in 70% of the trials , irrespective of their ratings , which were therefore not consequential . The probabilistic contingency was adjusted so as to match that of the effort task . In the force task , subjects had to squeeze a handgrip knowing that the chance to win the outcome was determined by the ratio of the force they produced during the trial and their maximal force measured beforehand . Note that previous experiments in the lab using the grip task with similar range of incentives showed that subjects produce on average about 70% of their maximal force [19] . In the choice task , participants had to choose between two composite options , the selected outcome being obtained in 70% of trials . The choice task followed an adaptive design [20] in which options were proposed so as to optimize the parameterization of an a priori value function ( linear integration of gain and donation with their interaction ) . As expected , explicit ratings , forces produced and subjective values inferred from choices all increased with incentives , i . e . with both gain and donation ( Fig 1 , bottom ) . Before going into more sophisticated models , we conducted linear regressions ( for ratings and forces ) or logistic regression ( for choices ) against the two main factors ( gain G and donation D ) and their interaction . Regression estimates obtained for main factors were significantly different from zero in all cases: in the rating task ( βR ( G ) = 0 . 07±6 . 10−3 , t ( 18 ) = 11 . 5 , p = 1 . 10−9; βR ( D ) = 0 . 06±7 . 10−3 , t ( 18 ) = 8 . 2 , p = 1 . 10−7 ) , in the force task ( βF ( G ) = 0 . 05±6 . 10−3 , t ( 18 ) = 8 . 5 , p = 1 . 10−7; βF ( G ) = 0 . 05±6 . 10−3 , t ( 18 ) = 7 . 2 , p = 9 . 10−7 ) and in the choice task ( βC ( G ) = 0 . 16±0 . 03 , t ( 18 ) = 5 . 6 , p = 2 . 10−5; βC ( G ) = 0 . 12±0 . 02 , t ( 18 ) = 5 . 4 , p = 4 . 10−5 ) . Interaction terms were significant for the rating and force tasks but not for the choice task ( βR ( G*D ) = -2 . 10–4±9 . 10−5 , t ( 18 ) = -2 . 7 , p = 0 . 01; βF ( G*D ) = -3 . 10–5±1 . 10−5 , t ( 18 ) = -2 . 6 , p = 0 . 02; βC ( G*D ) = 1 . 10–5±2 . 10−4 , t ( 18 ) = 0 . 1 , p = 0 . 95 ) . In none of the tasks did we find a significant difference between the weights of gain and for donation , although there was a trend in favor of selfishness ( R: t ( 18 ) = 1 . 79 , p = 0 . 089; F: t ( 18 ) = 1 . 10 , p = 0 . 29; C: t ( 18 ) = 1 . 70 , p = 0 . 11 ) . We also regressed the residuals of this regression against trial and session number , in order to test for fatigue effects . As none of these tests was significant ( all p>0 . 1 ) , we did not include any parameter accounting for fatigue in our computational models . Finally , we compared the distribution of forces and ratings , irrespective of gain and donation . As uncertainty was controlled by force production in the effort task , the distribution could be affected by risk attitude , relatively to the rating task in which uncertainty was constant . Indeed , subjects should avoid medium forces , if they are risk averse , or on the contrary favor them , if they are risk seeking . We thus fitted a second-order polynomial function to individual distributions of forces and ratings . The coefficients of quadratic regressors were significant for both tasks ( F: b = -0 . 31 ± 0 . 11 , t ( 18 ) = -2 . 75 p = 0 . 013 , R: b = -0 . 21 ± 0 . 06 , t ( 18 ) = -3 . 34 , p = 4 . 10–3 ) , with no significant difference between tasks ( t ( 18 ) = -0 . 85 , p = 0 . 41 ) . There was therefore no evidence that risk attitude created a difference between forces and ratings . However , these model-free analyses do not provide any formal conclusion about how value functions differ across tasks , so we now turn to a model-based Bayesian data analysis . In order to further investigate how changing the elicitation paradigm could affect the subjective value of potential outcomes , we defined a set of twelve value functions that could explain the observed behavior in each task ( see Methods ) . These value functions represent different ways of combining the two dimensions ( gain and donation ) composing the outcomes proposed in the tasks . They were used to generate forces and ratings with linear scaling ( with slope and intercept parameters ) and choices with logistic projection ( softmax function with temperature parameter ) . All value functions were fitted on behavioral responses for every subject and task using Variational Bayesian Analysis ( VBA ) [21 , 22] . The explained variance ( averaged across subjects ) was comprised between 43 and 70% in the force task , between 57 and 85% in the rating task and between 45 and 85% in the choice task . These results show that , for all three tasks , there were important differences in the quality of fit between value functions , which we compare below .
To our knowledge , this is the first study comparing direct elicitation of cardinal values ( rating and force tasks ) to ordinal rankings ( choice task ) for a same set of options . Those tasks are widely used in neuroeconomics and it is somewhat comforting that they reveal similar value functions driving the behavior despite trivial differences . They nonetheless present different advantages and drawbacks that may guide the design of future studies .
The study was approved by the Pitié-Salpétrière Hospital ethics committee . All subjects were recruited via e-mail within an academic database and gave informed consent before participation in the study . Participants were right-handed , between 20 and 30 years old , with normal vision and no history of neurological or psychiatric disease . They were not informed during recruitment that the task was about giving money to a charity , in order to avoid a bias in the sample . Nineteen subjects ( 10 females; age , 22 . 2 ± 1 . 4 ) were included in the study . They believed that the money won while performing the task would be their remuneration for participating , but eventually , their payoff was rounded up to a fixed amount ( 100€ ) . Subjects performed the three tasks , the order being counterbalanced across subjects for the force and rating tasks . The choice task was always performed after the two others , which were performed during MRI scanning for other purposes . The force task was preceded by maximal force measurement for the right hand [6] . Participants were verbally encouraged to squeeze continuously as hard as they could until a line growing in proportion to their force reached a target displayed on a computer screen . Maximal force was defined as the maximal level reached on three recordings . Then subjects were provided a real-time feedback about the force produced on the handgrip , which appeared as a red fluid level moving up and down within a thermometer , the maximal force being indicated as a horizontal bar at the top . Subjects were asked to try outreaching the bar and state whether it truly corresponded to their maximal force . If not , the calibration procedure was repeated . In the force and rating tasks , 121 trials were presented in a random order across three sessions of 40 or 41 trials . Each trial corresponds to one of the 121 combinations of the experiment design ( eleven possible incentives for themselves by eleven possible incentives for charity donation: from 0€ to 100€ with steps of 10€ ) . Subjects performed the three sessions with the right hand , with short breaks between sessions to avoid muscle exhaustion . In the force and rating tasks , each trial started by revealing the potential outcome , composed of two monetary incentives , with the inscriptions “YOU” followed by the amount for the subject , and “ORG” followed by the amount for the charity ( Fig 1 , top ) . The outcome was displayed for a duration jittered between 4 and 6 seconds . In the force task , subjects knew that the probability to win the outcome was proportional to the force they would produce after the display of the thermometer on the screen . More precisely , the probability of winning was equal to the percentage of their maximal force that they produced in the current trial . Subjects were also instructed to manage their forces in the effort task to avoid any frustration due to potential fatigue effect , and to use breaks between sessions to recover their muscular strength . During task trials , they were provided with online feedback on the exerted force ( via a fluid level moving up and down within a thermometer ) . They were also informed that they had to produce a minimal effort in every trial ( 10% of their maximal force ) and that the trial would be over when they stop squeezing the handgrip . Each trial ended with the display of the final outcome of their effort , for a duration jittered between 4 and 6 seconds , via the words “WON” ( with the proposed monetary earnings ) or “LOST” ( with null earnings for both subject and charity ) . The rating task only differed at the time of the motor response . Instead of a thermometer , a vertical rating scale from 0 to 10 units appeared after presentation of the potential outcome . Subjects were asked to rate the desirability of the outcome on the screen by moving the cursor through button presses with the right hand ( index and middle finger for moving the cursor left and right , and ring finger for validating the response ) . They were asked to use the whole scale across trials . They were also informed that their rating would have no impact on the final outcome . They were then shown the final outcome that was randomized to obtain a “WON” in 70% trials , and a “LOST” 30% of trials ( i . e . , a proportion similar to that obtained in the force task ) . The binary choice task included 200 trials , each presenting two composite options , one on each side of the screen . After considering the two options for 2 seconds , subjects could indicate the one they would prefer to win using their right hand ( index vs . middle finger for left vs . right option ) . This option was actually won in 70% of trials , which was indicated with a positive feedback ( “WON” ) accompanied by the selected earnings . In the other 30% of trials , a negative feedback ( “LOST” ) was shown with a null outcome ( 0€ ) for both receivers . Given the number of options in our design , there were 1212 ( 14641 ) possible binary choices . Constraints can be applied to reduce this number: choices are informative only if options are crossed ( attributes never dominate on both dimensions ) , if options differ on both dimensions , and if the pair of options was not previously presented . However , those constraints only reduced the number of choices to 3025 . Thus , we used an online optimization design to exploit the fact that some options are more informative than others to estimate a value function . At each trial , the design was optimized over a single dimension ( gain or donation ) . The chosen combination was the one that minimized the trace of the posterior covariance matrix over the parameters of an a priori value function defined as follows: V ( G , D ) = βG * G + βD * D + βGD * ( G * D ) , corresponding to a linear integration with interaction [20] . Contrary to the force and rating tasks , the amounts for subjects and charity could vary with steps of 1€ ( still between 0€ and 100€ ) , since options were optimized for each trial and subject . Subjects were informed that three trials would be randomly drawn ( one per task ) and that the average outcome would be actually implemented ( including both their gain and donation ) . They were aware that their responses in the rating task would have no influence on the outcome , whereas they would have an impact in the effort and choice task . The uncertainty about winning the outcome was fixed to 70% in the choice and rating tasks , but controlled by the force produced in the effort task . As expected , the average forces were not significantly different from 70% ( 65±3% , p>0 . 1 ) , and hence matched the uncertainty level of the other tasks . | In economic decision theory , value is a construct that provides a metric to compare options: agents are likely to select options leading to high-value outcomes . In neuroscience , different behavioral tasks have been used to elicit the subjective values of potential outcomes , notably rating tasks , which demand an explicit value judgment on the outcome , and effort tasks , which demand an energetic expense ( in order to increase outcome probability ) . However , it remains unclear whether the values elicited by these tasks are the same as the values that drive choices . Indeed , it has been argued that they involve different costs and consequences for the agent . Here , we compared value models that could account for behavioral responses in choice , rating and effort tasks involving the same set of options , which combined a monetary gain for the participant and a donation to a charity . We found that the most plausible model was that a same value function , with a similar selfishness parameter ( relative weight on gain and donation ) , generated the responses in all three tasks . This finding strengthens the notion of value as a general explanation of motivated behaviors , beyond standard economic choice . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"economic",
"agents",
"medicine",
"and",
"health",
"sciences",
"decision",
"making",
"engineering",
"and",
"technology",
"thermometers",
"behavioral",
"economics",
"social",
"sciences",
"limbs",
"(anatomy)",
"neuroscience",
"neuroeconomics",
"cognitive",
"psychology",
"co... | 2017 | Choose, rate or squeeze: Comparison of economic value functions elicited by different behavioral tasks |
Bdellovibrio bacteriovorus is a Delta-proteobacterium that oscillates between free-living growth and predation on Gram-negative bacteria including important pathogens of man , animals and plants . After entering the prey periplasm , killing the prey and replicating inside the prey bdelloplast , several motile B . bacteriovorus progeny cells emerge . The B . bacteriovorus HD100 genome encodes numerous proteins predicted to be involved in signalling via the secondary messenger cyclic di-GMP ( c-di-GMP ) , which is known to affect bacterial lifestyle choices . We investigated the role of c-di-GMP signalling in B . bacteriovorus , focussing on the five GGDEF domain proteins that are predicted to function as diguanylyl cyclases initiating c-di-GMP signalling cascades . Inactivation of individual GGDEF domain genes resulted in remarkably distinct phenotypes . Deletion of dgcB ( Bd0742 ) resulted in a predation impaired , obligately axenic mutant , while deletion of dgcC ( Bd1434 ) resulted in the opposite , obligately predatory mutant . Deletion of dgcA ( Bd0367 ) abolished gliding motility , producing bacteria capable of predatory invasion but unable to leave the exhausted prey . Complementation was achieved with wild type dgc genes , but not with GGAAF versions . Deletion of cdgA ( Bd3125 ) substantially slowed predation; this was restored by wild type complementation . Deletion of dgcD ( Bd3766 ) had no observable phenotype . In vitro assays showed that DgcA , DgcB , and DgcC were diguanylyl cyclases . CdgA lacks enzymatic activity but functions as a c-di-GMP receptor apparently in the DgcB pathway . Activity of DgcD was not detected . Deletion of DgcA strongly decreased the extractable c-di-GMP content of axenic Bdellovibrio cells . We show that c-di-GMP signalling pathways are essential for both the free-living and predatory lifestyles of B . bacteriovorus and that obligately predatory dgcC- can be made lacking a propensity to survive without predation of bacterial pathogens and thus possibly useful in anti-pathogen applications . In contrast to many studies in other bacteria , Bdellovibrio shows specificity and lack of overlap in c-di-GMP signalling pathways .
Predatory , “attack-phase” cells of Bdellovibrio bacteriovorus HD100 use flagellar motility in liquid environments; and gliding motility on solid surfaces , to encounter other Gram-negative bacteria [1] , [2] . Prey bacteria include a wide range of pathogens of man , animals and plants , [3] , [4] thus Bdellovibrio can be seen as “pathogens of pathogens” [5] . B . bacteriovorus attach to these prey and enter their periplasms , by mechanisms that remain to be fully understood . Once inside prey , B . bacteriovorus become non-motile and degrade prey macromolecules , using them for their own growth and replication ( Figure S1 in Text S1 ) . Growth occurs from both poles , giving rise to odd and even numbered progeny by synchronous septation of the elongated multi-nucleoid B . bacteriovorus filament , inside the infected , spherical , prey cell which is now called a “bdelloplast” [6] . At the end of predatory growth and septation , B . bacteriovorus induce motility once more , and use flagellar motility to emerge from prey in liquid media , or gliding motility to emerge from prey on solid surfaces , and move off , in a non-replicative , “attack phase” to seek more prey encounters . Cultures of B . bacteriovorus growing in this predatory or prey/host-dependent ( HD ) manner require entry to another prey cell to replicate as the “attack-phase” cells have replication suppressed ( by as yet unknown control mechanisms ) and do not grow using organic nutrients from the external media [7] . Attack phase cells are vibroid and 0 . 25 µm by 1 . 25 µm with a single polar sheathed flagellum , they attach to prey at the non-flagellar pole [8] . In the laboratory a small fraction: 1×10−7 , of attack phase B . bacteriovorus populations can also be cultured axenically without prey , as so-called HI ( host-independent ) cultures on peptone-rich artificial media; here they grow and divide as if inside prey cells ( Figure S1 in Text S1 ) [9] . Natural point mutations in the Bd0108 Type IVB pilus-like gene product are reported to account for the small fraction of predatory cells that can adopt this “wild type” HI phenotype [10] . Wild-type HI B . bacteriovorus HD100 convert readily back to HD cultures , via predatory invasion , if presented with prey cells in the absence of rich media . HI cells of B . bacteriovorus are typically longer than the attack phase HD cells , being 2–10 µm long but 0 . 25–0 . 3 µm in width; longer cells usually have a serpentine morphology and one or more flagella in polar or other cell sites . Cyclic di-GMP is a bacterial second messenger that controls various processes . In the Proteobacteria , c-di-GMP controls a lifestyle transition between single , usually motile , cells and surface-attached multicellular communities ( biofilms ) [11] , [12] . It also contributes to the switch between environmental and pathogenic lifestyles in other organisms [13] . C-di-GMP is synthesized by diguanylyl cyclases containing GGDEF domains , and degraded by phosphodiesterases containing EAL or HD-GYP domains . C-di-GMP acts via a variety of receptors ( effector proteins ) [14] . Among the most common receptors are PilZ domain proteins and a sub-group of GGDEF domain proteins containing so-called I-sites [15] . These sites are present in many diguanylyl cyclases where they function in feedback inhibition . However , a sub-group of the GGDEF domain proteins , which are enzymatically inactive , have evolved that bind c-di-GMP via the I-sites and function as c-di-GMP receptors . B . bacteriovorus HD100 is predicted to have a high c-di-GMP “intelligence” because its 3 . 8Mb genome contains as many as 15 PilZ domain proteins representing putative c-di-GMP receptors [16] , [17] , [18] . It is peculiar that the number of predicted diguanylyl cyclases and phosphodiesterases is relatively low , i . e . 5 GGDEF , 1 EAL and 6 HD-GYP domain proteins , which facilitates analysis of c-di-GMP signalling cascades as genes encoding enzymes for c-di-GMP synthesis and degradation can be deleted and the resultant phenotypes tested . In this study , we carried out a deletion analysis of genes encoding 5 GGDEF domain proteins that may initiate c-di-GMP signalling cascades . Unexpectedly , this analysis produced a discretely different phenotype for each GGDEF domain gene deletion strain . One GGDEF knockout strain was rendered obligately predatory , a phenotype that is very desirable for the future application of Bdellovibrio as anti-infective agents to kill pathogenic bacteria [5] . We discovered that individual c-di-GMP signalling pathways control each of the axenic and predatory lifestyles and B . bacteriovorus gliding and flagellar motility . The extreme degree of specificity and lack of the overlap among the signalling pathways involving a small diffusible molecule has not been previously observed in bacteria , or believed to be possible . The small ( 0 . 25 µm×1 . 25 µm ) B . bacteriovorus cell size , which might have been thought to facilitate rapid c-di-GMP equilibration , did not cause nonspecific cross-activation of c-di-GMP cascades . While the mechanisms involving pathway separation are yet to be uncovered , unique cellular localizations of some diguanylyl cyclases and c-di-GMP receptors observed here may contribute to the lack of pathway cross-reactivity . The pathways include one degenerate GGDEF domain protein , with a GVNEF motif , which we found to be a c di-GMP receptor required for the rapid entry of B . bacteriovorus into prey cells . Thus c di-GMP signalling has evolved to control the predatory ability of this naturally invasive killer of pathogenic bacteria .
We analyzed the sequences of the 5 GGDEF domain proteins encoded in the B . bacteriovorus genome using Pfam ( Figure S2 in Text S1 ) The sequences of GGDEF domains of four proteins were very similar to the consensus ( Pfam database [19] ) and did not contain residues predicted to interfere with the diguanylyl cyclase activity [20] . This suggested that these four proteins are likely enzymatically competent . We designated the B . bacteriovorus HD100 locus tags for these proteins [21] as encoding DgcA ( Bd0367 ) , DgcB ( Bd0742 ) , DgcC ( Bd1434 ) and DgcD ( Bd3766 ) , where Dgc stands for diguanylyl cyclase . A genome sequence comparison revealed that DgcA , DgcB and DgcC are conserved between different Bdellovibrio strains , whereas gene Bd3766 , encoding DgcD , is not conserved in the genome of another B . bacteriovorus strain tiberius ( Hobley et al . , in preparation ) . Pfam predicts that [19] DgcA has an N-terminal receiver domain typical of bacterial response regulators ( Figure S2 in Text S1 ) , which implies that its activity is regulated by phosphorylation by an unknown histidine kinase . DgcB has an N-terminal forkhead domain , which may be involved in binding of a protein containing a phosphorylated serine or threonine residue [22] . DgcC contains a large N-terminal domain of unknown function . DgcD has a predicted periplasmic N-terminal domain containing 4 tetratricopeptide repeats , TPR domains , flanked by the transmembrane domains . The GGDEF domain of the fifth protein , Bd3125 , was clearly degenerate . Among others , it had two substitutions at key residues in the most conserved GGDEF ( Gly-Gly-Asp-Glu-Phe ) motif , the catalytic half site required for substrate , GTP , binding , i . e . , GGDEF→GVNEF . Thus protein sequence analysis suggested that Bd3125 is not enzymatically competent . Each of the five proteins , including Bd3125 , contains an RxxD motif 5 residues upstream of the GGDEF motif , which is a typical sequence and position for the I-site . For the four Dgc proteins , this site likely represents a site for feedback inhibition . Given the likely lack of diguanylyl cyclase activity of Bd3125 but the presence of the I-site , we predicted that it may function as a c-di-GMP receptor ( effector ) protein , and as such designated it CdgA . To characterize activities of the GGDEF domain proteins , we first expressed them as fusions to the maltose-binding protein , MBP , in the highly motile strain , E . coli MG1655 . For DgcD that contains a large transmembrane domain , the fusion was made directly to the GGDEF domain . Ko and co-workers , followed by Ryjenkov and co-workers have shown that elevated c-di-GMP levels inhibit E . coli motility in swim agar plates [23] , [24] . Each MBP-fusion was tested for the effect it had on E . coli motility in swim agar plates . As expected , DgcA , DgcB and DgcC strongly inhibited E . coli swimming ( Figure S3 in Text S1 ) , consistent with their predicted diguanylyl cyclase activities . However , the MBP-fusion to the GGDEF domain of DgcD did not inhibit swimming , either because its GGDEF domain is enzymatically inactive , or , more likely , because it could not be activated in E . coli . Because the B . bacteriovorus dgcD knockout produced no observable phenotype , biochemical characterization of DgcD was not pursued . Expression of CdgA produced a significantly larger swim zone than that of the empty vector ( Figure S3 in Text S1 ) . Since CdgA cannot possibly act as a c-di-GMP degrading enzyme , we predict that it likely binds c-di-GMP and acts as a c-di-GMP sink , thus effectively decreasing the pool of available c-di-GMP in E . coli . We purified the MBP-fusions to DgcA , DgcB , DgcC and CdgA and tested diguanylyl cyclase activities and c-di-GMP binding in vitro . As expected , DgcA , DgcB and DgcC proved to function as diguanylyl cyclases ( Figure 1A–C ) . Because each of these proteins contains an N-terminal regulatory domain whose activation cannot be ensured in E . coli , it is likely that the observed activities of the purified fusions lie between their respective active and inactive states . Among the three cyclases , DgcB is most sensitive to feedback inhibition by c-di-GMP ( Figure 1B ) . DgcB purified from E . coli contained c-di-GMP at approximately 1∶1 protein:c-di-GMP molar ratio . No diguanylyl cyclase activity was observed following addition of the substrate , GTP , to the purified protein . Only after extended ( approximately 30 minutes ) incubation at 37°C , could the DgcB activity be detected , probably due to unfolding of the c-di-GMP-inhibited DgcB conformation and spontaneous formation of the enzymatically competent DgcB dimer [25] . The CdgA protein containing a degenerate GGDEF domain showed no diguanylyl cyclase activity , which is consistent with the sequence analysis and motility assays in E . coli ( Figure S3 in Text S1 ) . However , this protein bound c-di-GMP in vitro with an apparent Kd ∼2 µM ( Figure 1D ) , as measured by equilibrium dialysis [26] . This value is well within the range of intracellular c-di-GMP concentrations observed in bacteria [12] . Therefore , CdgA is a c-di-GMP receptor ( effector protein ) , which , according to the genetic analysis shown below , acts downstream of the diguanylyl cyclase DgcB . An in-frame deletion of dgcA was achieved by conjugation of a suicide vector pK18mobsacB into B . bacteriovorus HD100 and subsequent sucrose suicide screening for gene replacement and loss of the plasmid ( see materials and methods and Text S1; [2] , [27] ) . The ΔdgcA mutant strain could not grow predatorily and did not form plaques ( Figure 2Ai ) on lawns of prey bacteria , thus it was isolated solely by axenic culturing on peptone-rich media . The ΔdgcA mutant was seen by phase contrast microscopy to be non-motile in liquid media . Electron microscopy ( Figure 3AI , ii ) showed that although a membranous flagellar sheath was made , no functional flagellum was assembled within it . This appearance was reminiscent of a non-motile fliC3 flagellin mutant previously studied in Bdellovibrio [1] , [28] . Studies with that fliC3 mutant had shown it would enter prey cells if applied directly onto them on a solid surface , and could use gliding motility to exit the bdelloplast after septation and lysis [1] , [2] . Therefore cells of the ΔdgcA mutant were applied to E . coli prey ( that were constitutively expressing mCherry ) on a 1% agarose surface , the prey were invaded and the ΔdgcA mutant replicated and septated within them . However , after septation no induction of ( flagellar or ) gliding motility was seen in the ΔdgcA strain and the septated Bdellovibrio remained motionless inside the dead “shell” of the prey bdelloplast long after ( as long as 120 minutes was measured without any change being observed ) the prey mCherry fluorescence was dissipated by the lysis of the bdelloplast ( Figure 4 ) . In contrast , wild-type Bdellovibrio lyse bdelloplasts and immediately ( after 10 minutes ) , emerge , using gliding motility if on an agarose surface . Thus the deletion of the dgcA gene rendered the cells effectively unable to be predatory , as they could not escape the prey remains to search for new prey . A DgcA-mCherry strain was seen to be fluorescent throughout the cell at all times through the predatory cycle ( Figure 5Ai ) , and during Host-Independent growth ( Figure 5Aii ) , not only at times when the cells would be motile ( by either flagellar- or gliding- motility ) . Complementation of the ΔdgcA mutant with a full length but C-terminally mCherry-tagged wild type dgcA gene was achieved by single recombination into the chromosome from suicide plasmid pK18mobsacB . Exconjugant Bdellovibrio from this complementation regained the ability to form plaques on E . coli prey lawns ( Figure 2Bi ) and typically formed 2×03 plaques per conjugation , compared to zero for the parental ΔdgcA mutant strain ( Figure 2Ai ) . Gliding was restored to the ΔdgcA mutant by this complementation ( Figure S4 in Text S1 ) but interestingly flagellar motility was not . However , we have previously shown that flagellar motility is not required for predation [1] . We were able to observe , by video microscopy , gliding cells of the complemented strain having lysed and escaped from the prey bdelloplasts . Unfortunately the full-length dgcA gene without mCherry tag proved difficult to isolate ( possibly due to c-di-GMP effects in cloning strains of E . coli ) so the phenotype of the complemented ΔdgcA strain was only discernable from the mCherry tagged dgcA gene . We acknowledge that the presence of the mCherry tag may interfere with protein-protein interactions that could be required for full complementation , so we cannot test the flagellar effect . However as gliding motility and predatory growth were restored by this dgcA-mcherry complementation , we conclude that the DgcA protein controls gliding motility and thus successful predatory growth on prey bacteria . Inactivation of dgcB gave a mutant strain which did not form plaques ( Figure 2Aii ) and which could only be isolated by HI axenic growth ( Figure 3B ) but which retained expression of wild-type flagella . Some 50% of axenically grown wild type HI Bdellovibrio have a flagellum but only 5% of those HI cells swim ( in contrast to the 98% motility of predatory Bdellovibrio cultures ) . For the dgcB mutant HI cells 50% had flagella and 10–15% of the cells were motile , more than for the wild type HI strains . When 21 cultures , consisting of seven replicates of three independent isolates , of this strain were challenged with prey bacteria in liquid culture , versus 3 cultures of two wild type ( WT ) B . bacteriovorus HI strain controls , predatory invasion and prey bdelloplast formation was seen in the WT B . bacteriovorus HI strains after 9 hours and the prey cultures began to clear by prey-lysis . At this point no bdelloplasts were seen for the ΔdgcB mutant strains . At 24 hours of incubation there were no prey remaining in the WT HI cultures but the ΔdgcB mutant cultures were full of prey bacteria . In 4 of the 21 test cultures of the ΔdgcB mutant strains prey killing and bdelloplast formation were seen after an additional 120 hours incubation . In the remaining 17 cultures however , no bdelloplasts were seen despite prolonged further incubation . The 4 presumed suppressor strains were tested by PCR for the presence of the original ΔdgcB mutation and this was confirmed . Repeating this experiment on two different occasions , in a different experimental setting , luminescent prey bacteria [29] were incubated in 96 well Optiplates ( Porvair Biosciences ) in a BMG luminescent plate reader with 18 different cultures of two different isolates of the ΔdgcB mutant ( including the isolate that had given presumed suppressor strains before ) , were incubated for 72 hours . The predator:prey ratio in these latter experiments should have shown clearly , even low-levels ( 5% of total ) of prey-killing by a drop in prey luminescence . No such drop in luminescence was seen in any of the 18 isolates , ( data not shown ) showing that the phenotype of the ΔdgcB mutant is non-predatory , but that rare suppressor mutants may be isolated ( these suppressor strains are currently the subject of further study ) . A DgcB:mCherry fusion protein expressed in a wild type DgcB background was constitutively brightly fluorescent in the cytoplasm of B . bacteriovorus cells whether they were growing predatorily ( Figure 5 Bi-ii ) or axenically ( Figure 5 Biii ) . As the DgcB protein contains a predicted forkhead domain at the N terminus , we conclude that it may be activated via this domain when its c-di-GMP synthetic properties are biologically appropriate . This activation may be in response to prey sensing . Complementation of the ΔdgcB mutant with the wild type dgcB gene was achieved by single recombination into the chromosome from suicide plasmid pK18mobsacB . A single complementation conjugation gave rise to 1 . 7×104 plaques on prey lawns ( Figure 2Bii ) in stark contrast to the ΔdgcB mutant itself which did not form plaques ( Figure 2Aii ) and had to be cultured axenically as HI strains ( and in which only very rare suppressor strains emerged after many days of prey challenge in liquid cultures of those axenically growing HI ΔdgcB mutants ) . Further evidence that the diguanylyl cyclase activity of the DgcB itself was important for predation came from the conjugation of a GGAAF variant into the ΔdgcB mutant instead of the wild type conjugation . The GGAAF plasmid did not restore plaquing ability ( Figure 2C ) to the ΔdgcB strain . Thus we conclude that the diguanylyl cyclase activity of the DgcB protein controls , via c-di-GMP signalling , processes that are required for the invasion of prey bacteria . Inactivation of the dgcC gene produced attack phase Bdellovibrio cells that although normally flagellate , were smaller , but wider ( Figure 3Ci ) than wild type ( Figure 3C ii ) ( Figure 6 ) . The dgcC mutant was still predatory , and invaded and killed E . coli prey cells at wild type rates . However , the normal 1 in 107 conversion of predatory Bdellovibrio to axenically growing HI Bdellovibrio [9] was lost in this mutant and 4 . 5×1011 cells had to be applied to nutrient media to isolate and culture a rare single colony ( Figure 3Di ) , growing axenically . This suggested a secondary mutation was required to allow the ΔdgcC mutant to grow axenically . Bacteria from such suppressor strains were frequently ( 10% ) biflagellate , but not markedly different ( within the natural variation of HI Bdellovibrio from wild-type HI strains ) as exemplified by the wild-type strain HID13 ( Figure 3Dii ) . Fluorescent microscopy of DgcC-mCherry in B . bacteriovorus carrying a wild type copy of the dgcC gene , showed a cytoplasmic distribution in attack-phase B . bacteriovorus ( Figure 5Ci ) and only faint fluorescence when Bdellovibrio was elongating inside prey cells ( Figure 5Cii ) , with fluorescence gone when the predatory cells were septating in the bdelloplast ( Figure 5Ciii ) . Attack-phase cells of the fluorescent DgcC-mCherry strain ( which also carried a wild type dgcC gene ) readily converted ( at wild type , 1 in 107 frequency ) , into axenically growing HI cultures . In these HI cells there were strong fluorescent foci of DgcC-mCherry expression at one pole of each cell ( Figure 5Civ ) and in 60% of dividing HI cells DgcC foci were seen associated with division points at non-Dapi-staining regions ( Figure 5Cv ) . There were also fainter foci along the length of non-dividing longer HI cells ( Figure 5Civ , v ) . There was no overall cytoplasmic distribution , in contrast to that of the DgcC-mCherry expression in predatory attack-phase cells ( which do not divide outside of prey ) . We propose that DgcC may be required to regulate predatory cell division and/or cytoskeletal regulation during predatory growth . This is because the ΔdgcC progeny produced were shorter and fatter than wild-type ( Figure 3Ci , ii , Figure 6 ) . Also it was noteworthy that DgcC-mCherry fluorescence was present when the Bdellovibrio were outside prey or in the early stages of invasion ( Figure 5Ci , ii ) , but that it dissipated at septation ( Figure 5Ciii ) . , suggesting DgcC turnover at septation . Wild-type attack-phase Bdellovibrio cannot replicate outside prey , due to an unknown control mechanism [7] . It is interesting , in support of this , that the conversion to axenic growth mode of Bdellovibrio was abolished in the DgcC mutant . Axenic and predatory division of Bdellovibrio does not have to use a completely overlapping set of cell division proteins , as the Bdellovibrio genome is large , and as predatory division is by synchronous septation of a long filament , yet in axenically growing Bdellovibrio budding events ( of single cells from the tip of a long filament ) , binary fission events ( of shorter serpentine HI cells ) and other division patterns are seen . One gene , Bd0108 , at the hit locus has been reported by others as a hot spot for mutations which affect the ability of Bdellovibrio to convert from predatory to axenic growth [9] , [10] , [30] . Despite this the sequence of the Bd0108 gene in the ΔdgcC strain was found to be wild type , in addition the sequence of Bd0108 in the rare ( 1 in 4 . 5×1011 ) “HI suppressor of ΔdgcC” that grew was also wild type . Complementation of the ΔdgcC mutant with the wild type dgcC gene was achieved by single recombination into the chromosome from suicide plasmid pK18mobsacB . Exconjugants from this complementation were predatory , as was the original ΔdgcC mutant strain , and the cells returned to the wild-type cell width dimension of 0 . 23 µm and almost returned to the wild-type length of 1 . 25 µm ( Figure 6 ) . However when a version of the dgcC gene , encoding GGAAF instead of wild-type GGEEF ( to inactivate the catalytic site of the protein ) was used in identical conjugation experiments , the exconjugant strains were not fully complemented in terms of cell size ( Figure 6 ) , showing intermediate size distributions of length and width . The normal 1 in 107 conversion of wild type predatory Bdellovibrio to axenically growing HI Bdellovibrio ( Figure 2Di ) was restored to the ΔdgcC mutant by wild type dgcC complementation ( Figure 2Dii ) and colonies of axenically growing Bdellovibrio were readily isolated from nutrient media plates . Further experiments will determine whether the GGAAF encoding version of the dgcC plasmid restores the normal HI conversion rate . In contrast to the other GGDEF genes , an in-frame deletion of the dgcD gene did not cause an observable phenotype in the conditions that we measured and the mutant strains could be cultured in both HD ( Figure 2Aiii ) and HI growth modes without alteration in growth rates . Introduction of the wild type dgcD gene had no effects on growth ( Figure 2Biii ) . They also retained wild type morphology and motility ( Figure 3Ei , ii ) . Bright fluorescent foci of DgcD-mCherry were seen ( Figure 5D iv ) in wild type HI Bdellovibrio , but very weak constitutive fluorescence only , was seen in HD cells ( Figure 5 D I , ii , iii ) . Sequencing of another B . bacteriovorus isolate in our laboratory ( in preparation ) has shown that while the other 4 GGDEF genes are conserved across strains , that dgcD is not . Due to the lack of a significant phenotype for the dgcD deletion and the lack of conservation among Bdellovibrio strains , we did not investigate the role of DgcD any further . Exconjugants with an in-frame deletion of the cdgA gene were not predatory and did not form plaques on prey lawns ( Figure 2A iv ) and ΔcdgA Bdellovibrio could only be isolated in HI axenically grown plate cultures ( Figure 3F ) . However , when the ΔcdgA mutant was offered to prey cells in mixed liquid cultures , prey invasion and bdelloplast formation did proceed , but slowly , i . e . 40–90 minutes for ΔcdgA versus 30–40 minutes for wild-type HD strains ( Figure S5 in Text S1 ) . This slow predation resulted in liquid cultures taking two days to clear the majority of prey cells as compared to around 16 hours for wild-type strains . This slow prey invasion , likely accounted for the failure of the ΔcdgA mutant to be efficiently cultured predatorily on plates of prey lawns when originally isolated . Furthermore , when the ΔcdgA mutant , that had been growing slowly predatorily in liquid cultures , was returned to prey lawns on agar plates , they again failed to form plaques . Complementation of the ΔcdgA mutant with the wild type cdgA gene was achieved by single recombination into the chromosome from suicide plasmid pK18mobsacB . Exconjugants from this complementation gave rise to >105 plaques per conjugation , counted on dilution plates of prey lawns ( Figure 2B iv ) in contrast to none from matched numbers of the original ΔcdgA strain ( Figure 2A iv ) . Invasion of prey cells by the complemented strain , previously grown on prey , was found to occur after 30–40 minutes ( a wild type speed ) by timelapse microscopy ( Figure S5 in Text S1 ) . Fluorescent protein localization of CdgA-mCherry , in a wild type CdgA-expressing B . bacteriovorus HD100 , showed that it was expressed at the non-flagellar pole or prey-interacting “nose” of the predatory B . bacteriovorus cells and that this expression was seen when the Bdellovibrio were attached to the prey cell at the point of invasion ( Figure 5E i , ii ) . After invasion a single fluorescent polar focus of CdgA-mCherry persisted whilst the Bdellovibrio began to elongate , initially from this single pole . Then a second CdgA-mCherry focus developed at the second pole - during the period in which the Bdellovibrio grows via bipolar elongation ( Figure 5E iii , iv ) ; [6] ) . Upon septation of the predatory Bdellovibrio , within the prey bdelloplast , a single bright CdgA-mCherry focus was seen at one pole of each septated Bdellovibrio cell , prior to release from prey ( Figure 5E v ) . In axenically-growing , longer , HI cells polar and sometimes bipolar expression of this protein was also seen ( Figure 5E vi , vii ) . This pattern correlates with video-microscopy observations we have made where some unusually long HI Bdellovibrio wild-type cells , found by chance with each pole adjacent to a prey cell , simultaneously invade both prey cells at once . In such unusually long HI Bdellovibrio ( which do not have the usual polar flagellum ) , it seems that each pole can be competent for prey entry ( Figure S6 in Text S1 ) . Among mutations in the c-di-GMP signalling proteins , the ΔdgcB and ΔcdgA mutations were the only ones that impaired predatory growth , specifically at the prey entry stage , which suggests that DgcB and CdgA belong to the same regulatory cascade , where DgcB signals , via c-di-GMP synthesis , to CdgA . A more severe phenotype of the dgcB deletion versus the cdgA deletion implies that DgcB signals via more than a single target . The DgcB targets likely regulate activity/expression of host hydrolytic and prey modifying/degrading proteins involved in penetration of prey bacteria . It appears that diguanylyl cyclase activity of DgcB is activated via an encounter with prey , possibly via the forkhead domain of DgcB . Using the DgcB-mCherry fusion we determined that DgcB is apparently expressed at relatively high levels and located throughout the cell . The method described by Bobrov and co-workers [31] was used to determine the c-di-GMP levels in pure Bdellovibrio cells from the axenically grown ΔdgcA , B and ΔcdgA mutant HI strains , versus matched wet cell weights of wild type HI controls . A similar analysis was not possible for predatory strains due to possible contamination with prey-derived c-di-GMP . The mean level of c-di-GMP for wild type axenic HI B . bacteriovorus strains HID13 and HID26 was 1 . 4+/−0 . 1nM/mg wet weight of cells . For the ΔdgcA strain this value was considerably lower at 0 . 4nM/mg wet weight of cells , for the ΔdgcB strain the value was 1 . 5nM/mg wet weight of cells and for the ΔcdgA strain the value was slightly higher at 2 . 0nM/mg wet weight of cells . Thus the absence of the DgcA protein noticeably reduced the extractable c-di-GMP content of the B . bacteriovorus cells , consistent with the diguanylyl cyclase activity of DgcA . The absence of the CdgA protein slightly increased the extractable c-di-GMP content but the absence of the DgcB protein did not cause a significant difference to extractable c-di-GMP levels . The latter result is not surprising given the expectation that DgcB is activated when the Bdellovibrio are in contact with , or the immediate proximity of prey ( which was not the case in the axenically grown cells used here . In summary , ( Figure 7 ) we have found that each of 4 conserved GGDEF proteins in B . bacteriovorus HD100 contributes non-overlapping regulatory controls to different aspects of the predatory or axenic life cycles of this bacterium , and that these can be seen in single GGDEF gene mutants with the remaining GGDEF genes intact . We found that DgcB controls predatory invasion of prey bacteria , by signalling to c-di-GMP receptors , probably including CdgA , a degenerate GGDEF protein located at the prey-interacting “nose” of Bdellovibrio . As DgcB may receive information that indicates proximity or attachment to prey being sensed by the Bdellovibrio , we are now searching for the signals that activate DgcB . This will allow us to understand signals that “tell” Bdellovibrio that prey are present and aid the targeting of Gram-negative pathogens for destruction in anti-infective settings . DgcA controls gliding motility ( Figure 7 ) which is required for the Bdellovibrio to exit the exhausted prey debris after predation and to move off to regions where new prey encounters would be possible [1] , [2] . To our knowledge this is the first case where c-di-GMP controls gliding motility , whereas c-di-GMP control of flagellar motility is better understood in other bacteria . As only gliding motility and predation , not flagellar synthesis , were found to be restored by the DgcA-mCherry complementation experiment , it may be that DgcA does not regulate flagellar synthesis in B . bacteriovorus HD100 . DgcA has a response regulator domain at the N terminus and thus has some similarity to WspR which controls flagellar biogenesis versus pellicle formation in Pseudomonas [32] , [33] . Flagellar biogenesis is not however required for B . bacteriovorus predation [1] and further studies on DgcA will define whether it also regulates flagellar biogenesis . We do note that , at the 5′ end of one of 4 operons of gliding motility genes of B . bacteriovorus [2] , there is a gene , Bd1482 , encoding a PilZ domain protein , which may be a candidate for receiving signals from DgcA to effect gliding and thus we have tentatively labelled this ( Figure 7 ) as a candidate to receive c-di-GMP signals from DgcA . Gliding will be particularly important in the predation of pathogen biofilms where Bdellovibrio have been shown to be effective [34] . DgcC controls the transition between the predatory “attack phase” of Bdellovibrio when it is “locked into” a non-replicative phase , hunting prey; and the replicative , axenic , growth phase , on protein rich media . This is a mysterious transition [10] , [35] . It is noteworthy that DgcC-mCherry forms foci in growing axenic cells and we hope that studying any proteins with which DgcC forms a complex may illuminate this control of non-predatory replication . The DgcC deletion mutant is the gateway to producing a useful , obligately predatory Bdellovibrio for application as a self-limiting living antibiotic . We have revealed the impressive c-di-GMP intellect of this tiny bacterium and how it is employed in its unique intra-bacterial niche , and we can now see how this can be manipulated to use it as an anti-infective [18] . Studying the PilZ receptor domain proteins of Bdellovibrio , of which there are 15 [36] , will dissect further , the stages by which predatory and axenic growth processes are regulated , and will reveal the cyclase- receptor relationships . This study has contributed to our understanding of specificity of c-di-GMP signalling pathways; something where different models have been proposed from workers studying other bacteria . Some have suggested that little specificity exists between diguanylyl cyclases and their targets [37]; with all expressed diguanylyl cyclases contributing to the intracellular c-di-GMP pool that is sensed by all c-di-GMP receptors or targets . In an alternative , “c-di-GMP cloud” model [38] , each diguanylyl cyclase has a specific target; little crosstalk exists between different c-di-GMP signalling pathways because c-di-GMP bursts are localized , and the spillover is prevented by c-di-GMP phosphodiesterases . Several studies have supported the later model by data showing that intracellular levels of c-di-GMP do not necessarily correlate with the observed phenotypes , that c-di-GMP is produced at unique cellular locations , and most directly , specific c-di-GMP targets were linked to individual diguanylyl cyclases [39]-[43] . However , B . bacteriovorus is the first bacterium in which the strikingly different phenotypes of the diguanylyl cyclase mutants shows an exclusive cyclase- c-di-GMP target ( s ) relationship . It is ironic that the task of proving the existence of c-di-GMP signalling pathway specificity has fallen onto one of the tiniest bacteria , where diffusion rates could have been envisioned to make c-di-GMP spillover unavoidable . The small size of Bdellovibrio cells makes the exclusivity point most convincing . While some bacteria may have ‘general purpose’ diguanylyl cyclases , and activation of parallel c-di-GMP targets is perhaps advantageous for some organisms , under certain circumstances; specific c-di-GMP signalling from discrete diguanylyl cyclases is clearly at play in Bdellovibrio control pathways , and as this specificity occurs in such small bacteria; it should also be considered in others .
Bdellovibrio and E . coli strains are listed in Table S1 ( in Text S1 ) . Predatory B . bacteriovorus strains were routinely grown in Ca/HEPES with E . coli S17-1 as prey as previously described [1] , [44] . Host-Independent ( HI ) Bdellovibrio were isolated and grown in PY broth as previously described [44] , [45] . Kanamycin ( 25 µg ml−1 for E . coli and 50 µg ml−1 for Bdellovibrio ) , ampicillin ( 50 µg ml−1 ) and isopropyl-β-D-1-thiogalactopyranoside ( IPTG; 200 µg ml−1 for induction of fluorescence in E . coli ) were used where appropriate . Chromosomal deletions of the dgcA , dgcB , dgcD and cdgA reading frames ( Bd0367 , Bd0742 , Bd3766 and Bd3125 ) were created using a modified version of previously described methods [27] , [46] . Deletion of the dgcC ( Bd1434 ) ORF and replacement with a kanamycin resistance cassette was initially performed using a modified version of that described by Lambert and co-workers [29] and later followed up , to verify the cell size phenotype with the same dgcC chromosomal deletion using the silent deletion method as used for the four other genes above . Complementation analyses for dgcA , dgcB , dgcD and cdgA deletion strains used single recombination of either full-length mcherry tagged ( for dgcA ) or wild type dgc or cdg genes ( for the other strains ) into the Bdellovibrio chromosome from suicide plasmid pK18mobsacB . For strains dgcB and dgcC only , GGAAF expressing ( rather than wild type GGDEF or GGEEF ) genes were used in single recombination experiments to test the function of the GGDEF domain . ( Wild type CdgA does not have a conventional GGDEF domain so was excluded ) . Construction of each of the mutant and complemented strains is described fully in Text S1 . Bdellovibrio deletion strains were assayed for morphological changes by EM , after growth ( both predatory and axenic ) under standard conditions . Cells were stained with 1% phosphotungstic acid ( PTA; pH 7 . 0 ) and imaged with a JEOL 1200Ex electron microscope at 80 kV . Replicate cultures were grown independently on different days and flagellar patterns and cell morphologies were determined from 45–50 cells observed each time . Deletion mutants of dgcA , dgcB and cdgA could only be isolated using axenic ( HI ) growth . HI deletion strains were added in excess to prey in Ca/HEPES buffer and monitored by microscopy for signs of prey entry and replication at several time-points during prolonged incubation over multiple days ( up to 10 days for ΔdgcB; whilst wild-type HI strains [HID2 , HID13 and HID26 [47] , chosen for morphological similarities , see important considerations in Text S1] completed predation when checked after 24 hours of incubation ) . Both the ΔdgcA and ΔcdgA mutants were seen to enter prey within the first 24–48 hours of incubation in these liquid conditions and these were further analysed by time-lapse microscopy ( described below ) . Cells from prolonged incubation of ΔdgcB mutant strains with prey were re-confirmed to be ΔdgcB by PCR analysis . Deletion mutants of both dgcC and dgcD could be readily obtained from screening predatory cultures , to check whether these strains could grow axenically they were “turned HI” by filtration of attack-phase cells through a 0 . 45 µm filter to remove any remaining prey , and plating onto rich Peptone-Yeast Extract media [45] . The number of attack-phase cells of the ΔdgcC mutant had to be multiplied 1000 –fold ( to 4 . 5×1011 ) to obtain a single colony on the PY media . Complemented strains were tested by the same methods . The diguanylyl cyclase assays in vitro were performed by measuring the rate of GTP conversion into c-di-GMP by MBP-Dgc fusion proteins . GTP was added to the purified enzyme solution following the protocol by Ryjenkov et al . [48] . Each dignuanylyl cyclase reaction contained 5 micromoles of the MBP-Dgc fusion proteins . The equilibrium dialysis measurements of c-di-GMP binding to CdgA were performed using 10 micromole of MBP-CdgA and varying concentrations of c-di-GMP as described elsewhere [26] . Nucleotide separation and quantitation was accomplished by HPLC as described by Ryjenkov et al . [48] . The method of Bobrov [31] was used to determine the extractable levels of c-di-GMP in axenically grown Bdellovibrio cells . with analysis using liquid chromatography tandem mass spectrometry carried out by Lijun Chen and Bev Chamberlin at the Mass Spectrometry Core of RTSF ( Research Technology Support Facility ) at Michigan State University USA Axenically growing HI Bdellovibrio strains ( mutant and wild type HI ) were grown in 50ml of PY broth from a starting 0 . 2 OD600nm until they reached a final 0 . 6 OD600nm . Cells were then pelleted by centrifugation , the wet weights were determined and matched; they were frozen in liquid nitrogen and then processed using Bobrov's method and extraction buffer ( 40% methanol 40% acetonitrile in 0 . 1N formic acid ) which was later neutralised with NH4HCO3 . The levels of extractable c-di-GMP in the cell extracts were compared to known added standards of pure c-di-GMP . Time-lapse back-lit microscopy was performed as described fully by Fenton and co-workers [6] , with the exception that HI Bdellovibrio , to be tested for predatory capacity , were grown overnight in PY broth and then diluted to a starting OD600nm of 0 . 75 before addition to prey in the infection culture . Bdellovibrio strains containing GGDEF protein-mCherry fusions , in a wild-type GGDEF gene background , were constructed using a modification of the method described by Fenton and co-workers [49] , construction of each tag is described fully in Text S1 . Resulting strains were imaged using a Nikon Eclipse E600 epifluorescence microscope with a 100x objective lens and an hcRED filter block ( excitation 550 to 600 nm; emission 610 to 665 nm ) in conjunction with a Hamamatsu Orca ER camera and the Simple PCI software ( version 5 . 3 . 1 Hamamatsu ) . Gliding motility was assayed by timelapse microscopy as in Lambert and co-workers [2] . Bd0367 dgcA GenBank ID: NP_967362 . 1 Bd0742 dgcB GenBank ID: NP_967706 . 1 Bd1434 dgcC GenBank ID: NP_968330 . 1 Bd3125 cdgA GenBank ID: NP_969891 . 1 Bd3766 dgcD GenBank ID: NP_970474 . 1 | Bdellovibrio bacteriovorus is a tiny bacterium that preys upon other bacteria including pathogenic bacteria that cause infections in humans , animals , or crop plants . Bdellovibrio don't attack human , plant or animal cells and so could in future be used as “living antibiotics” . Here we have discovered , using genetics chemical analyses and microscopy , that proteins with a sequence in them called “GGDEF” control whether Bdellovibrio grow by preying upon other bacteria or whether they grow “normally” without attacking prey . The GGDEF proteins all synthesise the small signalling molecule cyclic- di GMP , but interestingly the production of this signal has different effects depending on which GGDEF protein makes it . If we remove one GGDEF protein this makes a Bdellovibrio that can't eat bacteria anymore and has to survive on environmental nutrients . Removing a different GGDEF protein gives Bdellovibrio that can only survive by eating prey bacteria such as pathogens- they lose the ability to eat “normal” nutrients . This is very useful when trying to produce Bdellovibrio as a therapy . The correct GGDEF mutant would have to “eat” pathogens only and couldn't grow using the nutrients present in the blood and serum of a wound , for example , so it would be a self-limiting treatment . | [
"Abstract",
"Introduction",
"Results/Discussion",
"Materials",
"and",
"Methods"
] | [
"biotechnology",
"systems",
"biology",
"medicine",
"infectious",
"diseases",
"drugs",
"and",
"devices",
"complementary",
"and",
"alternative",
"medicine",
"genetics",
"biology",
"microbiology",
"molecular",
"cell",
"biology",
"genetics",
"and",
"genomics"
] | 2012 | Discrete Cyclic di-GMP-Dependent Control of Bacterial Predation versus Axenic Growth in Bdellovibrio bacteriovorus |
Evidence of an adverse influence of soil transmitted helminth ( STH ) infections on cognitive function and educational loss is equivocal . Prior meta-analyses have focused on randomized controlled trials only and have not sufficiently explored the potential for disparate influence of STH infection by cognitive domain . We re-examine the hypothesis that STH infection is associated with cognitive deficit and educational loss using data from all primary epidemiologic studies published between 1992 and 2016 . Medline , Biosis and Web of Science were searched for original studies published in the English language . Cognitive function was defined in four domains ( learning , memory , reaction time and innate intelligence ) and educational loss in two domains ( attendance and scholastic achievement ) . Pooled effect across studies were calculated as standardized mean differences ( SMD ) to compare cognitive and educational measures for STH infected/non-dewormed children versus STH uninfected /dewormed children using Review Manager 5 . 3 . Sub-group analyses were implemented by study design , risk of bias ( ROB ) and co-prevalence of Schistosoma species infection . Influential studies were excluded in sensitivity analysis to examine stability of pooled estimates . We included 36 studies of 12 , 920 children . STH infected/non-dewormed children had small to moderate deficits in three domains—learning , memory and intelligence ( SMD: -0 . 44 to -0 . 27 , P<0 . 01–0 . 03 ) compared to STH-uninfected/dewormed children . There were no differences by infection/treatment status for reaction time , school attendance and scholastic achievement ( SMD: -0 . 26 to -0 . 16 , P = 0 . 06–0 . 19 ) . Heterogeneity of the pooled effects in all six domains was high ( P<0 . 01; I2 = 66–99% ) . Application of outlier treatment reduced heterogeneity in learning domain ( P = 0 . 12; I2 = 33% ) and strengthened STH-related associations in all domains but intelligence ( SMD: -0 . 20 , P = 0 . 09 ) . Results varied by study design and ROB . Among experimental intervention studies , there was no association between STH treatment and educational loss/performance in tests of memory , reaction time and innate intelligence ( SMD: -0 . 27 to 0 . 17 , P = 0 . 18–0 . 69 ) . Infection-related deficits in learning persisted within design/ROB levels ( SMD: -0 . 37 to -52 , P<0 . 01 ) except for pre-vs post intervention design ( n = 3 studies , SMD = -0 . 43 , P = 0 . 47 ) . Deficits in memory , reaction time and innate intelligence persisted within observational studies ( SMD: -0 . 23 to -0 . 38 , all P<0 . 01 ) and high ROB strata ( SMD:-0 . 37 to -0 . 83 , P = 0 . 07 to <0 . 01 ) . Further , in Schistosoma infection co-prevalent settings , associations were generally stronger and statistically robust for STH-related deficits in learning , memory and reaction time tests ( SMD:-0 . 36 to -0 . 55 , P = 0 . 003–0 . 02 ) . STH-related deficits in school attendance and scholastic achievement was noted in low ( SMD:-0 . 57 , P = 0 . 05 ) and high ROB strata respectively . We provide evidence of superior performance in five of six educational and cognitive domains assessed for STH uninfected/dewormed versus STH infected/not-dewormed school-aged children from helminth endemic regions . Cautious interpretation is warranted due to high ROB in some of the primary literature and high between study variability in most domains . Notwithstanding , this synthesis provides empirical support for a cognitive and educational benefit of deworming . The benefit of deworming will be enhanced by strategically employing , integrated interventions . Thus , multi-pronged inter-sectoral strategies that holistically address the environmental and structural roots of child cognitive impairment and educational loss in the developing world may be needed to fully realize the benefit of mass deworming programs .
Soil-transmitted helminthiases ( STH ) include infections with roundworm ( Ascaris lumbricoides ) , whipworm ( Trichuris trichiura ) and hookworm ( Ancylostoma duodenale and Necator americanus ) . STH infection affects one third of the world’s population [1 , 2] . Primarily due to poverty , poor personal hygiene , frequent outdoor exposures , a higher likelihood of indulging in high risk behavior such as soil-eating , and the presence of environmental conditions favoring transmission [3 , 4] , school-age children between the ages of 5 to 15 in mostly developing countries are at highest risk of chronic helminth infection and helminth-associated morbidities [2 , 4 , 5] . Of note , the designation of school-aged children as highest risk may be because there has been insufficient research of the neurodevelopmental and cognitive importance of STH in toddlers [6–8] and preschool children [6 , 9 , 10] . In this highly affected demographic , STH-related morbidities occur during critical periods of physiologic , mental and physical development . The chronicity of infection ensures that any STH-related small to moderate nutritional , growth and cognitive deficits is cumulated over extended periods of the developmental life course . Epidemiologic studies have reported lower health outcomes including- a higher prevalence of lethargy [11] , stunting , wasting and anemia [12 , 13] for helminth–infected relative to uninfected school-age children . In addition , some epidemiologic studies have reported a higher prevalence of school-absenteeism [13] , and lower performance on a range of cognitive tests [14] . The mechanisms of adverse effect on cognition and mental function of infected children are not well understood [15] but helminth-associated iron-deficiency is thought to be important [1 , 16 , 17] . Recent evidence suggests that helminth-associated morbidity and mortality is likely magnified for polyparasitized children [18–20] and those with chronic untreated moderate and heavy intensity of infections . Recent systematic reviews and meta-analysis of the health benefits of STH treatment based on intervention study designs concluded that the evidence for improvements in cognitive function was mixed and inconclusive and based on at best low quality evidence [15 , 21] . Conclusions from these meta-analyses raise fundamental questions about the expected educational and cognitive benefits that partly justify current deworming programs in helminth endemic regions . However , the following criticism have been made of their conclusions [22–26]: ( i ) existing trials included were of poor quality [25]; ( ii ) interpretation did not factor in realities such as the high frequency and rapidity of reinfection following treatment [22]; ( iii ) and the fact that by offering one time treatment combined with the short-term duration of most trials , they have limited ability to evaluate long-term benefits of regular deworming as envisioned and supported by the World Health Organization [5] , Unesco and the World Bank [23] . Proper assessment of long-term cognitive outcomes may require a sustained period of intervention and possibly post-treatment educational remediation in school-aged children [27–32] . Given the limitations of randomized trials noted above and a health policy environment that supports periodic deworming of school-aged children for expected growth , nutritional and perhaps cognitive benefits , clinical equipoise for unknown benefit of deworming is difficult to justify and the ethical landscape has coalesced around the understanding that randomizing children to deworming compared to placebo is largely unjust . Systematic reviews and meta-analyses remain an important tool for clarifying the possible cognitive impact of deworming . Including observational and intervention study designs in this synthesis provides robust scientific evidence to inform existing empirical gap in a way that enhances external validity beyond highly selected trial participants and conforms with the prevailing clinical , ethical and health policy environments that converge and contend with one another on the subject of deworming . Hence , we incorporate all available epidemiologic evidence and re-examine the hypothesis that helminth infections adversely affect child cognitive function and educational outcomes . Unlike prior reviews , we examine the impact of STH infection on specific cognitive domains–learning , memory , attention/reaction time and intelligence , and do not assume equal impact of STH on across domains . We specifically hypothesize that STH infections will be associated with educational deficits and lower performance in learning , memory and attentional processes because these domains of child function are more sensitive to environmental assaults . We define a fourth domain “innate intelligence” that we expect to be mostly genetically determined and possibly less affected by STH infection .
We performed a comprehensive search of databases including MEDLINE using PubMed ( http://www . ncbi . nlm . nih . gov ) , Science Direct and Google Scholar , as of April 22 , 2016 unrestricted by language . Throughout the data search , we used the key words , “soil-transmitted helminthiasis” and “cognitive” as well as "intestinal parasites" , "school performance" and "school attendance" . We included all studies–regardless of study design that had raw data for relationships between any of the 3 STH parasite species and educational indices ( attendance or achievement ) or neurocognitive outcomes . Articles were excluded if: ( i ) they were duplicates; ( ii ) completely lacked or incompletely presented needed raw data for relationships between STH and respective outcomes . We searched the body of included studies and their reference sections [32–35] for possible additional inclusions to this meta-analysis not identified by keyword searches . Our search yielded 7 , 709 citations and after a series of exclusions ( outcomes attributed to other than STH , absence of required data and presence of duplicate data ) , provided for the final pool of articles included in the meta-analysis which was [13 , 17 , 31 , 36–68] . We focused on the more recent publications ( from the year 1992 onwards ) in the hope of minimizing the high level of heterogeneity across studies prior to 1992 as noted by Watkins and Pollitt . [35] . These were categorized into four domains: ( i ) memory , ( ii ) reaction time , ( iii ) learning , and ( iv ) intelligence tests . Many studies used a suite of psychometric instruments to assess a single or multiple cognitive domains in enrolled children . The memory domain included tests of working ( short-term ) memory as well as those of long-term memory . Attention/reaction time tests were those that measured the ability of a child to sustain concentration on a particular object , action , or thought , including their capacity to manage competing demands in their environments . The learning/executive function domain included tests to evaluate children’s performance in goal-oriented behavior , particularly components that are important for scholastic advancement . Executive function included tests of cognitive processes that enable children to connect past experience with present action , and by so doing , engage in planning , organizing , strategizing , paying of attention to details , and to emotionally self-regulate , make necessary efforts to remember important details required for attainment of future goals . [69] We included in the ‘intelligence’ domain psychometric tests of intelligence quotient ( IQ ) that we believe measures largely biologically determined cognitive abilities , in contrast to cognitive performance measures that are environmentally pliable . [70] When multiple instruments were used to measure the same cognitive domain , a grand mean of scores and a grand mean of standard deviation ( SD ) across all instruments were calculated . Thus , for each publication , one overall mean and SD value was determined for each domain . A study could contribute data to different cognitive domains if it used tools spanning across several cognitive domains; however , each instrument only contributed to one single domain of function ( S1 Table ) . Overall effects were derived only if there were two or more studies in a given domain . Attendance rate was defined as the number of days children attended school over the past month ( in cross-sectional studies ) or over the study period ( in longitudinal studies ) . In case-control studies , the percentage of children enrolled versus not enrolled in school was calculated for STH-infected and non-infected children . The definition of scholastic achievement varied across studies . Children with high versus low attainment were identified based on: i ) pass rate on standardized teacher-made tests , ii ) percent of children that were in appropriate class for their age , iii ) the enrollment of children in elite versus non-elite schools , iv ) scores in the school function domain of pediatric quality-of-life inventory , v ) change in class position after treatment for STH infection , vi ) an above versus below average class performance as rated by a teacher , or vii ) pass rate in any kind of educational test , whether teacher administered or not . ( S1 Table ) . STH infection status was determined by microscopic examination of stool . Operationally , infection was defined based on study design as follows: 1 ) untreated/placebo versus Albendazole / Mebendazole / Dicaris-treated in a randomized controlled trial ( RCT ) , 2 ) any versus no STH infection in cross-sectional studies and 3 ) pre-versus post-Albendazole / Mebendazole / Dicaris treatment or infection-free versus persistent infection among STH-infected individuals in a longitudinal design study . Two investigators ( NP , AEE ) independently extracted data . Disagreements ( if any ) were resolved by a third person ( LT ) . If resolution was not attained by the third author alone , abstracted information was resolved by consensus between LT , NP and AEE . The following information was obtained from each publication: first author’s name , published year , country of origin , parasite ( s ) involved , age range of the subjects , effect outcome of the study , study design , study features as well as subject features and sample sizes . We followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses ( PRISMA ) and Meta-analysis Of Observational Studies in Epidemiology ( MOOSE ) guidelines in describing our findings and standard methodology [71–73] . We assessed study quality using the Cochrane Collaborations tool for assessing risk of bias in randomized trials . [74] To assess methodological quality of each observational study , we modified the Newcastle-Ottawa scale ( NOS ) [75 , 76] using the star * system based on the following: ( i ) representativeness of infected sample or case/control selection ( min = 0 , max = 3* ) ; ( ii ) comparability using known correlates of cognitive function/educational attainment ( min = 0 , max = 6* ) . Here , we accounted for the confounding effects of age ( score = 1* ) , sex ( score = 1* ) , nutritional ( score = 2* ) and socioeconomic ( score = 2* ) status in relation to STH infection and educational/ cognitive outcomes . ( iii ) Absence of bias in relation to outcome assessment in prospective cohort studies ( min = 0 , max = 3* ) or exposure assessment in cross-sectional and case-control studies ( min = 0 , max = 3* ) . For each study , the initial raw quality score ( max = 12* ) was rescaled to match the scale of 9* and then classified as low , high or very high risk of bias per prior literature precedent [76] . Data were analyzed using Review Manager 5 . 3 ( Copenhagen: Nordic Cochrane Centre , Cochrane Collaboration , 2014 ) and SigmaStat 2 . 3 and SigmaPlot 11 . 0 ( Systat Software , San Jose , CA ) . We sought the number of children on two levels: ( i ) those who took the cognitive tests and compared those who were infected with those who were not; and ( ii ) those monitored for school achievement and attendance . We expressed these relationships as standardized mean difference ( SMD ) and 95% confidence intervals ( CI ) . SMD estimates were classified as statistically significantly different if their confidence intervals did not cross zero . Most the included articles ( studies ) presented multiple means and measures of spread ( SD or standard error ( SE ) ) . From each article , we obtained the mean of the multiple means and SDs which we used as input in generating forest plots . However , Some papers presented median ( m ) and range ( a and b ) . These measures were converted into approximate mean and SD following Hozo et al [77] . Where mean and 95% CI or SE were reported , SD were derived as described by the Cochrane Collaboration [78] . For studies presenting data on differences in mean scores between two time points for treated/infected versus untreated/uninfected groups , appropriate SD for mean difference was calculated per the approach described by the Cochrane Collaboration [78] . Pooled SMDs were obtained using two analysis models based on presence or absence of heterogeneity: the fixed [79] and random [80] effects , respectively . Heterogeneity refers to diversity which may influence the manner in which the data are treated [81] . Significance was set at a P-value of <0 . 05 . We addressed heterogeneity between studies with the following: ( i ) estimated using the χ2-based Q test [82]; ( ii ) its sources identified using the Galbraith plot method [83] and ( iii ) explored using subgroup analysis [82] where we examined effects in observational and interventional studies . Intervention studies , often prospective , are specifically tailored to evaluate direct impacts of treatment or preventive measures on disease . Observational studies on the other hand are often retrospective and used to assess potential causation in exposure-outcome relationships and therefore influence preventive methods [84] . In addition , we also examined effects according to risk of bias ( low and high ) , all trial and pre-post studies as well Schistosomiasis coinfection . Influence of each study on robustness of the summary effects and heterogeneity was determined with sensitivity analysis , which involved omitting one study at a time followed by recalculation of the pooled SMD . Change in direction of association ( e . g . poor performance to better performance and vice-versa ) after study omission indicates non-robustness of the summary effect , otherwise the pooled SMD is considered robust , indicating stability of the results . We investigated publication bias in domains with ≥ 10 studies because with < 10 , sensitivity of the qualitative and quantitative tests of comparisons studies was low [85] . We implemented sensitivity analysis based on observational or experimental study design . Available data from studies that provided baseline treatment of children and post-test readings of the same subjects were included as intervention studies .
We included 36 studies ( Fig 1 ) of 12 , 920 children ( 5-20y ) that evaluated STH associated differences in psychometrically evaluated cognitive tests , educational attainment or school attendance ( Table 1 ) . Of these , 5 , 932 children were evaluated in the context of STH treatment ( with or without randomization ) and 6 , 978 were evaluated as part of observational study . In the latter , 5 , 538 and 586 children were part of cross-sectional and longitudinal studies , respectively . The remaining 854 subjects were part of comparative Epidemiologic surveys . The PRISMA checklist is included with further pertinent details for this meta-analysis ( S3 Table ) . Table 2 ranks included observational studies in terms of methodological quality using the NOS . Fourteen ( 56% ) of these were determined to have low risk of bias and 11 studies ( 44% ) were classified as having high risk or very high of bias . The ten intervention studies with featuring allocation of trial participants to deworming ( alone or in combination ) versus control groups were communicated across 11 publications . Seventy percent of trials had moderate or high risk of bias and thirty percent were uncertain to low risk of bias ( Table 3 ) . S2 Table shows that the vast majority of included studies examined educational or cognitive impact infection with multiple STH species ( up to 62 . 5% ) . Ascaris infections were the least studied ( none in three domains ) . Nineteen studies ( 52 . 7% ) included in this meta-analysis used Albendazole or Mebendazole for STH treatment with the exception of Boivin et al [41] where Decaris ( Tetramisole , Lavamisole ) was used . As shown in Table 4 , the overall analysis shows that STH-infected/non-treated children performed significantly worse than uninfected/dewormed children in the three of the six domains including: memory ( SMD: -0 . 31 , P = 0 . 01 ) ; learning ( SMD: -0 . 44 , P <0 . 0001 , Fig 2 ) and intelligence ( SMD: -0 . 27 , P = 0 . 03 ) . Because pooled effects in all six domains were obtained under highly heterogeneous conditions ( Pheterogeneity <0 . 01 , I2 = 66–99% ) , we identified investigations contributing to large variability in the pooled effect of STH-infection/non-treatment on performance in each of the cognitive and educational domains using the Galbraith plot technique ( Fig 3 ) . This technique did not significantly reduce heterogeneity in respective outcomes with the exception of the learning domain ( I2 = 33% , Fig 4 ) . However , application of this technique resulted in emergence of statistically robust estimates of effect in three domains: reaction time ( SMD: -0 . 21 , P = 0 . 004 ) , achievement ( SMD: -0 . 24 , P <0 . 01 ) and attendance ( SMD: -0 . 52 , P <0 . 01 ) . On the other hand , exclusion of one influential study ( Shang et al [61] ) resulted in loss of statistical significance in intelligence domain ( SMD: -0 . 20 , P = 0 . 09 ) . Stability of overall findings were evaluated in sub-groups distinguished by study design , presence of non-outlier studies and study quality ( Table 5 ) . In the context of observational study designs , STH infection/non-treatment was significantly and consistently associated with deficits in memory , learning , reaction time and intelligence with SMD ranging from -0 . 42 to -0 . 23 ( P <0 . 01–0 . 05 ) . Among studies featuring STH treatment- pre-post longitudinal study or experimental in design , moderate ( SMD: - 0 . 8 , P = 0 . 14 ) to large ( SMD: -0 . 46 , P<0 . 0001 ) deficits in memory and learning tests were respectively noted . was evident for untreated children . With the exception of the learning domain ( SMD = -0 . 40 , P<0 . 0001 ) , there was no association between non-treatment for STH and performance in any of the other five outcome domains for pooled estimates of experimental study designs only ( SMD -0 . 27 , 0 . 27 , P = 0 . 18–0 . 69 ) . Among observational studies featuring pre-post STH assessment of cognitive or educational outcomes , infection related deficits was evident in test of reaction time only ( SMD -0 . 37 , P<0 . 001 ) with corresponding substantial reduction in heterogeneity ( Pheterogeneity = 0 . 72 , I2 = 0% ) . Among studies classified as low risk of bias , STH infection/non-treatment was associated with significant deficits in memory , learning and school attendance ( SMD: -0 . 39 to -0 . 55 , P <0 . 0001–0 . 03 ) . Among studies classified as high or very high risk of bias statistical significance was achieved or maintained in the domains of learning , reaction time and intelligence ( SMD: -0 . 28 to -1 . 00 , P <0 . 001–0 . 03 ) as well as achievement ( SMD: -0 . 70 P <0 . 001 ) . To determine whether Schistosoma species co-infection to contributed to observed associations between respective outcomes and STH infection vs . no infection among observational studies , separate pooled estimates were derived for primary studies with and without co-prevalence of Schistosoma species ( Table 6 ) . STH infection related deficits in learning ( SMD = -0 . 32 , P = 0 . 02 ) and memory ( SMD = -0 . 32 , p = 0 . 001 ) was evident in pooled analyses of primary studies without co-prevalent Schistosoma species infection . Among observational studies with co-prevalent schistosoma co-infection , magnitudes of association were generally stronger and pooled associations were statistically for infection related deficits in tests of learning , memory and reaction time ( SMD -0 . 36 to -0 . 55 , P = 0 . 003–0 . 02 ) . Table 7 presents a summary of the impact of deleting individual publications on robustness of pooled SMD in overall analyses and within strata of design , outlier study status and study quality . Of the six domains , reaction time was most robust followed by learning . In most sub-group analyses ( post-outlier , observational , high risk of bias and those without Schistosomiasis coinfection ) pooled SMDs were similar in magnitude and direction ( i . e . robust ) in comparison with SMDs from overall analyses . Of the 47 robust estimates , 27 ( 62 . 8% ) provided statistically robust associations between STH infection/non-treatment and cognitive deficits or educational loss . We found no evidence of publication bias in all six outcome domains for the overall findings and subgroups with ≥ 10 studies ( Table 8 ) .
In line with our hypothesis , we found that non-treatment for STH infection was consistently associated with statistically , clinically and health policy relevant deficits in five of the six domains examined . The relationships observed were generally invariant to exclusion of influential individual studies but robustness of findings varied by study design–observational , STH treatment studies and experimental studies , and study quality . Specifically , infection related deficits were larger in magnitude , tended to be more statistically robust and stable in direction for pooled studies of observational compared to STH treatment ( with or without experimental ) studies and for high vs . low risk of bias studies . The direction of pooled effects from observational studies with and without potential for Schistosoma species coinfection were similar although the magnitude of associations tended to be higher for STH-related deficits in settings with potential for Schistosoma species coinfection . These non-treatment/STH-infection related disadvantages in cognitive function and educational measures correspond to small to moderate deficits per the Cohen criteria [88] . Of note , these small to moderate deficits are for a highly prevalent exposure affecting millions of children in the developing world . Because reinfection is rapid in helminth endemic settings [45] , many children are effectively chronically infected by these parasites and the majority of children are simultaneously polyparasitized–i . e . infected by two or more STH species at the same time [19] . Polyparasitic and multi-species infections may produce additive disadvantages for cognitive function and educational loss . Our finding of stronger pooled effects in Schistosoma prevalent settings is suggestive evidence in support of this thesis . Hence , we speculate that the modest deficits reported here may under-estimate the true magnitude of learning and memory deficits attributable to untreated STH infections in endemic regions . Working memory and learning tasks involve the brain’s frontal lobe [89] which continues to develop throughout school-age years [90] . Drake et al in prior systematic reviews [18 , 32] noted the lack of consistency in cognitive domains affected by helminth infections . The one exception was in STH-infection associated lower performance on tests of working memory [32] . Our review lends credence to that previous observation for working memory and provides new evidence of STH infection related deficits in learning tests . Also in line with our study hypothesis , we found that untreated STH infection was associated with slower reaction time . The association between STH infection/non-treatment and attentional processes/reaction time was evident in observational studies and persisted among intervention studies when two influential intervention studies were excluded . These suggest that STH infections may have an adverse effect on attention and/or cognitive speediness ( i . e . reaction time ) . Two findings from this meta-analyses were either not consistent with our study hypothesis or were altogether surprising . The finding that school attendance and scholastic achievement did not differ significantly in overall analyses by infection/treatment status , though similar to results from prior meta-analyses on this subject , [21 , 34 , 91–93] was contrary to our study hypothesis . However , we noted that the overall findings were not stable . With exclusion of influential studies and design stratified analyses we noted suggestive evidence of an adverse impact of infection/non-treatment on these outcomes among observational studies with high potential for bias but not among intervention or low risk of bias studies . Of note , scholastic achievement , as operationally defined in this study , includes children’s pass rate on standardized or teacher-made tests , the percent of children that were in appropriate class for age , child enrollment in elite versus non-elite schools , scores in the school function domain of pediatric quality-of-life inventory , change in class position after treatment , an above average versus average/below average teacher rating of scholastic performance and pass/fail rate in any kind of educational test , whether teacher administered or not . Hence , our finding of significant adverse effect of STH infection on this domain could have broader implications for the educational achievement in complex real-world settings . The effect on achievement likely reflect to varying degrees deficits in cognitive domains such as learning , memory and reaction time [94] . The observation of infection related deficits for tests of innate intelligence was surprising because we conceptually conceived innate intelligence as mostly a “genetically determined finite quantity” relatively insensitive to postnatal environmental assaults such as STH infections . Of note , this finding was robust only among observational studies and those classified as “high risk of bias” . Furthermore , the impact of infection/non-treatment for STH on innate intelligence became statistically insignificant with the exclusion of an influential study suggesting the need for abundant caution in the interpretation of this data . Here , we have classified neurocognitive performance assessed with variety of instruments in the primary literature into four major cognitive domains based on extant knowledge of the primary cognitive domains covered by respective tools . We supplemented as needed with content area guidance from a neuropsychologist . Innate intelligence included measures of general intelligence based on IQ tests such as the Philippine non-verbal intelligence test , Kaufmann assessment battery for children , Wechsler intelligence test for children ( where overall scores and no subscales are specified ) . Prior reviews [32] have eloquently described the inherent challenges with neurocognitive assessments that complicates empirical efforts to understand the impact of helminth infections on disparate domains . For example , performance in respective domains likely correlate with one another to varying degrees in most individuals . Hence , it is possible that the STH-associated deficits in innate intelligence measures partly reflect STH-related effect on other domains where performance may be more sensitive to environmental factors including educational quality and health factors [95] . For this systematic review and meta-analysis , we have intentionally taken the approach of combining assessment instruments and educational measures into one of six pre-identified domains that contribute in varying degrees to the ability of children to learn , be educated and potentially their future productivity as adults . These classifications are guided by the literature and underlying theoretic constructs assessed . In spite of our best efforts , this remains an imprecise science . We provided extensive detail of our classification approach as supplemental material for critical evaluation and further refinement . In sum , the high heterogeneity [96] , limited statistical power in sub-group analyses [95 , 97] and the potential for residual confounding- particularly in non-randomized interventions and observational studies , [98] should lead to cautious interpretation of findings from sub-group and sensitivity analyses . Our review is subject to several limitations that should be considered in the interpretation of our findings . Firstly , heterogeneity was high across studies . We analytically addressed this using random-effects model and explored the specific role of influential studies using the Galbraith plot method . Secondly , as noted earlier , our overall findings were sensitive to observational versus intervention study design and study quality . Thirdly , we are unable to fully delineate potentially STH species-specific differences in the domains evaluated as studies did not always distinguish between STH species . Additionally , across individual studies , the number of stool samples used for diagnosis of STH infection varied from single to multiple . Where single samples were tested , the chance of misclassifying lightly infected children increases . Thus in some studies , “the uninfected” may include an unknown number of mostly lightly infected children . Similarly , information on joint Schistosoma and STH infections as well as data on STH infection intensity was not consistently provided making it difficult to robustly examine potential for beyond additive adverse effects for multi-species infected children and dose-response by infection intensity . Last but not least , variations in the primary literature did not allow for subgroup analyses by the following factors–STH species , malnourished versus normally nourished children , type of deworming agent , and treatment strategy ( mass deworming versus test and treat only infected children ) . These subgroup analyses–if possible , could give insight with respect to the kinds of children most likely to derive a cognitive benefit from deworming and whether mass or targeted deworming is the best approach [99] to deworming school-aged and potentially preschool children [100] . Potential for heterogeneity in the association of deworming with cognitive and educational outcomes by child nutritional status , treatment strategy , deworming agent and parasite species requires further elucidation . In light of considerable heterogeneity , varying quality of underlying studies and sensitivity to experimental versus observational study design in some outcomes , we agree with prior expressed need for appropriate caution in the interpretation of findings from synthesis of epidemiologic literature especially where associations are primarily driven by observational studies and the risk of bias in underlying studies is high or very high [101] . In spite of these limitations , our study has a number of important strengths that allow us provide additional context for understanding the influence of study design , specific influential studies and risk of bias within present literature and STH-associated cognitive and educational deficits . We address two issues noted by the recent Campbell review- i . e . the use of a complicated array of different cognitive tools and a lack of understanding of STH impact on absenteeism [101] . We enhance clarity regarding the possible impact of STH infection/non-treatment on educational and cognitive loss by systematically collapsing the variety of cognitive tools used within four cognitive domains and conducting specific analyses on educational loss–including scholastic achievement and absenteeism . Our approach implicitly recognizes STH-infections may have unequal impact in different domains despite potential over-lap in abilities tested within respective domains . It is noteworthy that in spite of major differences in our empirical approach relative to prior systematic reviews and meta-analyses on this subject , our finding of no STH-infection related deficits for most cognitive and educational outcomes in sub-group analyses restricted to experimental study designs confirms previously reported findings based on prior systematic reviews focused on RCTs only . [21 , 34 , 91–93] This meta-analysis allows for evaluation of consistency in primary relations investigated across key factors by including all relevant epidemiologic studies regardless of design , conducting explicit analyses to evaluate impact of influential studies , and evaluating the risk of bias in underlying studies . It has been noted that carefully implemented meta-analyses based on observational studies generally produce estimates similar to those from meta-analyses based on RCT and thus supports evidence-based medical decision-making [102 , 103] . Our approach provides empirical evidence to evaluate an important health policy and yet allows allows for appropriate qualification as warranted on the basis of design , influential studies and risk of bias . We provide evidence of small to modest deficit in five of six evaluated domains although there were few influential studies and variations for associations existed by study design . Prior meta-analyses of RCTs that evaluated cognitive impacts of STH infections had different conclusions about the cognitive and scholastic effects of STH infection [15 , 21 , 34] . Key differences in our approach–namely inclusion of all relevant epidemiologic studies and evaluation of effects within educational and cognitive domains , may partly account for the inferential difference between this and prior reviews . Despite the empirical debate regarding the cognitive benefit of deworming for STH , the current ethical , clinical and health policy environments remain strongly skewed in favor of deworming for child growth , prevention of anemia and potentially avoidance of preventable cognitive deficits . Deworming efforts have expanded as a strategy to control the prevalence and intensity of infections . Ongoing operational intervention research is poised to provide valuable insight regarding the optimal approach-whether school based , community based or a combination of infection control via mass deworming will additionally examine the net benefit of bi-annual versus annual deworming [104] . Mass deworming campaigns are unlikely to interrupt infection in most settings , but infection profile is expected to shift towards low-intensity single and multi-species co-infections . Prior research has demonstrated that polyparasitism is consequential for anemia- an important determinant of cognitive deficit in children [19 , 105] . Specific investigations of the cumulative impact of polyparasitic STH infection on cognitive and educational outcomes are lacking . Future epidemiologic studies that explicitly examine the cognitive impact of multi-species parasitic infections by number as well as by intensity of coincident infections will provide useful information to guide health policy and may inform the optimal frequency of deworming in the context of low intensity infections . Currently , preschool-age children are not treated as part of routine deworming programs for STH [106] , and yet , the evidence suggests that children are infected shortly after weaning and remain persistently infected throughout childhood and adolescence . The adverse impact of infection on children begins way before school age and compounds the cumulative health disadvantage associated with STH infection . This meta-analysis , and indeed most short-term study outcomes meta-analyzed , does not include this demographic . Recent review of investigations have demonstrated the safety and efficacy of STH deworming in preschool children [107] . Existence of an adverse developmental impact of STH infection on cognitive/educational domains would justify expanding the age-bracket of children who should be subjected to STH deworming . Educational and cognitive interventions will likely be more effective if initiated earlier in life for STH-infected children . However , more investigations among preschool children may be needed to understand the risks and potential benefits of early deworming in these children . Given the chronicity of infection during childhood and adolescence in helminth endemic regions , it is possible that any cognitive and educational loss associated with STH infection will not be resolved by deworming alone without prevention of reinfection and specific interventions to address the environmental and structural determinants of parasitic infections . Thus , effective future interventions are likely to be those that emphasize multi-pronged inter-sectoral strategies to holistically address challenges to child wellbeing in the developing world[108] . | Previous systematic reviews of the effect of STH infection on cognitive and educational performance were either inconclusive or found little to no evidence of associated benefit . Lack of consensus in prior reviews and their narrow emphasis on randomized controlled trials prompted this comprehensive assessment of whether educational and cognitive benefits are associated with the common practice of preventive chemotherapy among children in STH-endemic regions . This literature synthesis included all relevant epidemiologic studies regardless of design and investigated associations between STH infections and two outcomes: i ) educational loss ( attendance and scholastic achievement ) and ii ) psychometrically performance in four neurocognitive domains—memory , learning , reaction time and intelligence . Pooled results across 36 observational and intervention studies suggest that STH infection/non-treatment is associated with deficits in five of six domains evaluated and provide evidence in support of a cognitive and educational benefit of deworming for STH infections . This overall finding was sensitive to study design and risk of bias across studies . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"learning",
"children",
"cognitive",
"neurology",
"medicine",
"and",
"health",
"sciences",
"education",
"sociology",
"tropical",
"diseases",
"social",
"sciences",
"parasitic",
"diseases",
"neuroscience",
"learning",
"and",
"memory",
"pediatrics",
"age",
"groups",
"cogni... | 2018 | Soil-transmitted helminth infection, loss of education and cognitive impairment in school-aged children: A systematic review and meta-analysis |
An important goal in molecular evolution is to understand the genetic and physical mechanisms by which protein functions evolve and , in turn , to characterize how a protein's physical architecture influences its evolution . Here we dissect the mechanisms for an evolutionary shift in function in the mollusk ortholog of the steroid hormone receptors ( SRs ) , a family of biologically essential transcription factors . In vertebrates , the activity of SRs allosterically depends on binding a hormonal ligand; in mollusks , however , the SR ortholog ( called ER , because of high sequence similarity to vertebrate estrogen receptors ) activates transcription in the absence of ligand and does not respond to steroid hormones . To understand how this shift in regulation evolved , we combined evolutionary , structural , and functional analyses . We first determined the X-ray crystal structure of the ER of the Pacific oyster Crassostrea gigas ( CgER ) , and found that its ligand pocket is filled with bulky residues that prevent ligand occupancy . To understand the genetic basis for the evolution of mollusk ERs' unique functions , we resurrected an ancient SR progenitor and characterized the effect of historical amino acid replacements on its functions . We found that reintroducing just two ancient replacements from the lineage leading to mollusk ERs recapitulates the evolution of full constitutive activity and the loss of ligand activation . These substitutions stabilize interactions among key helices , causing the allosteric switch to become “stuck” in the active conformation and making activation independent of ligand binding . Subsequent changes filled the ligand pocket without further affecting activity; by degrading the allosteric switch , these substitutions vestigialized elements of the protein's architecture required for ligand regulation and made reversal to the ancestral function more complex . These findings show how the physical architecture of allostery enabled a few large-effect mutations to trigger a profound evolutionary change in the protein's function and shaped the genetics of evolutionary reversibility .
Understanding the mechanisms by which protein functions evolve is a major goal in studies of molecular evolution . A question of particular interest is how the biophysical architecture of a protein shapes its evolutionary potential . This question is a specific form of a general issue long-discussed by evolutionary biologists: whether and how the structure of complex biological systems – the process of organismal development or physiology , for example – influences evolution's capacity to modify those systems and produce new phenotypes [1]–[3] . This issue is typically discussed in terms of constraints on evolution , which are conceived as limits placed on the types of phenotypic variation that can be produced through mutation . In principle , however , the structure of biological systems could also act positively , facilitating the evolution of certain new phenotypes via simple genetic changes [4] . Proteins , although small in scale , are also complex biological systems , because their functional behavior is determined by specific interactions between large numbers of differentiated elements – in this case , thousands of atoms in three-dimensional space . How a protein's structure determines its evolutionary potential has been studied to only a limited extent . Prior work on suggests general patterns of constraint imposed by structure and function: residues in the protein core are generally less amenable to substitution than those on the surface [5]–[7] , and interacting residues in proteins can exert site-specific constraints on each other [7]–[9] . To understand how structure affects the evolution of new functions , however , it is necessary to directly trace the mechanisms by which functional changes occurred during historical evolution . Vertical approaches that experimentally dissect historical evolution through time are particularly useful , because they elucidate the genetic and structural mechanisms by which historical shifts in protein function took place . In cases of very recent functional evolution , population genetic methods can be used to identify which sequences , functions , and structures are ancestral and which are derived [10]–[12] . For more ancient divergences , ancestral protein reconstruction ( APR ) allows the functions and structures of ancient proteins to be experimentally characterized and provides a platform for identifying the historical mutations that mediated shifts in these properties [13] , [14] . In only a few cases have historical shifts in function been analyzed using experimental genetic and structural analysis of ancestral proteins [4] , [15]–[18] , although several additional studies have attempted to model ancestral structures based on their extant descendants [19]–[22] . In virtually all of these cases , the shift in function under study was a relatively subtle change in specificity . How fundamental changes in function evolve , such as the gain/loss of a biochemical activity or mode of regulation , remains largely unstudied ( but see ref . [19] ) . An ideal group of proteins for studying the structural mechanisms of functional evolution would be functionally diverse , contain adequate phylogenetic signal for reconstructing ancestral protein sequences and their historical mutational trajectories , and have well-developed methods for functional and structural characterization . The steroid receptor ( SR ) protein family fulfills these criteria in general [16]–[18]; in this paper , we focus on the mechanisms for a lineage-specific change in the mode of allosteric regulation by ligands in the SR ortholog of mollusks . In vertebrates [23] and some invertebrates [24] , [25] , members of the SR gene family are hormone-activated transcription factors , which regulate developmental , reproductive , and physiological processes . The protein's ligand-binding domain ( LBD ) serves as an allosteric switch controlled by the hormone , which binds in an internal hydrophobic cavity deep in the protein's LBD; ligand binding shifts the domain's thermodynamic equilibrium from the inactive conformation when ligand is absent to the active conformation when ligand is bound . In the inactive conformation , a C-terminal “activation-function” helix ( AF-H ) is disordered or extended away from the rest of the protein . In the active conformation , AF-H packs against the body of the protein , contributing to the assembly of a new surface that attracts coactivator proteins that alter chromatin or otherwise potentiate transcription of nearby target genes . Ligand binding stabilizes the position of the other helices against which AF-H packs and thus increases the stability of the active conformation relative to the inactive conformation [26] , [27] . In contrast to vertebrate SRs , mollusks contain a single SR ortholog that is unique in being a ligand-independent constitutive transcriptional activator . ( Because the mollusk receptors are most similar in sequence to the estrogen receptors of vertebrates , they are commonly referred to as ERs , although the phylogenetic analyses reported in this paper indicate that they are equally orthologous to the entire clade of vertebrate SRs . ) The ligand-binding domains from diverse mollusk ERs , including those of the sea slug Aplysia californica , the cephalopod Octopus vulgaris , the clam Thais clavigera , and the oyster Crassostrea gigas , have all been shown experimentally to activate transcription at high levels in the absence of any added ligands; further , they do not bind estrogens , and no increase in transcriptional activation is observed when the receptor is treated with hormones or other substances [28]–[31] . This constitutive function is thought to be evolutionarily derived , because the LBDs of other invertebrate SRs – including two from species in the closely related phylum of annelid worms , as well as the cephalochordate Branchiostoma floridae – lack constitutive activity and can be activated by addition of estrogens [24] , [25] . Further , phylogenetic reconstruction , synthesis , and experimental characterization of the ancestral gene from which the entire SR family descends ( AncSR1 ) showed that AncSR1 was estrogen activated with very little constitutive activity [28] , [32] . The mechanisms by which the mollusk ERs' constitutive transcriptional activity evolved are unknown . No structures of these proteins are available , and there has been no genetic or evolutionary work identifying the key substitutions that confer on these proteins their unique functions . Numerous questions are therefore unanswered: whether this derived function required many or few mutations , whether it was brought about by additive contributions from historical sequence changes or by a complex of epistatically interacting mutations , and what sort of remodeling of the protein structure was required to confer it . Constitutive activity has been observed in some distantly related members of the nuclear receptor superfamily; comparing crystal structures among constitutive and non-constitutive receptors suggests that the underlying mechanisms are diverse . In some cases , the transcriptionally active conformation appears to be stabilized in the absence of ligand by the acquisition of bulky residues that fill the hydrophobic cavity [33] , [34] . In others , electrostatic interactions between side chains or improved packing interactions between structural elements in the protein appear to stabilize the active conformation [35] , [36] . In still others , an omnipresent ligand fills the pocket , causing activity even when no exogenous ligand is added and suggesting ligand-independence before this structural information was available [37] . No studies , however , have identified the historical mutations that caused constitutive activity to evolve . Here we characterize the evolution of the mollusk ER's lineage-specific function by combining structural and genetic analysis of the constitutive ER of the oyster Crassostrea gigas with experimental reconstruction , manipulation , and characterization of ancestral proteins in the lineage leading to mollusk ERs . This combination of structural and evolutionary genetic approaches allows us to analyze in detail the mechanisms by which receptor function evolved and how the protein's structure shaped its functional and genetic evolution .
To understand the mechanisms by which constitutive activity evolved in the Crassostrea gigas ER ( CgER ) , we used X-ray crystallography to determine the three-dimensional structure of its LBD at a resolution of 2 . 6 Å in the absence of any added ligands ( Table S1 , Fig . 1A ) . The domain is in the classic active conformation , and the backbone conformation is similar to that of the human ERα ( 1 . 36 Å RMSD for all atoms ) . Unlike other SRs previously studied , there is no electron density of a ligand in the interior of the CgER LBD ( Fig . 1B ) . Rather , the internal cavity where ligand binds in other steroid receptors is occupied in CgER by several bulky hydrophobic side-chains , including F425 and F525 , and – to a lesser extent – F524 ( using human ERα numbering to facilitate comparison ) . These hydrophobic residues would strongly clash with estradiol as it is oriented in the human ERα LBD ( Fig . 2 ) . The resulting cavity has a total volume ( 168±8 Å3 ) much smaller than that of human ERα ( 402 Å3 ) and too small to accommodate estradiol ( 245 Å3 ) and other steroids . This structure has several implications . First , it indicates that CgER is an authentically ligand-independent transcriptional activator , which exists in the active conformation in the absence of ligand or other apparent modifications . If estrogen-like compounds , endogenous or environmental , affect mollusk reproduction and physiology , as has been reported previously [38]–[42] , our findings indicate that these impacts must be mediated by mechanisms other than ER activation . Second , the presence in the protein interior of bulky residues that occlude the ligand cavity provides a physical rationale for the inability of mollusk ERs to bind or be activated by ligands [28]–[31] . Although we cannot rule out the possibility that some unknown substances might be bound by the mollusk ER , the extremely small size of the cavity when the protein is in the active conformation suggests that if any such ligands exist , they would have antagonist rather than agonist effects on transcriptional activation . Third , the structure confirms that the constitutive activity of mollusk ERs is a derived evolutionary character , because no other members of the steroid receptor family have filled ligand cavities . A more distantly related clade within the nuclear receptor superfamily – the ERRs – are constitutive activators with partially filled internal cavities , but the residues that fill the pocket are at different sites in the sequence [34] , [43] . The CgER structure alone is insufficient to determine the mechanisms by which constitutive activity evolved . Ligand-independent activity may have been caused by the mutations that filled the receptor's ligand pocket; alternatively , it may have been caused by genetic changes that stabilized the active conformation by different mechanisms , followed by the substitutions that occluded the cavity . To identify specific historical substitutions that caused constitutive activity to evolve in the mollusk ERs , we identified candidate mutations that occurred during the historical interval in which constitutive activity emerged and tested them in the context of the relevant ancestral sequence . We began by reconstructing the amino acid sequence of the SR protein as it existed in the last common ancestor of the various lophotrochozoan phyla , including annelids and mollusks . We aligned 135 present day receptor sequences from a wide range of invertebrates and vertebrates ( Table S2 ) , determined the best-fit evolutionary model , and inferred the maximum likelihood phylogeny ( Fig . 3A , Fig . S1 ) . This phylogeny is largely consistent with the results of previous analyses [18] , [28] , [29] , except that in this case the lophotrochozoan ERs are not a sister group to the chordate ERs but are instead placed outside of all chordate SRs , including the vertebrate ERs , androgen receptors ( ARs ) , progestagen receptors ( PRs ) , glucocorticoid receptors ( GRs ) , and mineralocorticoid receptors ( MRs ) . This topology is more parsimonious than the previous one , because it implies a single receptor in the ancestor of all bilaterally symmetric organisms , followed by the minimum possible number of gene duplications – all in the chordates – and no subsequent losses . In contrast , placing the lophotrochozoan ERs as sister to the chordate ERs requires an additional earlier gene duplication and a subsequent loss of the AR/PR/GR/MR group from the lophotrochozoans . The present analysis is also based on more complete taxonomic sampling than previous efforts and therefore represents the best current hypothesis of SR phylogeny . To test the hypothesis that ligand-independent activation evolved on the branch leading to the ancestral mollusk ER , we inferred the maximum likelihood sequence of the ancestral lophotrochozoan SR ( AncLophoSR ) ; this is the reconstructed sequence at the node in the phylogeny that represents the SR ortholog in the last common ancestor of all extant lophotrochozoans ( Fig . 3 ) . AncLophoSR is 58 . 4% identical to CgER-LBD , and 57 . 9% identical to the Human ERα ( Fig . S2 ) . Support for the reconstruction was only moderate ( Fig . S3 ) : the mean posterior probability over all sites was only 0 . 73 , with 35 sites with a plausible alternate reconstruction ( defined as a second-best state with PP>0 . 25 ) ( Table S3 ) . At the 16 sites that line the internal cavity where ligands bind , however , confidence is a higher , with a mean PP of 0 . 91 and only one ambiguously reconstructed site . We tested the ligand-dependence of the AncLophoSR LBD by expressing it as a Gal4-DBD fusion protein and characterizing its transcriptional activity using a reporter gene assay in transfected cultured cells . In contrast to the constitutively active mollusk ERs , AncLophoSR had virtually no ligand-independent transcriptional activity , and it exhibited a clear dose-responsive increase in activity as estrogen concentrations were increased . AncLophoSR is highly sensitive to estradiol with an EC50 of 12 nM , although its maximal activation in this assay is lower than that of the human ERα [29] ( Fig . S4 ) . We next sought to determine the sensitivity of this result to uncertainty about the ancestral sequence . The number of sites with alternative plausible reconstructions is too large for us to test them all individually . We therefore synthesized a radically different version of AncLophoSR containing all 35 plausible alternative amino acids . Although this sequence is far less likely to be correct than the ML reconstruction ( with a likelihood that is 2 million times lower ) , it represents the “far edge” of the cloud of plausible ancestral reconstructions and allows a very conservative test of the robustness of inferences about the ancestral protein's functions . When expressed and assayed , this sequence was also estrogen-activated , although its baseline activity was somewhat higher than that of AncLophoSR-ML ( Fig . S5 ) . This result indicates that the estrogen-sensitivity of AncLophoSR is robust to statistical uncertainty about the ancestral reconstruction . Taken together , these findings corroborate the conclusion , supported by other lines of evidence , that ligand-independence evolved in the mollusks from an estrogen-sensitive ancestral state [18] , [24] , [25] , [28] . To identify the historical sequence changes that caused the evolution of constitutive activity and the loss of ligand regulation , we combined phylogenetic and structural analysis . These events must have occurred on the branch of the phylogeny leading from AncLophoSR to the ancestral mollusk ER , because all mollusk ERs are constitutively active and ligand-insensitive [28]–[31] . Seventy-nine amino acid replacements took place on this branch . To narrow down the list of candidate substitutions for a causal historical role , we reasoned that functionally important residues are most likely to be conserved among descendant sequences and to be located in the regions of the protein structure that differ between the mollusk ERs and other SRs . Of the 79 substitutions , 44 are conserved in all or all but one mollusk ER sequences ( Fig . S6 ) . We plotted these historical replacements on the crystal structures of CgER and human ERα and on a homology model of AncLophoSR . Four historical sequence changes emerged as top causal candidates , because they contributed to filling the ligand pocket or improved packing among helices near the pocket and/or the coactivator interface . These four sites form a ring around the lower portion of the ligand cavity in the elements of the structure that are stabilized by interactions with the ligand in vertebrate SRs . Specifically , a415W ( on the loop between helices H6 and H7 ) , h524F and l525F ( both on H10 ) , and l536F ( on the loop between H10 and AF-H ) place large hydrophobic side chains into open spaces within the ligand cavity or in smaller spaces between key helices ( Fig . 3B , using lower and upper case to denote the states in AncLophoSR and CgER , respectively ) . To test the hypothesis that these four sequences changes conferred on the evolving mollusk ER its constitutive activity , we introduced the CgER states into the AncLophoSR background and determined their impacts on activation and allosteric regulation by ligand . When introduced singly , none of the four was sufficient to fully recapitulate the derived phenotype . l536F , however , caused a very large increase in constitutive activity ( and on maximal ligand-dependent activity , as well ) . Neither a415W and nor h524F had an observable effect on its own . l525F abolished all activity , indicating that it is incompatible with the ancestral background ( Fig . 4 ) . We next assayed all possible two-fold combinations of the four candidates . We found that two pairs – h524F/l536F and a415W/l536F – each completely recapitulated evolution of the CgER-like phenotype , with very high constitutive activity and no additional activity induced by ligand . Despite the strong contributions of h524F and a415W when combined with l536F , the pair a415W/h524F had little effect on either constitutive or ligand-induced activity . All combinations containing l525F abolished both types of activity . These data indicate that large-effect mutations played key roles in the evolution of constitutive activity , with just two substitutions required to recapitulate the entire shift in function in the likely ancestral background . Further , significant epistasis is present , because the effects of h524F or a415W differ radically depending on whether site 536 has the ancestral leu ( in which case these substitutions have no apparent effect ) or the derived Phe ( in which case they abolish ligand-dependent activity and yield a solely constitutive activator ) ( Fig . 4 ) . We next prepared all three-fold combinations and found that combining the three substitutions that in pairs contribute to the derived phenotype – h524F , a415W and l536F – causes no functional difference compared to h524F/l536F or a415W/l536F ( Fig . 4 ) . Epistasis is again apparent in the redundant effects of these mutations on function: adding either h524F or a415W to l536F abolishes ligand-regulation and further enhances constitutive activity , but adding both causes the same effect as adding either one . We therefore conclude that two historical amino acid replacements from the mollusk lineage are sufficient to recapitulate the evolution of constitutive activity in the ancestral background: l536F makes a major , independent contribution , and adding either h524F or a415W is sufficient to explain the evolution of total ligand-independent activity . The fourth candidate substitution – l525F – abolishes all transcriptional activity , both constitutive and ligand-dependent , when introduced in isolation or in any combination with the other substitutions ( Fig . 4 ) . This result is surprising , because the derived state F525 is conserved among all known mollusk ERs and does not render them nonfunctional . Other historical mutations that occurred in the stem mollusk lineage must interact epistatically with l525F , exerting a permissive effect that allows mollusk ERs to tolerate l525F without losing function . The residues found in CgER at three of these four sites – W415 , F525 , and F536 – are conserved among all known mollusk ERs . The fourth , F524 in the CgER , is a tyrosine in the ancestral mollusk ER and most extant mollusks , suggesting that mutation Y524F occurred later in the lineage leading to Crassostrea . To determine the effect of having a tyrosine at this position , we repeated all experiments using genotypes containing tyrosine instead of phenylalanine at site 524 . In every background , the tyrosine yielded nearly identical functional behavior as the phenylalanine ( Fig . S7 ) , indicating that either of the two bulky aromatic states found in mollusk ERs can make a similar contribution to the evolution of constitutive activity . Taken together , these data indicate that two historical substitutions – l536F and either a415W or h524Y – were sufficient to cause the evolution of constitutive activity , and the subsequent substitution Y524F in the lineage leading to CgER did not affect this activity . Given the fact that two pairs of three historical substitutions are sufficient to recapitulate the evolution of constitutive activity in AncLophoSR , we asked whether reversing them to their ancestral states in CgER could restore the ancestral ligand-regulated function . The answer is no . When either pair is reversed , full constitutive activity remains present and no ligand regulation is apparent . This result indicates that additional “restrictive” mutations occurred , which made reversal of the mutations that were once sufficient to cause the evolution of the new function no longer sufficient to restore it [16] . The three-fold revertant also remained fully constitutive and ligand-independent . ( Fig . 5 ) When l525F , the fourth candidate substitution near the ligand pocket , was also reversed , however , the four-fold revertant became ligand-regulated , manifesting a >3-fold , dose-responsive increase in activation upon administration of estradiol ( Figs . 5 , S4B ) . This change is accomplished by knocking down constitutive activity substantially while maintaining high levels of activation only when hormone is added . Although more than 90 other replacements occurred between AncLophoSR and CgER , reversing just four of them is therefore sufficient to restore an allosteric response to hormone . Additional derived states in CgER must have had a further restrictive effect , because reversing these four substitutions does not fully abolish constitutive activity . We next determined whether it is necessary to reverse all four states to restore ligand dependence to CgER by assaying all possible combinations of the four ancestral and derived residues in the CgER protein . The answer is yes: only when all four states are reversed is full ligand-dependent activity restored , and just one of the triple revertants displayed even partial ligand-dependent activation ( Fig . 5 ) . Thus , although only two of the four historical substitutions are necessary to trigger the evolution of ligand-independent activity , two others must also be reversed for the ancestral function to be reacquired . Specifically , the historical pairs l536F/a415W or l536F/h524F are sufficient to yield constitutive activity; whichever pair came first , adding the third redundant substitution has no further effect on function but prevents reversing the other two from restoring the ancestral function . The fourth substitution , l525F , is also restrictive , but instead of having no effect on function , it requires permissive mutations to be tolerated . These mutations' effects are unlike those of previously observed restrictive mutations that impede irreversibility [16]: in those cases , the restrictive substitutions created a genetic background in which reverting the function-switching mutations to the ancestral state renders the protein nonfunctional . In the mollusk ER , the restrictive substitutions cause reversal of the key substitutions to be functionally inconsequential , not deleterious . These inferences about reversibility are robust to uncertainty about the ancestral states . At three of the four key sites , the ancestral state in AncLophoSR are reconstructed with little or no ambiguity . At the other – ala415 – serine is a possible alternate state , with PP = 0 . 20 . Reversing Trp415 in CgER to ser instead of ala along with the three other ancestral states also restores ligand-sensitive activity to CgER . Moreover , switching ala415 to ser in AncLophoSR does not change the ancestral receptor's estrogen-sensitive activity ( Fig . S5 ) . These experiments confirm the importance of the key substitutions l536F , a415W , and h524F in the evolution of constitutive activity . In the genetic background of CgER , the derived states at these sites confer strongly increased constitutive activity and a loss of ligand regulation compared to having the ancestral states , just as they do in AncLophoSR , although there are some subtle differences in their interactions in the two different contexts . Thus , the capacity of these historical substitutions to recapitulate the evolution of constitutive activity is robust to whether or not the genetic background contains the many other sequence changes that occurred along the lineage leading to extant mollusks — an interval of more than 500 million years and over 100 substitutions . Finally , we sought to understand the structural mechanisms that mediated the functional effects of the four key historical mutations during evolution and their epistatic interactions with each other . Ligand-activated receptors serve as allosterically controlled transcriptional regulators: they exist in a thermodynamic equilibrium , in which the inactive conformation is favored in the absence of ligand and the active conformation is favored in its presence . Examination of the CgER structure suggests that the historical substitutions confer constitutive activity not by filling the ligand cavity per se but by stabilizing the active conformation in the absence of ligand enough to remove allosteric control . The additional restrictive mutations prevented reacquisition of ligand-sensitivity by occluding binding of the ligand and conferring excess stability to the active conformation . Specifically , the large-effect substitution l536F – which potentiates transcriptional activity in the presence or absence of ligand and is required for the evolution of constitutive activity – stabilizes the interaction of the activation function helix ( AF-H ) with the rest of the protein by improving packing interactions . The small side chain of the ancestral leu 536 leaves a small secondary cavity open within the protein interior – one spatially distinct from the ligand pocket – at the crucial point where AF-H , H3 , and H10 meet to stabilize the tertiary structure of the active conformation ( Fig . 6A ) . Replacing leu with the much bulkier Phe fills this cavity and acts as a sort of linchpin that improves packing of these helices against each other , presumably stabilizing the active conformation and providing a structural explanation for this mutation's potentiating effect on activation by the receptor . Substitution a415W – which has little effect on its own but enhances constitutive activity and abolishes the response to estrogen once F536 is present – stabilizes the active conformation by improving packing interactions and contributes indirectly to occlusion of the ligand cavity . This residue lies outside of the ligand cavity on helix H6 , but it interacts with F425 , a conserved residue on H7 , the side chain of which lines the cavity . Replacing the small side chain of the ancestral a415 with the much bulkier derived Trp causes a clash with F425 , which in turn moves directly into the ligand cavity and packs against F404 , another conserved residue on the beta-turn . The result is to both occlude the ligand pocket and to result in stronger packing interactions between H6 and H7 and between H7 and the beta-turn , increasing stability along that face of the protein ( Fig . 6B ) . These structural effects are consistent with a415W's effects of preventing activation by hormone and increasing constitutive activity in the absence of ligand . Substitution h524F – which also does not strongly affect function in isolation but increases constitutive activity when combined with F536 – contributes to the loss of ligand activation and the evolution of constitutive activity more directly . In the estrogen- activated receptors , the ancestral histidine side chain accepts a hydrogen bond from estradiol's 17β-hydroxyl; replacing this residue with the larger , nonpolar Phe eliminates this interaction and causes a clash with the ligand ( Fig . 6C ) , explaining this substitution's negative effect on ligand activation . In addition , Phe's aromatic ring packs strongly against hydrophobic side chains on helices 3 and 6/7 , forming a “bridge” across the bottom of the pocket from helix 10 that stabilizes the receptor in the absence of ligand , explaining its enhancement of constitutive activity ( Fig . 6C ) . An aromatic Tyr residue , as found in other mollusk ERs , is anticipated to behave similarly . The redundancy of h524F and a415W presumably occurs because either derived residue is sufficient to clash with ligand , abolishing activation by hormone , and – if F536 is present – to achieve maximal activation in the absence of ligand . Finally , substitution l525F , when introduced into AncLophoSR , destroys receptor activation whether ligand is present or not . The bulky sidechain of the derived Phe points directly into the ligand cavity , clashing with estrogen and packing against residues on the AF-H loop and helix H3 ( Fig . 6D ) , explaining its deleterious effect on estrogen activation . In CgER , Phe 525 also contributes to constitutive activity , because it serves as a structural hub that makes van der Waals contacts to numerous residues around it , connecting H10 to residues on H3 , H7 , and AF-H . Why the Phe replacement also eliminates constitutive activity – and does so in all combinations of ancestral and derived states when introduced into AncLophoSR ( Fig . 4 ) – is unclear . The opposite effect of this substitution in AncLophoSR implies that the effects of this residue , which packs against so many structural elements important for activation , depends on the specific position and angle of the backbone and the rotamers at nearby sites . The evolution of the mollusk ER can be understood as a vestigialization of the protein's allosteric regulatory mechanism . This process is analogous to the gradual and neutral evolutionary degradation of unused morphological characters – such as the hindlimbs of whales or the eyes of cavefish [44] , [45] – which then leave underlying structural traces of their past existence . In this case , the LBD became “stuck” in the active conformation whether or not ligand was bound , due to two large-effect mutations that stabilized the active conformation , shifting the equilibrium towards that conformation even in the absence of ligand . Once allosteric regulation was lost , the architecture that had been required for ligand-dependence – such as the large internal cavity for binding ligand and the dependence of AF-H's position on ligand – degraded further , without apparent consequence for the receptor's transcriptional function or allosteric regulation . Vestiges of this architecture , however , have persisted in the extant mollusk ER in a nonfunctional state since the molluscan ancestor , >500 million years ago [46] , as demonstrated by the fact that CgER can regain ligand-dependence by reversing a small number of historical mutations in and around the ligand cavity and AF-H . One consequence of vestigialization is that regaining the feature becomes more genetically complex than merely reversing its initial loss , because of additional decay in the underlying architecture . In the case of the mollusk ER's allosteric mechanism , two ancient mutations conferred full constitutive activity , shifting the equilibrium towards the active conformation even in the absence of ligand . Once these two mutations were in place , additional mutations further filled the ligand cavity and further stabilized the activation conformation . These mutations caused no apparent functional effect on the receptor's functional output , because the allosteric mechanism was disabled anyway , but they further degraded the underlying architecture of allostery . As a result , restoring the ancestral function into CgER now requires at least two additional “de-vestigializiating” mutations to remove bulky residues from the pocket and partially shift the equilibrium in the absence of ligand back towards the inactive conformation . A similar “ratchet-like” mechanism has been observed during the evolution of other proteins , with substitutions occurring after a functional shift that make reverse evolution to the ancestral structure and function more genetically complex and evolutionarily unlikely than before [16] . If only two mutations were required to trigger the evolution of constitutive activity in the ancestral background , why would the additional redundant/restrictive substitution ( s ) have evolved and then been conserved ? One possibility is that they evolved neutrally , a result consistent with the finding that they have no apparent effect on receptor function . Neither W415 or F525 , the potentially redundant mutations , is conserved in all mollusk ERs , a result consistent with at least partial neutrality . In many mollusks , however , these states have persisted over a long time , an observation that could be explained by the fact that reversion to the ancestral states cannot be accomplished with a single nonsynonymous mutation , and the intermediate amino acids may be deleterious . Alternatively , although the derived states have no discernible effects on allosteric regulation or transcriptional activity , they could contribute to other properties , such as folding stability , that might affect function or fitness in certain lineages or environments . Our findings show how the physical architecture of the steroid receptor LBD shaped its evolutionary potential . Allosteric regulation of SRs involves a shift in the thermodynamics of receptor activation upon ligand binding . In the absence of ligand , the inactive conformation is more stable than the transcriptionally active conformation , so the majority of receptor molecules are in the inactive conformation . When the ligand is bound , however , the active conformation is more stable than the ligand-independent conformation , so the presence of ligand drives most receptor molecules into the active form . Structurally , the difference between the two conformations is relatively subtle , primarily involving the ordering and packing of one helix against the protein's body . Because of the delicate energetic balance among these functionally distinct but structurally similar conformations , relatively small perturbations in the stability of one conformation vis-à-vis the other have the potential to make a receptor active even in the absence of ligand , or inactive even in its presence . Stabilizing the active conformation in the absence of ligand can make it more stable than the inactive conformation and cause the allosteric switch to become stuck in the “on” conformation , resulting in a constantly active transcription factor . Conversely , destabilizing the active conformation relative can make a receptor unable to activate transcription whether or not ligand is present . Because of the simple physical basis to evolve constitutive activity , the minimal genetic architecture required to trigger such a shift is simple . Just one or a few mutations can shift the thermodynamic equilibrium among states and radically change the protein's capacity to be regulated allosterically . In contemporary SRs , for example , clinical and laboratory single point mutations are known that make nuclear receptors constitutively active by stabilizing the active conformation in the absence of ligand [47]–[52] . Similarly , our findings indicate that during the historical evolution of the mollusk ER , acquiring only two substitutions was sufficient to confer constitutive activation . The biophysical architecture of allostery therefore influenced the process of genetic evolution in the mollusk ER . Because of the relatively delicate energetics and subtle structural basis of allosteric regulation , a very small number of mutations triggered a large shift in its functional behavior . In this way , the evolving mollusk ER's structural properties influenced the processes of genetic evolution and made the evolution of a radical shift in function relatively simple in genetic terms . Subsequently , the architecture of the receptor's new constitutive activity allowed additional substitutions to make evolutionary reversal to the ancestral function increasingly complex . Taken together , these findings illustrate how the structural basis of protein function shapes genetic evolution , not only by imposing constraints but also by facilitating the emergence of certain radical changes in function .
Alignments of 135 nuclear receptors , identified using BLAST and downloaded from GenBank and the JGI genome browser ( Table S2 ) , were made using MUSCLE [53] , [54] , followed by manual editing . ProtTest [55] was used to determine the best-fit model of evolution ( the JTT substitution matrix [56] with gamma-distributed rate variation , a proportion of invariant sites , and observed amino acid frequencies ) . The phylogeny was inferred using PHYML [57] , and statistical support for each node was evaluated by obtaining the approximate likelihood ratio ( the likelihood of the best tree with the node divided by the likelihood of the best tree without the node ) . The ancestral reconstruction was performed using PAML and Lazarus software [58] , [59] using the ML tree edited to place the Strongylocentrotus purpuratus ERR in the expected position . This tree editing had very little impact on the ancestral state reconstruction . When we compared the ancestors generated from the edited tree to the ancestors generated from the unedited tree , the ancestral lophotrochozoan ER sequence differed at only two residues , neither of which had high levels of support on either tree ( in both cases the ML state had PP<0 . 22 , and the second most likely state was identical to the ML state on the other tree ) ( Table S3 ) . We used the sequence obtained on the edited tree to resurrect the ligand-binding domain ( LBD ) of the ancestral lophotrochozoan SR ( Genbank ID KC261633 ) . The CgER LBD ( a gift of T . Matusumoto , National Research Institute of Aquaculture , Japan ) was subcloned into pMCSG9 , which includes a His-tag , MBP , and TEV cleavage site . The CgER/MBP/His protein was expressed in BL21DE3 pLysS cells , and induced with 200 mM IPTG . The protein was purified using a nickel affinity column . The MBP/His fusion was cleaved from the protein using TEV protease , and MBP/His fusion tag was purified from the CgER with a second run on the nickel column . Fractions containing CgER LBD were dialyzed into 150 mM NaCl , 20 mM TrisHCl ( pH 7 . 4 ) and 2 mM CHAPS , and concentrated to 3 mg/mL . Chemical purity was assessed by SDS-PAGE . Crystals were grown by hanging drop vapor diffusion from solutions containing 1 µl of protein and 1 µl of 25% Peg 4000 , 10% glycerol , 01 . M TrisHCl ( pH 4 . 8 ) . Crystals were cryoprotected in 25% Peg 4000 , 20% glycerol , 01 . M TrisHCl ( pH 4 . 8 ) and were flash-cooled in liquid nitrogen . Data were collected at 100 K and wavelength of 1 Å the South East Regional Collaborative Access Team at the Advanced Photon Source ( Argonne , Illinois , USA ) , and were processed and scaled with HKL2000 [60] . Molecular replacement [26] using human ERα ( 1ERE ) was used to determine initial phases for CgER . Structures were refined using COOT [61] and Refmac [62] . All residues were either Ramachandran-favored ( 98 . 77% ) or allowed ( 1 . 23% ) . The structure has been deposited with the Protein Data Bank ( PDB 4N1Y ) . VOIDOO was used to calculate the probe-occupied volume of the ligand-binding pocket , using a probe radius of 1 . 4 angstroms [63] . Cavities were calculated 10 times , with molecules rotated to different orientations prior to the VOIDOO cavity calculation . Values are shown as mean ± standard deviation . Pymol ( Schrödinger , LLC ) was used to construct all structure figures , and LOVOalign was used to calculate the RMSD between human ERα ( 1GWR ) and the CgER LBD ( 4N1Y ) [64] . MODELLER [65] , [66] was used to make a homology model of the mutated CgER LBD . The mutated CgER LBD was modelled onto the CgER LBD structure . Ten homology models were created , and the model with the lowest DOPE score was used for structural comparisons . A homology model for the AncLophoSR:estradiol complex was generated using the functionally similar human ERα:estradiol complex ( 1ERE ) as a guide . Residue replacements were performed using the program COOT [67] and rotamers that approximated the side chain positions in either human ERα ( 1ERE ) . Four regions of the AncLophoSR model were built using CgER ( 4N1Y ) as a guide due to either gaps or insertions in the sequence alignment ( Ala19-Thr28 , Asp-87-Lys91 , Cys105-Met111 ) or in one case due to a lack of ordered residues in human ERα ( Ala146-Asp166 ) . The AncLophoSR aligns without gaps or insertions with CgER , suggesting that these structural differences have no functional impact with respect to estrogen activation . Rotamers were again corrected by hand using COOT . The model was then subject to 500 rounds of energy minimization in the program Phenix software [68] to correct geometry . Structures were rendered for display using Pymol ( Schrödinger , LLC ) . The hinge and LBD of the Crassostreas gigas ER were cloned into the pSG5-Gal4DBD vector ( a gift of D . Furlow ) . The resurrected ancestral lophotrochozoan SR LBD sequence was synthesized as a fusion construct containing the hinge domain and C-terminus of the CgER ( Genscript , Piscataway , NJ ) , and cloned into the pSG5-Gal4DBD vector ( Fig . S2B ) . CHO-K1 cells were transfected using Lipofectamine and Plus with 1 ng LBD plasmids , 100 ng of luciferase reporter plasmid ( pFRluc ) and 0 . 1 ng of a normalization plasmid ( phRLtk ) . After 4 hours , the transfection mixture was replaced with medium supplemented with stripped serum , and allowed to recover . The cells were treated with 17β-estradiol ( Steraloids , Newport , RI ) diluted in medium/serum and incubated for 24 hours . Luciferase assays were performed using DualGlo luciferase ( Promega , Madison , WI ) . Mutations were made using QuikChange Lightning Site-directed mutagenesis ( Agilent , Englewood , CO ) , and were verified by sequencing . | An important goal in evolutionary genetics is to understand how genetic mutations cause the evolution of new protein functions and how a protein's structure shapes its evolution . Here we address these questions by studying a dramatic lineage-specific shift in function in steroid hormone receptors ( SRs ) , a physiologically important family of transcription factors . In vertebrates , SRs bind hormones and then undergo a structural change that allows them to activate gene expression . In mollusks , SRs do not bind hormone and are always active . We identified the genetic and structural mechanisms for the evolution of constitutive activity in the mollusk SRs by using X-ray crystallography , ancestral sequence reconstruction , and experimental studies of the effects of ancient mutations on protein structure and function . We found that constitutive activity evolved due to just two historical substitutions that subtly stabilized elements of the active conformation , and subsequent mutations filled the hormone-binding cavity . The structural characteristics required for a hormone-sensitive activator were thus vestigialized , much the same way that a whale's hindlimbs became vestiges of their ancestral form after they became dispensable . Our findings show how the architecture of a protein can shape its evolution , allowing radically different functions to evolve by a few large-effect mutations . | [
"Abstract",
"Introduction",
"Results",
"and",
"Discussion",
"Methods"
] | [
"biochemistry",
"biology",
"evolutionary",
"biology",
"biophysics",
"evolutionary",
"genetics"
] | 2014 | Vestigialization of an Allosteric Switch: Genetic and Structural Mechanisms for the Evolution of Constitutive Activity in a Steroid Hormone Receptor |
Brucellosis and leptospirosis are among neglected tropical zoonotic diseases particularly in the resource limited countries . Despite being endemic in these countries , there is paucity of information on its magnitude . This study investigated seropositivity of Brucella spp . and Leptospira spp . , and associated factors among abattoir workers and meat vendors in the city of Mwanza , Tanzania . A community based cross-sectional study was conducted in Mwanza city from May to July 2017 . Socio-demographic and other relevant information were collected . Detection of Brucella spp . and Leptospira spp . antibodies were done using slide agglutination test and microscopic agglutination test , respectively . Data were analyzed using STATA version 13 Software . A total of 250 participants ( 146 abattoir workers and 104 meat vendors ) were enrolled with median age of 31 ( IQR: 25–38 ) years . The overall , seropositivity of Brucella spp . antibodies was 48 . 4% ( 95% Cl: 42–54 ) . Seropositivity of B . abortus was significantly higher than that of B . melitensis ( 46 . 0% , 95%Cl: 39–52 vs . 23 . 6% , 95% Cl: 18–28 , P<0 . 001 ) while seropositivity of both species was 21 . 2% ( 95%Cl: 16–26 ) . The seropositivity of Leptospira spp . was 10 . 0% ( 95% CI: 6–13 ) with predominance of Leptospira kirschneri serovar Sokoine which was detected in 7 . 2% of the participants . Being abattoir worker ( OR: 2 . 19 , 95% CI 1 . 06–4 . 54 , p = 0 . 035 ) and long work duration ( OR: 1 . 06 , 95%CI: 1 . 01–1 . 11 , p = 0 . 014 ) predicted presence of both B . abortus and B . melitensis antibodies . Only being married ( p = 0 . 041 ) was significantly associated with seropositivity of Leptospira spp . Primary education was the only factor independently predicted presence of Brucella spp . antibodies among abattoir workers on sub-analysis of occupational exposure . None of factors were found to be associated with presence of Brucella spp . antibodies among meat vendors on sub-analysis . Seropositivity of B . abortus antibodies among abattoir workers and meat vendors is high and seem to be a function of being abattoir worker , having worked for long duration in the abattoir and having primary education . In addition , a significant proportion of abattoir workers and meat vendors in the city was seropositive for Leptospira kirschneri serovar Sokoine . There is a need to consider ‘one health approach’ in devising appropriate strategies to control these diseases in the developing countries .
Brucellosis and Leptospirosis are among neglected tropical diseases which are endemic in resource limited countries including those in the sub-Saharan African region [1–3] . They are major public health concern due to their epidemiological patterns which involves animal-human interfaces resulting into economic losses and sub-clinical infections among human population . In human , these infections present with nonspecific symptoms , as a result they are misdiagnosed with other febrile illnesses such malaria , typhoid fever and rheumatic fever [4] . Leptospirosis is worldwide distributed particularly in tropical and some temperate regions . It is an occupational disease affecting individuals working close with animals . Leptospirosis outbreaks often occur after floods whereby the infected urine from animals such as rodents , dogs and cattle easily contaminate the water and environment hence spread the infection to humans [5 , 6] . The annual incidence of human Leptospirosis is estimated to be 1 . 03 million cases worldwide with 58 , 000 deaths being attributed to the disease [7] . In East African region the annual incidence is estimated to be 25 . 6 cases per 100 , 000 population [7] . Brucellosis is a contagious bacterial zoonotic disease of public health importance . Abattoir workers and others that work closely with animals or animal products have a high risk of contracting the disease[8 , 9] . The disease is endemic in the south and the Central America , Mediterranean , Africa , Indian subcontinent , Asia , Arab peninsula and Middle East . The annual incidence is estimated to range from 214 . 4 to 1603 . 4 cases per 100 , 000 population [10–12] . In Tanzania , the prevalence has been reported to range from 0 . 7 to 23 . 9% among the high risk groups [13–18] . Livestock brucellosis and leptospirosis [3 , 19–21] are endemic in the lake zone that supply animals destined for slaughter in Mwanza city . Abattoir workers and meat sellers may be at high risk if biosafety measures are not in place . Despite livestock brucellosis and leptospirosis being common in Tanzania especially in the Lake Victoria zone , there is paucity of data regarding the seropositivity of these pathogens among human population in the city of Mwanza , Tanzania . This study was designed to provide baseline information regarding the seropositivity of these pathogens , the information that may be useful for designing control interventions and provide insights for future research in this area .
This was a community based cross-sectional study ( S1 File ) that was conducted between May and June 2017 in the city of Mwanza , Tanzania . The study was conducted among Igoma abattoir workers and meat vendors in the city . The abattoir has 250 workers , and about 200 cows and 50 goats are slaughtered per day . The city of Mwanza is the second largest in Tanzania with a total population of 706 , 453 according to 2012 National census [22] . The city possesses two districts namely; Ilemela and Nyamagana with a total population of 343 , 001 and 363 , 452 , respectively . The city depends on Igoma abattoir to supply meat to more than 90 meat selling shops in the city . The sample size was estimated by using Kish Leslie formula ( 1965 ) , using the Brucella spp . seropositivity of 14 . 1% from a previous study by Mngumi et al . , [17] . The Minimum sample size obtained was 186 , however a total of 250 participants were enrolled . After obtaining a written informed consent , participants working in abattoir and meat retail shops were serially enrolled . Socio-demographic and other relevant information were collected using interview standard questions ( S2 File ) . Data collected included: age , sex , occupation ( abattoir worker , retail meat seller ) , residence , education and work duration in years . About 4 to 5 mls of venous blood was collected in a plain vacutainer tubes ( Becton , Dickinson and Company , USA ) and transported to the Catholic University of Health and Allied sciences ( CUHAS ) multipurpose laboratory within 4 hours of collection . The sera were extracted by centrifugation at 2500 rpm for 10 minutes , decanted into cryovials in duplicates and stored at -40°C until processing . One set of the samples was transported to the Pest management centre at the Sokoine University of Agriculture ( SUA ) whereby the detection of Leptospira spp . antibodies was made . The detection of specific Brucella spp . antibodies was done at Catholic University of Health and Allied Sciences ( CUHAS ) . Detection of specific Brucella antibodies for B . abortus and B . melitensis was done using commercial rapid agglutination test according to the manufacturer’s instructions . The Eurocell A was used for B . abortus and Eurocell M for B . melitensis ( Euromedi equip LTD . UK ) . In each run the positive and negative control were used . Results were interpreted as positive if the agglutination reaction was similar to that of positive control . The test has been found to have sensitivity and specificity of 95% and 100% , respectively [23] . Regarding the detection of Leptospira spp . antibodies , local Leptospira serovars previously isolated from animals ( domestic animals and rodents ) in Tanzania namely: Leptospira kirschneri serovar Sokoine , L . interrogans serovar Lora , L . kirschneri serovar Grippotyphosa , L . borgpetersenii serovar Kenya and L . interrogans serovar Hebdomadis were selected and used in microscopic agglutination test ( MAT ) [24 , 25] . The selected serovars were cultured into fresh Leptospira Ellinghausen and McCullough , modified by Johnson and Harris ( EMJH ) culture medium incubated at 30°C for 4 to 7 days before using as live antigen in ( MAT ) . Culture purity and density was checked using dark-field microscope whereby a density of 300×108 Leptospires/ml was considered adequate for MAT . Serum samples were serially diluted with phosphate buffered saline ( pH 7 ) in a ratio of 1:10–1:80 in U–bottomed microtiter plate and 50 μl was used in MAT . Prepared live Leptospires antigen ( 50 μl ) was added to the diluted serum to give final dilutions of 1:20–1:160 ( 100 μl total volume ) of serum-antigen mixture in each microtiter well . The first row was used for negative control while positive control was added in the same row as the sample . The plates with serum–antigen mixture were incubated at 30°C for 2–4 hours before being examined for agglutination under dark field microscope . A sample was considered positive for a specific serovar if more than 50% of the microorganisms in the microtiter well were agglutinated at the titer of ≥ 1: 80 . Data collected was entered into a Microsoft excel sheet then analyzed using STATA version 13 software . The categorical variables were presented as proportions while continuous variables ( age and work duration ) were summarized as median with interquartile ranges . Cross-tabulation was done to determine factors with collinearity using Pearson Chi squared test . The median age and median work duration of Brucella spp . seropositive and seronegative participants were compared by Wilcoxon Mann-Whitney / ranksum tests . Logistic regression model was used to determine factors associated with the presence of specific Brucella spp . and Leptospira spp . antibodies . Factors with p value of less than 0 . 2 on univariate analysis were subjected to multivariable regression analysis . Odds ratio ( OR ) and 95% Confidence interval ( Cl ) were noted . A P value of < 0 . 05 was considered statistically significant . The protocol for conducting this study was approved by the joint CUHAS/BMC research ethics and review committee with ethical clearance number CREC/336/2017 . The permission was further granted by the city council director , village leaders and abattoir manager . Written informed consent was obtained from each participant prior recruitment to the study . All participants in the current study aged 18 years and above
All 250 participants were available for analysis . The median age of the study participants was 31 ( inter quartile range ( IQR ) : 25–38 ) years . One participant was female ( 0 . 4% ) and the majority 212 ( 84 . 8% ) of the participants were from urban areas . The majority of the participants 192/250 ( 76 . 80% ) were married . Out of 250 participants , only 51 ( 20 . 40% ) attended secondary education as shown in Table 1 . The median work duration ( years ) of those with low education was 7 ( IQR ) 3–12 and that of those with high education was 3 ( IQR ) 2–5 , p<0 . 001 . Overall seropositivity of Brucella spp . antibodies was found to be 48 . 4% ( 121/250 , 95% Cl: 42–54 ) . Seropositivity of B . abortus and B . melitensis was found to be 46 . 0% ( 115/250 , Cl: 39–52 ) and 23 . 6% ( 59/250 , 95% Cl: 18–28 ) , respectively while seropositivity of co-infection of B . abortus and B . melitensis was 21 . 2% ( 53/250 , 95% Cl: 16–26 ) On univariate analysis , age ( p = 0 . 029 ) , residing in rural areas ( p = 0 . 021 ) , having primary education ( p = 0 . 001 ) , being abattoir worker ( p<0 . 001 ) and work duration ( p = 0 . 002 ) were significantly associated with the presence of specific B . abortus antibodies . The median work duration of B . abortus seropositive participants was 7 ( IQR4-15 ) years compared to 5 ( IQR 3–10 ) of those who were seronegative ( p = 0 . 0016 ) , Fig 1 . By multivariable logistic regression analysis , having primary education ( OR:2 . 64 , 95% CI:1 . 25–5 . 55 , p = 0 . 011 ) , being abattoir worker ( OR:2 . 66 , 95% CI:1 . 49–4 . 77 , p = 0 . 001 ) and having long work duration ( OR:1 . 05 , 95% CI:1 . 015–1 . 09 , p = 0 . 041 ) were found to predict B . abortus seropositivity ( Table 2 ) . By univariate analysis , residing in rural areas ( p = 0 . 037 ) , being abattoir worker ( p = 0 . 023 ) and long work duration ( p = 0 . 014 ) were significantly associated with B . melitensis seropositivity . However , only long work duration ( OR: 1 . 05 , 95% CI: 1 . 00–1 . 10 , p = 0 . 024 ) was found to predict B . melitensis seropositivity on multivariable logistic regression analysis ( Table 3 ) . On univariate analysis residing in rural areas ( p = 0 . 030 ) , being abattoir worker ( p = 0 . 012 ) and long work duration ( p = 0 . 004 ) were significantly associated presence antibodies for both species . By multivariable logistic regression analysis; being abattoir worker ( OR: 2 . 19 , 95% CI: 1 . 06–4 . 54 , p = 0 . 035 ) and having long work duration ( OR: 1 . 06 , 95% CI: 1 . 01–1 . 11 , p = 0 . 014 ) were found to predict presence of antibodies for both species ( B . abortus and B . melitensis ) ( Table 4 ) . When sub analysis was done among the two groups ( abattoir workers and meat vendors ) the following was observed . Among the variables tested in the abattoir workers group for B . abortus; increased in age ( p = 0 . 001 ) , long work duration ( p = 0 . 004 ) , being married ( p = 0 . 007 ) and having primary education ( p<0 . 001 ) were significantly associated with B . abortus seropositivity . By multivariable logistic regression analysis only having a primary education ( OR: 3 . 77 , 95% CI: 1 . 45–9 . 76 , p = 0 . 006 ) was found to predict B . abortus seropositivity . Regarding B . melitensis . only long work duration ( p = 0 . 013 ) was significantly associated with B . melitensis seropositivity . By multivariable logistic regression analysis none of the factors was found to predict B . melitensis seropositivity . Primary education ( OR: 2 . 93 , 95% CI: 1 . 19–7 . 23 , p = 0 . 019 ) independently predicted Brucella spp . seropositivity . None of the factors tested was found to be associated with Brucella spp . seropositivity among meat vendors . ( S3 File ) Overall seropositivity of Leptospira spp . antibodies was found to be 26/250 ( 10 . 0% , 95% CI: 6–13 ) . When categorized by occupation , there was no significant difference on seropositivity among abattoir workers and meat vendors ( 11/46 ( 7 . 7% ) vs . 14/104 ( 13 . 5% ) , p = 0 . 124 ) . Among the five Leptospira serovars tested , the most prevalent was Leptospira kirschneri serovar Sokoine ( 7 . 2% ) , followed by L . interrogans serovar Lora ( 2 . 0% ) and L . kirschneri serovar Grippotyphosa ( 1 . 2% ) ( Fig 2 ) . Other serovars tested were L . borgpetersenii serovar Kenya and L . interrogans serovar Hebdomadis in which none of the samples were found to be seropositive . Among the factors tested , only being married was significantly associated with seropositivity of Leptospira spp . ( p = 0 . 041 ) .
Despite high specificity of serological test used in the current study , B . abortus and B . melitensis antigens are not specific when it comes to antigen-antibody assays because these two spp are more than 95% in structural homology . It should be noted that this study was done in the specific groups therefore results cannot be generalized to the general population of Mwanza city and being a cross-sectional study the trend of the outcome by time could not be established . In addition , the high seropositivity of Brucella spp . might be due to cross-reactivity of Brucella antigens with varieties Enterobactericeae antibodies . Brucella species identification through serology is markedly affected by cross reactions[41] and this is associated with high false positive . B . abortus seropositivity among abattoir workers in Mwanza city is alarmingly high and is predicted having long work duration and having primary level of education . In addition , a significant proportion of this population is seropositive to Leptospira kirschneri serovar Sokoine . Being important zoonoses and neglected tropical diseases , there is a need to emphasize on biosafety measures during slaughtering , surveillance strategies , and treatment across the country particularly in high risk groups . Moreover , this calls for the need to adopt “One Health Approach” in controlling these diseases across the country . Further studies focusing on molecular detection of the pathogens to provide opportunities for understanding the infection patterns and the epidemiological implications of the pathogens to the high risk communities are highly warranted . | Brucellosis and leptospirosis are among neglected diseases in many low-income countries affecting both animals and human populations . Despite being common , the information on their distribution are scarce . In a view of that , this study investigated the proportion of participants with positive antibody test specifically for the two diseases among slaughter house workers and meat sellers in Mwanza city . The study involved 250 participants aged between 25 and 38 years . Overall , 48 . 4% of participants were Brucella spp . seropositive . Proportion of participants who were positive for B . abortus specific antibodies was higher than that of B . melitensis while 21 . 2% of them found to have antibodies for both species studied . Furthermore , about 10% of the participants had Leptospira antibodies . Being abattoir worker , long work duration and having primary education were associated with the presence of Brucella antibodies while only being married was associated with the presence of Leptospira antibodies . The findings from this study emphasize the need for multisectoral approach in devising control strategies for these pathogens . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"medicine",
"and",
"health",
"sciences",
"leptospira",
"pathology",
"and",
"laboratory",
"medicine",
"pathogens",
"sociology",
"tropical",
"diseases",
"microbiology",
"diet",
"social",
"sciences",
"animal",
"products",
"social",
"systems",
"regression",
"analysis",
"bac... | 2018 | Seropositivity of Brucella spp. and Leptospira spp. antibodies among abattoir workers and meat vendors in the city of Mwanza, Tanzania: A call for one health approach control strategies |
The biotrophic basidiomycete fungus Ustilago maydis causes smut disease in maize . Hallmarks of the disease are large tumors that develop on all aerial parts of the host in which dark pigmented teliospores are formed . We have identified a member of the WOPR family of transcription factors , Ros1 , as major regulator of spore formation in U . maydis . ros1 expression is induced only late during infection and hence Ros1 is neither involved in plant colonization of dikaryotic fungal hyphae nor in plant tumor formation . However , during late stages of infection Ros1 is essential for fungal karyogamy , massive proliferation of diploid fungal cells and spore formation . Premature expression of ros1 revealed that Ros1 counteracts the b-dependent filamentation program and induces morphological alterations resembling the early steps of sporogenesis . Transcriptional profiling and ChIP-seq analyses uncovered that Ros1 remodels expression of about 30% of all U . maydis genes with 40% of these being direct targets . In total the expression of 80 transcription factor genes is controlled by Ros1 . Four of the upregulated transcription factor genes were deleted and two of the mutants were affected in spore development . A large number of b-dependent genes were differentially regulated by Ros1 , suggesting substantial changes in this regulatory cascade that controls filamentation and pathogenic development . Interestingly , 128 genes encoding secreted effectors involved in the establishment of biotrophic development were downregulated by Ros1 while a set of 70 “late effectors” was upregulated . These results indicate that Ros1 is a master regulator of late development in U . maydis and show that the biotrophic interaction during sporogenesis involves a drastic shift in expression of the fungal effectome including the downregulation of effectors that are essential during early stages of infection .
The basidiomycete Ustilago maydis is a biotrophic pathogen colonizing maize . The resulting disease , the so-called smut disease , is characterized by the formation of large tumors on all aerial parts of the plant . In these tumors fungal hyphae proliferate profusely and eventually produce massive amounts of dark pigmented , diploid teliospores . U . maydis is a dimorphic fungus which can grow by yeast-like budding in the absence of a host . On the leaf surface compatible haploid yeast-like cells mate and generate a dikaryon . The dikaryon switches to a filamentous form which is cell cycle arrested [1–3] . Upon perception of surface cues [4] the dikaryon differentiates infection structures and penetrates the maize epidermis . Following penetration , the cell cycle arrest is released [2] , the dikaryon invades the plant tissue , proliferates with the help of clamp formation and triggers the development of large tumors [2 , 5] . In the late stages of infection , after tumors are formed , the sporogenesis program is initiated . Although the chronology of events leading to teliospore formation is not yet fully understood , the first step is likely to be the fusion of the two haploid nuclei , followed by extensive mitotic divisions of the diploid hyphal cells leading to the formation of large hyphal aggregates [5 , 6] . Concomitantly , a mucilaginous matrix of undefined composition and origin is formed embedding the fungal cells during the subsequent hyphal fragmentation and maturation stages . Eventually , the tumors rupture and release the diploid spores in the environment . The cycle is completed when the teliospores germinate and give rise to haploid progeny after meiosis [6] . To overcome PAMP-triggered plant defense responses , and to establish a biotrophic interaction , U . maydis secretes a large panel of effector proteins which may function in the apoplast ( e . g . Pep1 ) or be translocated to the host cells ( e . g . Cmu1 , Tin2 ) [7–9] . Expression of the vast majority of the about 300 effectors lacking known protein domains is tied to the biotrophic stage [10] . About 25% of all effectors are arranged in gene clusters and many of these affect virulence either generally or in an organ-specific manner [10–13] . So far , the molecular basis for the virulence function of only a few U . maydis effectors ( Pep1 , Pit2 , Cmu1 , Tin2 , See1 ) has been elucidated [8 , 9 , 14–16] . U . maydis pathogenic development requires fusion of haploid cells and is initiated by the a and b mating type genes . Their expression is induced by the pheromone response factor Prf1 in response to pheromone and host derived signals transmitted via a cAMP-dependent and a MAPK pathway . The a locus encodes a pheromone/receptor system mediating cell-cell recognition and fusion . The b locus encodes two homeodomain proteins which , when derived from different alleles , form the bE/bW heterodimer which acts as master regulator for the switch to filamentous growth , host tissue colonization and tumor induction [10 , 17] . bE/bW induces a regulatory cascade which influences the expression of over three hundreds genes [3 , 17] . The majority of these genes are regulated via the zinc-finger protein Rbf1 , a direct target of bE/bW which acts as a central node in the b-regulatory cascade [3] . Because effector gene expression coincides with pathogenic development , the induction of most effector genes was initially considered to depend on the b cascade [10] . However only few effector genes are subject to direct regulation via components of this cascade . Rbf1 induction in axenic culture leads to activation of only a small subset of effector genes suggesting that plant signals and additional regulators are required for their expression [3 , 17] . The membrane proteins Sho1 and Msb2 induce effector gene expression in response to surface cues prior to penetration . They act via the transcription factors Biz1 and Hdp2 , two Rbf1 targets , with a specific function in appressorium development [1 , 18] . Biz1 might also regulate effector gene expression in conjunction with Mzr1 in the later stages of colonization [19] . Proteins of the WOPR family constitute a novel class of fungal-specific transcriptional regulators that bind DNA via their N-terminal WOPR box . The WOPR box consists of two highly conserved domains , WOPRa and WOPRb , predicted to adopt a globular structure and which are separated by a linker region of variable length and sequence . Both WOPR domains are required for DNA binding activity [20] . Most fungal genomes contain two paralogous WOPR genes that phylogenetically fall into two distinct clades [21 , 22] . To date , WOPR proteins have been studied exclusively in ascomycetes where they fulfill a conserved function in the control of developmental processes . Mit1p in Saccharomyces cerevisiae is a core regulator of invasive growth in haploid cells as well as pseudohyphal growth in diploid cells [23 , 24] . Wor1 , the best characterized member of the WOPR family , is the master regulator of the white-opaque phenotypic switching allowing Candida albicans to adapt to niches in the human host [20] . Similarly , the Ryp1 protein in Histoplasma capsulatum is a key regulator of the temperature-dependent mycelia-to-yeast transition critical for virulence [25] . Sge1 was the first WOPR protein studied in a plant pathogenic fungus . Sge1 supports parasitic growth of the tomato wilt pathogen Fusarium oxysporum f . sp . lycopersici by inducing the expression of at least four of the SIX effector genes [26] . WOPR proteins have later been linked to virulence in other ascomycete plant pathogens [27–31] . Several members of the WOPR family in plant pathogens also positively regulate the production of secondary metabolites potentially involved in pathogenicity [27–29] . In addition to their role in plant colonization , most WOPR proteins regulate sexual/asexual reproduction in phytopathogenic fungi [26–28 , 30–32] . Here we investigate the function of a WOPR regulator in the pathogenic development of the basidiomycete pathogen U . maydis . We show that Ros1 ( Regulator of sporogenesis 1 ) is not required for plant colonization but is essential for teliospore production occurring late during the biotrophic life cycle . ros1 deletion strains are locked in the dikaryotic , filamentous stage of infection . They fail to undergo karyogamy , subsequent mitotic cell divisions and are unable to form the mucilaginous matrix in which teliospores are embedded . We show that Ros1 affects the expression of many genes via direct interaction with their promoter regions . Remarkably , Ros1 triggers a dramatic switch in gene expression of the vast majority of effector genes .
The genome of U . maydis is predicted to encode two members of the WOPR family , pac2 ( UMAG_15096 ) and ros1 ( UMAG_05853 ) . As the deletion of pac2 ( UMAG_15096 ) was previously shown to not have any effect on U . maydis virulence or reproduction [33] , only ros1 was investigated here . A BLAST search revealed that Ros1 is conserved in other smut species belonging to the four genera of the class Ustilaginomycetes ( Ustilago , Sporisorium , Pseudozyma , and Melanopsichium ) . Outside of the Ustilaginomycetes , conservation is restricted to the N-terminal part comprising the WOPR box ( S1 Fig ) . In Ros1 the N-terminal WOPR box ( amino acids 8 to 305 ) contains both WOPRa ( amino acids 8 to 90 ) and WOPRb ( amino acids 240 to 305 ) domains separated by a rather long linker region of 156 amino acids ( S1A Fig ) . With the exception of Thr210 , all residues critical for DNA binding in Wor1 of C . albicans are conserved in Ros1 ( S1B Fig ) [34] . With a length of 879 amino acids Ros1 is the longest WOPR protein described so far ( S1A Fig ) . Several members of the WOPR family share a conserved nuclear localization signal ( NLS ) ( PGEKKRA ) ( S1 Fig ) . This motif is absent in Ros1 and we could not identify any other canonical NLS . However , bioinformatic tools predict Ros1 to localize in the nucleus . In addition Ros1 displays an unusually long polyglutamine stretch ( 46 residues , amino acids 707–752 ) in its C-terminal domain . Several other members of the WOPR family possess one or more glutamine-rich regions ( S1A Fig ) which might mediate the interaction with components of the transcriptional machinery [35] . To study the localization of Ros1 , a haploid FB1 strain was generated which constitutively expresses a C-terminal mCherry fusion of Ros1 together with the nuclear envelope marker Nup107eGFP . The red signal from Ros1mCherry was detected in the nucleus , surrounded by the green fluorescence of the nuclear envelope , demonstrating that Ros1 is targeted to the nucleus ( Fig 1A ) . The nuclear localization and the presence of the long polyQ stretch are consistent with a potential function of Ros1 as transcriptional regulator . To study its function we deleted ros1 in the two compatible haploid strains FB1 and FB2 . FB1Δros1 and FB2Δros1 could successfully mate and produce dikaryotic filaments on charcoal containing medium ( S2 Fig ) . When injected into maize seedlings , nearly all plants infected with the ros1 deletion strains developed tumors . However , compared to the FB1 x FB2 mixture , ros1 mutants exhibited reduced virulence ( Fig 1B ) . In particular , only 4 . 8% of the plants infected with the ros1 mutant mixture were dead after 12 days compared to 53 . 7% for the FB1 x FB2 infection . Remarkably , dark pigmented teliospores were absent in the tumors induced by the ros1 deletion strains ( Fig 1C ) and softening of the tumor tissue which usually becomes evident when teliospores accumulate in wild type infections did not occur . Spore formation could be restored by reintroducing ros1 in single copy in the ip locus of FB1Δros1 and FB2Δros1 strains ( Fig 1B ) . Because constructs containing a promoter region of 1 kb did not complement , the entire 7 . 6 kb region separating ros1 from the upstream gene UMAG_05850 was included in the complementation construct . While the complementation strains had regained the ability to produce spores ( Fig 1C ) , virulence was only partially complemented ( Fig 1B ) . We speculate that this reflects a position effect resulting from integrating the construct into the ip locus . Alternatively , the expression of ros1 might partially depend on distal regulatory elements that are missing in the complementation construct . To determine at which stage of development ros1 deletion strains are affected , we stained fungal hyphae with wheat germ agglutinin-Alexa Fluor 488 and followed the sequence of events leading to the formation of mature teliospores in wild-type infections by confocal microscopy ( Fig 2 , left panel ) . Until 4 days after infection , growth of the ros1 deletion strains was indistinguishable from the growth of wild type strains . In both cases uniform spreading of fungal hyphae within the leaf was observed ( Fig 2 ) . After 4 days wild type hyphae started to form aggregates which became visible at 6 dpi , indicating that the mucilaginous matrix in which the cells are embedded during spore formation [37] was produced ( Fig 2 ) . Between 6 and 8 dpi the hyphal aggregates continued to expand reaching diameters of up to 250 μm at 8 dpi . Around 10 dpi , hyphae in these aggregates underwent fragmentation and individual cells entered the spore maturation process . Finally , at 12 dpi groups of mature teliospores with their characteristic ornamentation became clearly visible . This sequence of events parallels what has been described [6] . By contrast , in plants infected by ros1 deletion strains neither hyphal aggregates nor fragmented hyphae could ever be observed ( Fig 2 , right panel ) . Thus , in the absence of ros1 , U . maydis development was locked in the filamentous stage . To see how this failure to aggregate was linked to expression of the ros1 gene , the expression pattern of ros1 in haploid strains in axenic culture and of FB1 x FB2 mixtures during plant infection was analyzed using quantitative RT-PCR ( Fig 3 ) . ros1 was expressed at a very low basal level in axenic culture and during the early steps of maize infection . Expression was then upregulated 35-fold at 6 dpi when sporogenesis was initiated and reached a maximum 70-fold induction at 8 dpi as hyphal aggregates expanded . Between 8 and 12 dpi , ros1 transcript abundance slowly decreased , concomitantly with the accumulation of mature teliospores ( Fig 3 ) . To evaluate in more detail the contribution of Ros1 to the regulation of spore development , we analyzed karyogamy and matrix formation in the ros1 mutant . To be able to visualize nuclei , the nuclear envelope marker Nup107eGFP as well as the plasma membrane marker Sso1mCherry were introduced into FB1 , FB2 and the corresponding ros1 deletion strains . Infected plant tissues were then analyzed by confocal microscopy between 6 and 8 dpi . At this stage wild type filaments contained only one nucleus while pairs of nuclei were visible in filaments of the ros1 deletion strains ( Fig 4A ) , indicating that karyogamy had not occurred ( even at later time points ) . In addition , the mucilaginous matrix observed in tumors induced by wild type strains was absent in tumors induced by the ros1 deletion strains ( Fig 4B ) . Teliospore differentiation takes place in large hyphal aggregates [5 , 38] . As such aggregates were absent in tissue infected by ros1 mutant strains , we also studied the accumulation of fungal biomass in plants infected with wild type and ros1 deletion strains by quantitative RT-PCR ( Fig 4C ) . In line with the microscopic data ( Fig 2 ) , until 4 dpi wild type strains as well as ros1 mutants showed a comparable small increase in fungal biomass ( Fig 4C ) , likely reflecting coordinated mitotic divisions of the dikaryon with the help of clamp connections [2 , 3] . However , while in wild type infected tissue the ratio U . maydis/plant biomass increased dramatically ( 7 . 5 to 42 . 3 ) between 6 dpi and 12 dpi , the ratio remained at the 4 dpi level in plants infected with the ros1 deletion strains ( Fig 4C ) . These data show that Ros1 affects karyogamy as well as matrix formation and is needed for massive late proliferation in the infected tissue . To further characterize the function of Ros1 , we studied the effect of expressing ros1 prematurely . To mimic early pathogenic development of U . maydis , we used strain AB33 in which filamentous growth and cell cycle arrest can be triggered in liquid culture via nitrate-inducible expression of bE1 and bW2 homeodomain genes [39 , 40] . Ros1 was placed under the control of the arabinose-inducible crg1 promoter [41] . In nitrate minimal medium supplemented with glucose ( NM + Glucose ) , the b genes were expressed and cells switched from budding to filamentous growth ( Fig 5A ) . However , when ros1 was induced simultaneously ( NM + arabinose ) , cells failed to filament , increased their diameter and formed septa ( Fig 5A ) . DAPI staining revealed that each section contained one nucleus , indicating that mitotic divisions had resumed . When filamentation was induced prior to ros1 expression , the resulting filaments stopped elongating and became septated with one nucleus per segment ( Fig 5B ) . This shows that Ros1 counteracts the activity of the bE/bW heterodimer , inhibits filamentation and triggers mitotic divisions . To study the effect of premature ros1 expression during plant colonization , we generated compatible ros1 deletion strains expressing ros1 under the control of the mig2-6 promoter . mig2-6 is an effector gene whose expression is strongly upregulated shortly after penetration [42] . Compared to wild type infections and infections with ros1 deletion strains ( Fig 6A ) the strains expressing ros1 prematurely caused severely attenuated disease symptoms ranging from chlorosis to very small tumors ( Fig 6C ) . These strains penetrated the plant surface but invasion of the plant tissue rapidly stopped and hyphae with an abnormally high number of septa developed ( compare Fig 6B and 6D ) . These results show that the timing of ros1 expression is critical for biotrophic development and tumor induction of U . maydis . To identify genes regulated by Ros1 , we combined two approaches: RNA sequencing to evaluate the global effect of Ros1 on gene expression and ChIP sequencing to identify which of the differentially regulated genes are direct targets . Both experiments were carried out on samples collected at 8 dpi when the ros1 expression level is maximal . For the transcriptomic analysis , we compared maize tumor tissue infected by strains FB1 x FB2 and the corresponding ros1 deletion strains . RNA-seq data showed that 2005 genes were differentially regulated ( fold change ( FC ) ≥ 1 , 5; p-value < 0 . 01 ) . Of these genes , 1091 were expressed at lower levels in the wild type strains compared to ros1 mutants and 914 were higher expressed in wild type strains compared to ros1 mutant strains ( S1 Table ) . This shows that about 30% of the 6766 protein-encoding U . maydis genes are differentially regulated by Ros1 . RNAseq results were confirmed by qRT-PCR for 15 genes encoding two glycoside hydrolases ( UMAG_05550 , UMAG_04503 ) , a trehalase ( UMAG_02212 ) , a cyclopropane fatty acid synthase ( UMAG_01070 ) , a polyketide synthase ( pks1 [43] ) , transcription factors ( UMAG_04101 , biz1 [1] , rbf1 [3] , fox1 [44] , UMAG_02775 ) and secreted effectors ( mig2-3 [45] , UMAG_04096 , dik1 [46] , UMAG_02473 [10] , UMAG_03046 ) ( S3 Fig ) . The deletion of ros1 mostly affects metabolic processes and cellular transport ( Fig 7 ) . Functional categories “C compound and carbohydrate metabolism” , “lipid fatty acid and isoprenoid metabolism” as well as “secondary metabolism” were predominantly enriched in both upregulated and downregulated gene sets ( Fig 7 ) . In contrast , categories related to mitochondrial function ( respiration , electron transport , mitochondrion biogenesis and mitochondrial inner membrane ) as well as protein synthesis ( translation , ribosome biogenesis ) were enriched only among Ros1-upregulated genes . About 170 genes belonging to “cell cycle and DNA processing”and 55 genes belonging to “cell growth and morphogenesis” were differentially regulated . Although they did not show significant enrichment , both categories were highlyrepresented in the upregulated gene set . Unexpectedly , Ros1 was also shown to cause a massive shift in secreted effector gene expression . 128 effectors genes were downregulated including 126 effector genes without functional domains , cmu1 [8] and UMAG_01130 [13] , two effectors containing a functional domain . In addition , 70 effector genes were upregulated by Ros1: 68 without functional domains and two with functional domains , UMAG_03615 [10] and UMAG_11763 [13] ) ( Fig 8 , S1 Table ) . To identify the genes directly targeted by Ros1 , we carried out a ChIP-seq analysis . Maize seedlings were infected with a compatible pair of complemented ros1 deletion strains expressing an HA-tagged version of Ros1 . A pair of compatible strains complemented with the native version of Ros1 without an HA tag was used as negative control . After sequencing the output DNA from three biological replicates , 1907 peaks showed high reproducibility , a significant peak shape score ( > 20 ) and a low p-value ( p < 0 . 01 ) ( S2 Table ) . 1441 distinct intergenic regions including 620 intergenic regions for divergently transcribed genes were found to be targeted by Ros1 . Only considering promoter regions , this brings the number of genes potentially targeted by Ros1 to at least 1913 ( S3 Table ) . Of the 2006 genes which were differentially expressed in the RNA-seq data , only 790 ( 40% ) displayed at least one ChIP peak in their upstream region ( S1 Table ) suggesting that a significant part of Ros1 regulation depends on intermediate regulators . In total 80 transcription factor-encoding genes were differentially expressed in the RNA-seq dataset and of these 42 were downregulated and 38 were upregulated by Ros1 . 25 upregulated transcription factor genes were predicted to be direct targets from the ChIP-seq analysis ( S1 Table ) and six of these including ros1 displayed multiple Ros1-ChIP peaks in their promoters ( S2 Table ) . For example , in the long intergenic region between ros1 and UMAG_05850 ( Fig 9A ) we detected six regions bound by Ros1 , suggesting a rather complex regulation including autoregulation . The ChIP analysis also revealed that Ros1 binds the promoters of genes previously identified as regulators of sporogenesis rum1 , hgl1 , tup1 and ust1 , suggesting that they could be direct targets . These genes did not show differential regulation by Ros1 in the RNA-seq dataset generated at 8 dpi . However , a time-resolved analysis of their expression pattern showed that rum1 , hgl1 and ust1 are slightly but significantly induced by Ros1 at 10 and 12 dpi ( S4 Fig ) . In general , Ros1 binding is detected more frequently upstream of upregulated genes than upstream of downregulated genes ( 50% of the upregulated genes against 30% of the downregulated genes ) ( S1 Table ) . This suggests that Ros1 acts primarily as a transcriptional activator but can also function as a repressor . A more detailed discussion of the RNA-seq and ChIP-seq analysis is found in the discussion to avoid redundancy . The WOPR regulators Wor1 , Ryp1 and Mit1 all recognize a similar DNA binding motif [20 , 24] . A search for the corresponding 14 bp consensus sequence in Ros1 ChIP-seq data ( FIMO online tool , http://www . meme-suite . org ) , identified 975 motifs ( p-value < 0 , 001 ) corresponding to 763 ChIP peaks ( S4 Table ) . To test whether one of these regions is directly bound by Ros1 , we carried out electrophoretic mobility shift assays ( EMSA ) using a recombinant His-tagged version of Ros1 containing the WOPR domain only ( Ros1WOPR-His ) . As target we used a 237 bp biotinylated probe corresponding to the ros1 promoter region between 2854 and 3090 bp upstream of the ros1 gene ( Fig 9A and 9B ) containing three putative binding motifs ( WT-probe ) . In presence of Ros1WOPR-His in a molar ratio of 2700:1 of protein to DNA the probe was completely shifted . This shift could be abolished by addition of a 500-fold excess of unlabeled WT-probe competitor , indicating that Ros1 interacts specifically with the WT-probe ( Fig 9C ) . In comparison , incubation of Ros1WORP-His with a fragment of the same length from the ros1 open reading frame ( ORF-probe ) did not lead to a mobility shift ( Fig 9C ) . To narrow down the region bound by Ros1 , mutations were introduced in the probe sequence . Two of the predicted Wor1-like binding motifs ( m1 and m2 ) located at the center of the peak ( Fig 9A ) , were mutated in probes mut-m1 and mut-m2 ( Fig 9B ) and tested for binding by Ros1WOPR-His . For both probes we observed discrete , significantly smaller shifts than for the WT-probe and the interactions could again be competed by an excess of unlabeled WT-probe ( Fig 9C ) . When a fragment containing both mut-m1 and mut-m2 ( mut-m1+2 ) mutations was used , an even less shifted complex was observed which could be competed ( Fig 9C ) . This strongly indicates that the predicted m1 and m2 sites are indeed bound by Ros1 and suggests furthermore , that the WT-probe fragment contains an additional binding site , most likely the m3 site ( Fig 9B ) . Using similar conditions , we also tested the binding of Ros1WOPR-His to other promoter regions identified by ChIP which contain at least one predicted binding site . Probes were designed for promoters of a gene encoding a transcription factor ( UMAG_02775 ) upregulated by Ros1 , four effector genes downregulated by Ros1 ( UMAG_02854 , UMAG_04040 , UMAG_02538 [10] , cmu1 [8] ) and three effector genes upregulated by Ros1 ( UMAG_03138 , UMAG_12258 , UMAG_03046 ) . All probes were specifically bound by Ros1WOPR-His ( S5 Fig ) , confirming that these genes are direct targets of Ros1 . RNA-seq and Chip-seq analysis had shown that 47 transcription factor genes may represent direct targets of Ros1 while 33 may represent indirect targets . The expression pattern of two of these , UMAG_02775 ( presumed to be directly regulated by Ros1 ) and UMAG_01390 ( presumed to be indirectly regulated by Ros1 ) was additionally determined in maize plants infected with FB1 x FB2 or the corresponding ros1 deletion strains in a time course experiment ( S6 Fig ) . Results show that Ros1 is responsible for the late upregulation ( between 8 dpi to 12 dpi ) of these two genes . To follow up on these two transcription factors , the corresponding genes were deleted in FB1 and FB2 and mutant strains were then tested for virulence in maize seedlings and for their ability to produce teliospores ( Figs 10 and S7 ) . Tumor induction was not affected by the deletion of UMAG_02775 and the mutant hyphae produced aggregates ( Fig 10 ) . However , in these aggregates only few spores developed and these were misshaped and were missing the ornamentation characteristic of mature spores ( Fig 10 ) . The deletion of UMAG_01390 attenuated virulence to a comparable extent to what had been observed in the ros1 mutant strains ( Fig 10 ) . Contrary to the ros1 mutant , UMAG_01390 deletion strains showed hyphal aggregation and reached the fragmentation stage of spore development ( Fig 10 ) . However , fragmented hyphal cells failed to enter the maturation process and did not give rise to ornamented teliospores ( Fig 10 ) . Taken together these results illustrate that UMAG_02775 and UMAG_01390 genes both affect discrete steps in spore development downstream of Ros1 .
Previous light microscopy studies had seen paired nuclei commonly in hyphae outside the aggregates but not in the aggregates of sporogenous hyphae [5] . This is consistent with our analysis using fluorescent nuclear markers which shows that hyphae in aggregates are monokaryotic . In contrast to wild type strains , ros1 deletion strains are unable to form such aggregates , fail to accumulate matrix material and remain dikaryotic . This could suggest that karyogamy precedes the formation of hyphal aggregates and may be required to initiate the synthesis of the mucilaginous matrix ( Fig 11 ) . In wild type strains hyphal aggregates develop within the plant intercellular space and considerably expand over time due to a dramatic increase of fungal biomass which is not observed in plants infected with ros1 deletion strains . When expressed ectopically in axenic culture , Ros1 triggered mitotic divisions suggesting that aggregate expansion during colonization is due to multiple rounds of mitotic divisions of the diploid cells as was earlier hypothesized [5] . Moreover , the phenotype of cells expressing ros1 prematurely in hyphae indicates that the resulting cell divisions do not involve clamps . We speculate that without clamps diploid cells proliferate faster than the dikaryon which could explain the rapid and massive increase of fungal biomass late in infection . Starting at six dpi U . maydis cells begin to aggregate and form large , ball-like structures . What glues cells together is presently unknown , but based on the finding that the matrix can be stained with basic fuchsin [49] , it is likely that the matrix contains polysaccharides . In addition , it was reported that hyphae containing diploid nuclei are partially refractory to chemical fixation and this was attributed to lysis of the cell wall and its conversion to a gelatinous material [5] . Incidentally we noticed that during sporogenesis staining of the cell wall , but not of the septa , with wheat germ agglutinin-Alexa Fluor 488 becomes fainter with time ( Fig 2 , 8 dpi time point ) which could reflect chitin degradation or modification of the cell wall . Cell wall loosening might facilitate the changes in cell morphology which are associated with hyphal fragmentation and spore maturation . The matrix could connect the cells and shield them against biotic and abiotic stresses . Among the 255 Ros1-regulated genes belonging to the functional category “C compound and carbohydrate metabolism” , 55 are predicted to be involved in polysaccharide metabolism . Most of them encode glycoside hydrolases . They are enriched in both up and downregulated gene sets . The two most upregulated genes encode enzymes targeting the fungal cell wall , an EXG1 beta-glucanase ( FC = 831 ) and a chitinase A ( FC = 441 ) ( S1 Table ) , and both could conceivably be involved in the gelatinization process . The strong Ros1-dependent upregulation of an UDP glucose dehydrogenase ( UMAG_00118 ) suggests an increased glucuronic acid production at the onset of sporogenesis . Bacterial extracellular matrices as well as the capsule of Cryptococcus neoformans contain highly polar glycosaminoglycans which are rich in glucuronic acid [50] . Among the genes upregulated by Ros1 , we found a gene related to C . neoformans CAP59 . This C . neoformans gene is involved in capsule synthesis by supporting polysaccharide export [51] . It is conceivable that the related gene in U . maydis ( UMAG_11017 ) could fulfill a similar function in the export of matrix material . We also observed that the repellent gene rep1 , which is already upregulated in hyphae [52] is further induced by Ros1 during sporogenesis . Repellents are structural proteins forming amyloid fibrils at the cell surface which mediate hyphal adhesion to hydrophobic surfaces [53] . These amyloid fibrils could be of importance during sporogenesis , could aid in connecting hyphae with each other and assume a structural function in the formation of the matrix . Consistent with this hypothesis hydrophobins which are functionally similar to repellents [54] play a structural role in the formation of fruiting bodies [53] . Since rep1 deletion mutants still produce viable teliospores [52] Rep1 is unlikely to be essential for teliospore formation in U . maydis . However , it cannot be excluded that the efficiency of spore formation is affected in rep1 mutants . In addition to genes involved in cell wall modification we observed the Ros1-dependent upregulation of several genes involved in the synthesis / modification of membrane lipids ( Fig 10 ) . Among them were genes encoding ergosterol biosynthetic enzymes , sphingolipid biosynthetic enzymes and fatty acid synthesizing / modifying enzymes like cyclopropane fatty acid synthase ( S1 Table ) . The likely ensuing alterations of plasma membrane composition might reflect that the plasma membrane in spores has a different composition from that of vegetative cells . In Schizosaccharomyces pombe and S . cerevisiae , sporulation involves de novo synthesis of the forespore membrane within the cytoplasm of mother cells , which subsequently becomes the plasma membrane of the developing ascospores [55] . Cyclopropane fatty acid synthesis was also reported to be essential for fruiting body development in the basidiomycete Coprinus cinerea [56] . Sphingolipids have important roles in membrane and lipoprotein structure and in cell regulation as signaling molecules for growth and differentiation . They have been shown to be required for proper cell growth and morphology in U . maydis [57] and the upregulation of sphingolipid synthesis by Ros1 might be prerequisite for teliospore differentiation . Many genes involved in fatty acid beta oxidation were downregulated by Ros1 . This may reflect that the predominant spore storage fatty acids of U . maydis are linoleic and palmitic acid [58] . In line with this , a gene encoding a caleosin-like protein ( UMAG_02753 ) is strongly upregulated by Ros1 ( FC = 220 ) . Caleosins are involved in the structural maintenance and turnover of lipid storage organelles , so-called lipid droplets [59] . The strong upregulation of this gene at 8 dpi might thus indicate lipid storage in spores . Most WOPR regulators characterized so far in plant pathogenic ascomycetes regulate both plant invasion and sexual / asexual spore production [26–28 , 30–32] . Ros1 is the first WOPR protein characterized in a basidiomycete . In comparison to Wor1 from C . albicans and Ryp1 from Histoplasma capsulatum for which ChIP-chip identified only about 200 and 700 targets [60 , 61] , Ros1 might directly regulate a much larger set of genes ( 1900 identified by ChIPseq ) . Binding sites for all WOPR proteins characterized to date are conserved [20 , 24] , and we have shown here that Ros1 can also bind the 14 bp consensus sequence identified for Wor1 [20] . However , not all Ros1-bound regions identified by ChIP-seq ( 765 out of 1913 ) harbor this motif . This could indicate that Ros1 can recognize additional sequences more distantly related to the Wor1 binding site or that it can bind additional sequences via interaction with other transcription factors . Recent studies in C . albicans and H . capsulatum have provided evidence that WOPR proteins can bind promoters in complex with other core regulators and have suggested that the formation of these complexes might be mediated by glutamine-rich regions [61 , 62] which serve as protein interaction domains also in other proteins [63] . The presence of an exceptionally long poly-glutamine tract in the C-terminal domain of Ros1 might reflect such a role . Similarly to Wor1 [60] and based on finding that direct Ros1 targets include both up and downregulated genes Ros1 may function both as an activator and a repressor . In C . albicans , S . cerevisiae and H . capsulatum , Wor1 , Mit1 and Ryp1 are part of core regulatory networks in which each transcription factor regulates and is regulated by the others [24 , 61 , 62] . In these networks , WOPR regulators fulfill a critical function because they bind most of the target promoters [23 , 24 , 61 , 64] . Similar to the genes encoding core regulators in these species , ros1 exhibits an unusually long promoter and binds to its own promoter , most likely reflecting autoregulation . Moreover , Ros1 directly regulates many transcriptional regulators which remain to be investigated and which could potentially be core regulators . WOPR regulated genes have functionally diverged considerably during evolution and show very poor overlap even in closely related species [24] . However , several classes of genes / processes regulated by WOPR proteins in plant pathogenic fungi appear conserved: these include spore formation as well as secondary metabolism and effector gene expression ( all discussed below ) . Premature expression experiments in hyphae showed that Ros1 is antagonistic to bE/bW and inhibits b-dependent filamentation ( Fig 11 ) . Analysis of Ros1 dependent gene expression revealed that the negative effect of Ros1 on filamentous growth could originate from an alteration of the bE/bW regulatory cascade . About 50% of the 345 genes which are regulated by overexpressing bE1/bW2 in axenic culture [3] were found to be affected by Ros1 . 75 genes upregulated by bE/bW are repressed by Ros1 and this includes several regulators of the b cascade: Rbf1 , the central regulator of pathogenic development responsible for inducing the majority of the 345 b-regulated genes , and two of its downstream targets , Biz1 and Hdp1 , which modulate the cell cycle and regulate the growth of filaments [3 , 65] . Conversely , 40 genes downregulated by bE/bW are induced by Ros1 and 36 genes upregulated by bE/bW were also induced by Ros1 . Among the 170 b-dependent genes differentially expressed , Ros1 is likely to directly regulate 71 . Interestingly , bE , bW and prf1 , the main regulators of the b cascade , are only slightly repressed by Ros1 , suggesting that the Ros1 induced inhibition of filamentation does not re-establish the budding program This is also apparent when b-expressing cells are microscopically observed after premature expression of Ros1 . In these cases we observe the formation of septated , compartmentalized cells each containing a single nucleus . Such structures , where cell segments containing a single nucleus become deliminated by thick septa , are reminiscent to structures in sporogenous hyphae [5] . This suggests that Ros1 targets the b cascade at certain nodes without downregulating the entire cascade and this may then be prerequisite for entering this septation program . During infection , inhibition of parts of the b cascade would occur at a specific stage of biotrophic development , when the cell cycle has been released and the U . maydis dikaryon is proliferating in the infected tissue by clamp formation . This developmental stage is different from the filamentous stage achieved by overexpressing bE1/bW2 in axenic culture . Therefore , data sets generated by Heimel et al . ( 2010 ) [3] and the 8 dpi time point studied here after infection with the dikaryon are not fully comparable . Our studies have illustrated that ros1 mutants do not initiate nuclear fusion and fail to trigger massive proliferation . This shows for the first time that the strong increase in fungal biomass late in infection may require karyogamy to be completed . An inspection of the RNA-seq data revealed no differential regulation by Ros1 of the four U . maydis genes related to genes KAR7 , KAR2 , KAR3 and KAR4 implicated in karyogamy in S . cerevisiae . While 168 genes involved in DNA processing and cell cycle ( Fig 11 ) were differentially regulated by Ros1 , we did not observe any significant enrichment for these categories in the ros1 upregulated gene set . One likely explanation is that at the 8 dpi timepoint chosen for the RNA-seq analysis genes involved in karyogamy are already shut off . Among the 89 Ros1 upregulated genes in the category “DNA processing and cell cycle” we detect clear indicators for proliferation like DNA helicases , DNA primase , PCNA and several DNA mismatch repair proteins ( S1 Table ) . However , the ros1 deletion strain is also able to replicate its DNA in the dikaryotic phase , and this could explain why no significant enrichment is observed for the category “cell cycle and processing” ( Fig 7 ) . In line with this interpretation , protein synthesis and mitochondrial function ( respiratory chain ) were both overrepresented functional categories among the genes upregulated in sporogenous hyphae compared to the dikaryon of the ros1 mutant . Late proliferation of diploid hyphae might rely primarily on plant sugars as suggested by the strong upregulation of several glycoside hydrolases located at the surface of the cells , e . g . a secreted trehalase ( UMAG_02212 ) , a membrane located glucoamylase ( UMAG_04064 ) , the secreted invertase SUC2 and several other non characterized secreted glycoside hydrolases ( e . g . UMAG_00102 , UMAG_06434 ) . Infection experiments in maize seedlings showed that ros1 mutants cause significantly fewer dead plants compared to the wild type . We speculate that the massive late proliferation of the wild type in hyphal aggregates might negatively affect plant fitness and account for the high percentage of plants not surviving under our glasshouse conditions . We have shown that Ros1 controls the early events of spore development by inhibiting b-dependent filamentation and inducing karyogamy and hyphal aggregate formation as well as the initiation of spore formation . Among the upregulated genes in wild type infections are 38 putative transcription factors which are Ros1-regulated . Mutants in two of these transcription factors genes ( UMAG_02775 and UMAG_01390 ) revealed that both mutant strains were able to reach the gelatinization stage and to trigger hyphal fragmentation but failed at discrete subsequent steps of spore maturation . This suggests that these genes are regulators of the spore maturation process downstream of Ros1 . Other regulatory proteins previously shown to interfere with teliospore formation are Hgl1 , the histone deacetylase Hda1 , and the transcription factors Rum1 , Ust1 , and Tup1 [33 , 43 , 66–68] . Hgl1 , Rum1 and Tup1 are all required for proper spore development after the fragmentation stage [33] . Ust1 ( UMAG_15042 ) is a transcriptional repressor in haploid cells which represses filamentation as well as formation of spore-like structures . [68 , 43] . The ChIP analysis revealed that Ros1 binds the promoters of rum1 , hgl1 , tup1 and ust1 , suggesting that they could be direct targets . rum1 , hgl1 and ust1 were slightly but significantly induced by Ros1 at 10 and 12 dpi ( S4 Fig ) . The small effect of Ros1 could indicate that these genes are regulated by a combination with other TFs . For rum1 and hgl1 this is in line with their requirement for spore formation . To explain the upregulation of the negative regulator ust1 by Ros1 , we consider its role in controlling the budding program [68] may be required during sporogenesis when we observe its upregulation . This would imply that the observed formation of spore-like structures in haploid cells of the ust1 mutant [68] , could be a default pathway and not reflect what is happening during the sporulation program after infection . Consistently , we do not observe expression levels of ust1 during the U . maydis life cycle which are below the levels in axenic culture . tup1 expression levels were not influenced by Ros1 . As we did not find evidence that Ros1 differentially regulates the divergently transcribed UMAG_10827 gene ( S1 Table ) , we speculate that Ros1 binding to the promoter is not essential for the activity of this promoter under the tested conditions . Having shown that Ros1 participates in regulating hgl1 and rum1 expression and having shown that Ros1 is required at an earlier stage that precedes the formation of the sporogenous hyphal aggregates prior to karyogamy places Ros1 upstream of these genes and processes and highlights the importance of Ros1 for the regulation of sporogenesis . The functional analysis of the other Ros1-regulated transcription factors identified here is a promising avenue to elucidate the entire regulatory network controlled by Ros1 during sporogenesis . Secondary metabolism is commonly associated with sporulation processes in microorganisms , including fungi [69] . Among the genes upregulated by Ros1 , we found pks1 and laccase I responsible for the synthesis of the melanin pigment in teliospores [43] . In addition , the most upregulated gene involved in secondary metabolism is related to versicolorin B synthase , an enzyme involved in the synthesis of aflatoxine in Aspergillus sp . [70] . In U . maydis this gene belongs to a newly identified gene cluster ( E . Reyes-Fernàndez and M . Bölker , personal communication ) and we speculate that it is responsible for an intermediate step in the biosynthesis of an antimicrobial compound and / or a pigment . A gene cluster involved in the synthesis of itaconic acid [71] is also upregulated by Ros1 and in this case it appears to be the regulatory gene ria1 which is directly targeted while the promoters of the other genes in the cluster are not bound by Ros1 . Itaconic acid inhibits the bacterial glyoxylate shunt essential for many bacteria to survive during infection of mammalian hosts . The gene cluster for the production of the mannosylerythritol lipids which have biosurfactant and antimicrobial activity [72] is also upregulated by Ros1 . The antimicrobial properties of itaconate and mannosylerythrol lipids might serve to combat against competing microbes during late stages of U . maydis development . The modulation of effector gene expression seems to be a common trend shared by many WOPR proteins from various plant pathogenic fungi [26 , 29–32] . However , in C . fulvum and V . dahliae the deletion of the respective WOPR protein is associated already with growth defects in axenic culture [31 , 32] , and in Zymoseptoria tritici abnormally swollen cell structures are observed during axenic growth in the wor1 mutant [30] . This suggests that in these cases the respective WOPR protein might be more involved in developmental processes than in specific regulation of effector genes . Furthermore , in cases where effector gene expression has been analyzed , effectors are usually upregulated by the respective WOPR protein and the upregulation concerns a relatively small number of effectors ( six in F . oxysporum , 14 in F . verticilloides , six in V . dalhiae ) [26 , 29 , 32] . One of the main findings emerging from our transcriptional analysis is that Ros1 in U . maydis induces a massive switch in the effector repertoire affecting about 60% ( 194 genes ) of the predicted set of effector genes without functional domains [47 , 48] as well as 4 genes encoding effectors with functional domain . 82 of the 198 differentially regulated effector genes are likely to be directly targeted by Ros1 . These include cmu1 [8] , pit2 [15 , 73] , mig1 [74] , 23 genes residing in effector clusters ( 2A , 2B , 5A , 6A , 10A , 19A , 22A , ) [10] as well as 45 of the late-induced effectors . The differential regulation of the genes which are not direct targets likely results from the alteration of the bE/bW cascade by Ros1 . Contrary to Sge1 which induces effector genes required for virulence in F . oxysporum [26] , Ros1 mostly downregulates effector genes . 26 of the 128 downregulated effector genes ( Fig 7 ) reside in clusters where the cluster deletion causes a virulence defect ( 5A , 5B , 6A , 10A , 19A ) [10] and 10 are members of the eff1 effector family that also contributes to virulence [75] . Unexpectedly , among the effectors downregulated by Ros1 are also critical effectors involved in the inhibition of plant defense responses like Stp1 , Cmu1 and Pit2 [8 , 15 , 73 , 76] . In addition , of the 14 leaf-specific effector genes without functional domain [13] ( of which many contribute to virulence ) all are downregulated by Ros1 . While it is easy to conceive that see1 , an effector involved in cell expansion associated with tumor formation [16] is downregulated because tumor formation has happened already at the stage when ros1 is induced , the downregulation of effectors with critical functions in plant defense suppression is more difficult to explain . Either plant defenses at this late stage of U . maydis development could be distinct from the early stages of infection and require a different effector set for suppression . Or alternatively the massive production of matrix material which is associated with the formation of the fungal aggregates , could shield the aggregated hyphae from plant detection or from the action of antimicrobial compounds associated with plant defenses . Reduced effector expression could then be sufficient to maintain the inhibition of plant defense at the periphery of the aggregates . As a third possibility it could be the combined action of late induced effectors plus matrix that is effective . In total , there are 70 effector genes that are induced by Ros1 ( 68 without functional domains ) . Four of them ( UMAG_03138 , UMAG_05926 , UMAG_03046 , UMAG_12258 ) exhibited a 50-fold higher expression in infections with wild type compared to the ros1 mutant strains . 13 of these late-induced effector genes reside in effector clusters ( 19A , 10A , 2B , 5A , 6A , 9A ) including clusters involved in virulence ( 5A , 6A , 10A , 19A ) [10] . Since the cluster deletions were all generated in a solopathogenic haploid strain [10] in which spore formation does not follow the same sequence of events as in the dikaryon , we cannot presently assess whether these late effectors affect processes facilitating U . maydis sporulation , inhibit late plant defense responses and / or are participating to the formation of the matrix . Alternatively , these late effectors might , together with secondary metabolites induced at that stage , be used as a cocktail to defend the spores against other microbes which could colonize tumor tissue when it dries up and ruptures , releasing the spores . One could also consider that late effectors could fulfill a signaling function inside the aggregates to control the spore maturation process . In conclusion , Ros1 emerges as the central regulator of a major developmental reprogramming leading to teliospore production and completion of the life cycle in U . maydis ( Fig 10 ) . Of particular interest is how U . maydis can survive in the hostile plant environment with reduced expression of a large set of effector proteins of which many have a critical virulence function early during colonization . In addition , elucidating the role of the “late effectors” which are specifically induced by Ros1 promises to provide new insights into how this facultative biotrophic fungus has established itself in its natural environment . Moreover , we are confident that deciphering the structure and dynamics of the regulatory network in which Ros1 functions will provide an understanding how this master regulator achieves such a broad control over gene expression . Another yet unresolved task is the identification of the upstream signals triggering ros1 induction during biotrophic development and thereby inducing late development including spore formation .
The Escherichia coli strain Top10 ( Life technologies ) and BL21 ( DE3 ) pLysS ( Promega ) were used for cloning purposes and for expression of recombinant Ros1 protein respectively . U . maydis strains used in this study are listed in S5 Table , they are derivates of haploid strains FB1 and FB2 [77] or AB33 [39] . Cells were grown in liquid YEPSL ( 0 . 4% yeast extract , 0 . 4% peptone , 2% sucrose ) at 28°C on a rotary shaker at 220 rpm . For virulence assays , compatible haploid strains were grown separately in YEPSL to an OD600 of 1 . 0 , transferred to the same volume of sterile water and mixed in equal amounts prior to injection into maize seedlings . For premature ros1 expression studies , AB33 and strains derived from AB33 were grown in complete medium ( CM ) [78] supplemented with glucose ( 2% ) to an OD600 of 0 . 5 . Cells were collected by centrifugation , washed with H2O and resuspended in nitrate minimal medium ( NM ) [78] containing arabinose ( 2% ) as sole carbon source . Cells were subsequently grown for 12h for microscopic observation . To induce ros1 after the switch to filamentous growth , AB33 or AB33 derived strains were incubated for 6 h in NM + glucose and then shifted to NM + arabinose for 6 h . All chemicals used for media preparation were of analytical grade and were obtained from Sigma-Aldrich . PCR reactions were performed using the Phusion High-Fidelity DNA Polymerase ( New England Biolabs ) . Templates were either FB1 genomic DNA or indicated plasmid DNAs . Point mutations were generated using the Quick change lightning kit ( Agilent Technologies ) . Restriction enzymes were all supplied by New England Biolabs . U . maydis was transformed by protoplast-mediated transformation [79] Gene replacements and integrations into the ip locus [80] were verified by Southern blot analysis . All primer sequences used to generate plasmids are listed in S6 Table . To generate the Ros1mCherry fusion construct , plasmid p123 [81] conferring resistance to carboxin and allowing integration into the U . maydis ip locus was used . mCherry-HA was amplified from plasmid p1139 ( kindly provided by A . Djamei ) using primers mCherry_EcoF / mCherry_NotR and cloned in place of gfp between EcoRI and NotI sites of p123 to generate pPotef-mCherry-HA . The ros1 open reading frame was amplified with primers ros1_XmaF / ros1_EcoR and cloned between sites XmaI and EcoRI of pPotef-mCherry-HA to generate pPotef-ros1-mCherry-HA . pPotef-ros1-mCherry-HA was linearized by SspI prior to transformation of U . maydis . For the deletion of ros1 , a PCR-based strategy [82] and the SfiI insertion cassette system [79] were used . 1 kb long left border and right border fragments adjacent to ros1 were PCR-amplified using primer pairs Dros1LB_F/R and Dros1RB_F/R and FB1 genomic DNA as template . The resulting fragments were ligated to the hygromycin resistance cassette of pBS-hhn [82] via SfiI restriction sites and cloned into pCRII-TOPO ( Life technologies ) to generate pDros1 . The deletion construct was PCR amplified from plasmid pDros1 using primers Dros1_F/R and transformed into U . maydis strains FB1 and FB2 to generate FB1Δros1 and FB2Δros1 . The drag and drop cloning method [83] was used to generate plasmids pDUMAG_02775 and pDUMAG_01390 for the deletion of UMAG_02775 and UMAG_01390 , respectively . Primer pairs D02775LB_F/R and D02775RB_F/R and primer pairs D01390LB_F/R and D01390RB_F/R were used to amplify left and right border fragments from UMAG_02775 and UMAG_01390 , respectively . Left and right border fragments and the hygromycin resistance cassette were integrated in plasmid pSR426 ( kindly provided by S . Reissmann ) by homologous recombination in S . cerevisiae . The deletion constructs were excised from plasmids pDUMAG_02775 and pDUMAG_01390 after cleavage by Bsu36I and used for transformation of FB1 and FB2 to generate strains FB1ΔUMAG_02775 , FB2ΔUMAG_02775 , FB1ΔUMAG_01390 and FB2ΔUMAG_01390 . For complementation of ros1 mutant strains , plasmid p123-Bsu was generated from plasmid p123 [81] by introducing a silent point mutation in the cbx gene to create a Bsu36I site using primers p123Bsu_F/R . The 7 . 5 kb long genomic region separating ros1 from the upstream gene UMAG_05850 was used as promoter in complementation constructs . This region might contain an additional gene ( UMAG_05852 ) . To make sure that complementation constructs would not have two copies of this gene , the putative start codon of UMAG_05852 was deleted . To this end , the 6 kb intergenic region between UMAG_05850 and UMAG_05852 was amplified from genomic FB1 DNA with primers Cros1_KpnF / Cros1_SbfR and a second fragment containing UMAG_05852 and ros1 open reading frame was amplified using primers Cros1_SbfF / Cros1_NotR . Both fragments were ligated and cloned into pCRII-TOPO ( Life technologies ) . In the resulting plasmid pCRII-TOPOros1 the sequence GCTGACGCATG containing the start codon of UMAG_05852 is changed to TAGCATAG . Using primers TOPOros1_mF / TOPOros1_mR , a silent point mutation was introduced in pCRII-TOPOros1 to remove a KpnI site located in UMAG_05852 . The complementation construct was then excised from pCRII-TOPOros1 and cloned into p123-Bsu between KpnI and NotI sites . The resulting plasmid pCros1 was linearized with Bsu36I and transformed into ros1 deletion strains to generate strains FB1Δros1-ros1 and FB2Δros1-ros1 . For complementation of UMAG_02775 mutant strains , fragments corresponding to the UMAG_02775 promoter and open reading frame were PCR amplified from FB1 genomic DNA using primer pairs P02775_F/R and orf02775_F/R respectively and cloned in p123 between KpnI and NotI Sites . The resulting plasmid pCUMAG_02775 was linearized by SspI and transformed into FB1ΔUMAG_02775 and FB2ΔUMAG_02775 to generate strains FB1ΔUMAG_02775-UMAG_02775 and FB1ΔUMAG_02775-UMAG_02775 . For complementation of UMAG_01390 mutant strains , a fragment corresponding to UMAG_01390 promoter followed by the open reading frame was PCR amplified from FB1 genomic DNA using primer pairs C01390_F/R and cloned in p123 between KpnI and NotI sites . The resulting plasmid pCUMAG_01390 was linearized by SspI and transformed into FB1ΔUMAG_01390 and FB2ΔUMAG_01390 strains to generate strains FB1ΔUMAG_01390-UMAG_01390 and FB1ΔUMAG_01390-UMAG_01390 . For the generation of strains expressing a C-terminal Ros1HA fusion for the ChIP analysis , the ros1 open reading frame without the stop codon was amplified from FB1 genomic DNA using primers Ros1_XmaF / Ros1_NotR and subcloned in vector p1306 ( Kindly provided by A . Djamei ) between XmaI and NotI sites upstream of a triple HA sequence to generate pRos1-3HA . A fragment containing the last 1 kb of ros1 orf and the triple HA tag sequence was then amplified from pRos1-3HA using primers Endros1_NcoF / Endros1_PspR and cloned in pCros1 between NcoI and NotI sites . The resulting vector pCros1-3HA was linearized by Bsu36I and transformed in ros1 deletion strains to generate FB1Δros1-Ros1HA and FB2Δros1-Ros1HA . To allow expression of ros1 from the crg1 promoter , Pcrg1 was integrated upstream of ros1 in the native locus of AB33 using pRU11 [39] . Right border ( corresponding to the first 1000 bp of ros1 ) and left border fragments ( corresponding to the 1000 bp upstream of the ros1 start codon ) were amplified from FB1 genomic DNA using primer pairs crg-ros1LB_F/R and crg-ros1RB_F/R , respectively . The generated fragments were cloned between NdeI and EcoRI sites of plasmid pRU11 . The resulting plasmid pPcrg1-ros1 was linearized with NcoI and transformed into AB33 to generate strain AB33Pcrg1Ros1 . To express ros1 from the mig2-6 promoter , Pmig2-6 [42] was amplified from FB1 genomic DNA using primers Pmig2-6_NdeF / Pmig2-6_XmaR and cloned into NdeI and XmaI sites of p123 to generate pPmig2-6 . The ros1 gene was amplified with primers Ros1_XmaF / Ros1_NotR and cloned via XmaI and NotI sites downstream of Pmig2-6 in pPmig2-6 . The resulting plasmid pPmig2-6-ros1 was linearized by SspI and transformed into ros1 deletion strains to generate strains FB1Δros1Pmig2-6Ros1 and FB2Δros1Pmig2-6Ros1 . To allow expression of the Sso1 nuclear membrane marker fused to mCherry , Psso1 was amplified from FB1 DNA with primers Psso1_KpnF and Psso1_NcoR . Potef was replaced by Psso1 in p123 using KpnI and NcoI sites to generate pPsso1 . A fragment containing the mCherrySso1 fusion gene followed by the terminator of sso1 ( Tsso1 ) was amplified from plasmid pBS-otef-mCherry-sso1-hyg [84] using primers McSso1_NcoF and McSso1_HpaR . The fragment containing mCherrySso1-Tsso1 was then integrated between NcoI and HpaI sites in place of GFP-Tnos of pPsso1 to generate pPsso1-mcherry-sso1 . This vector was linearized by SspI and transformed into FB1 , FB2 , FB1Δros1 and FB2Δros1 . To introduce the nuclear marker nucleoporin Nup107 fused to GFP in the resulting strains , pNup107GFP-ble [85] was used . The insert from this plasmid encoding the Nup107GFP including the regulatory sequences was amplified with primers Pnup_F/R and integrated in the native nup107 locus to generate strains FB1Pnup107Nup107eGFP-Psso1Sso1mCherry , FB2Pnup107Nup107eGFP-Psso1Sso1mCherry and the corresponding ros1 deletion strains . The sequence of all PCR-amplified regions was verified . Haploid strains were grown in YEPSL medium to an OD600 of 1 . 0 , washed and resuspended in sterile H2O . Compatible strains were mixed in a 1:1 ratio and syringe-inoculated into seven-day-old maize seedlings of the variety Gaspé Flint ( originally provided by B . Burr , Brookhaven National Laboratories ) . Three independent infections were carried out for each strain and disease symptoms were evaluated after 12 days according to established disease rating criteria [10]http://journals . plos . org/plospathogens/article ? id=10 . 1371/journal . ppat . 1004272—ppat . 1004272-Kamper1 . Significant virulence differences between strains were assessed by one-way ANOVA applying the Tukey-Kramer test [36] . Total RNA was extracted from cells grown in axenic culture or from infected plant samples . Leaf material was ground in liquid nitrogen to a fine powder while cells in culture were pelleted by centrifugation and frozen in liquid nitrogen . Samples were resuspended in TRIzol reagent ( Life technologies ) and homogenized in a FastPrep-24 ( MP Biomedicals ) . Total RNA was isolated according to the manufacturer's protocol . Genomic DNA contaminants were eliminated using the Ambion Turbo DNA free Kit ( Life Technologies ) . For RNA sequencing RNA samples were further purified using the RNeasy Mini Kit ( Qiagen ) and the RNA quality was controlled using an Agilent 2100 Bioanalyzer . Total RNA samples from plants infected in three biological replicates with FB1 X FB2 strains or the corresponding ros1 deletion strains were used to prepare sequencing libraries with the Illumina TruSeq RNA sample preparation Kit . Library preparation started with 2 μg total RNA . After poly-A selection ( using poly-T oligo-attached magnetic beads ) , mRNA was purified and fragmented using divalent cations under elevated temperature . The RNA fragments were reverse transcribed using random primers . A second strand cDNA synthesis was carried out with DNA Polymerase I and RNase H . After end repair and A-tailing , indexing adapters were ligated to the cDNA . The products were then purified and amplified ( 14 PCR cycles ) to create the final cDNA libraries . After library validation and quantification ( Agilent 2100 Bioanalyzer ) , equimolar amounts of library were pooled . The pool was quantified using the Peqlab KAPA Library Quantification Kit and the Applied Biosystems 7900HT Sequence Detection System . The pool was sequenced using an Illumina TruSeq PE Cluster Kit v3 and an Illumina TruSeq SBS Kit v3-HS on an Illumina HiSeq 2000 sequencer with a paired-end ( 101 x 7 x 101 cycles ) protocol . Sequence reads were mapped to U . maydis protein encoding genes ( ftp://ftpmips . gsf . de/fungi/Ustilaginaceae/Ustilago_maydis_521/ ) using CLC Genomics Workbench 7 . 5 ( CLC bio ) . The unique gene reads for all of the 6970 annotated U . maydis genes from the 6 libraries were combined and analyzed in R using the Differentially Expressed Genes ( DEG ) algorithm edgeR [86] . Differentially expressed genes between FB1 x FB2 and FB1Δros1 x FB2Δros1 were selected on the basis of their fold change ( FC ≥ 1 . 5 ) and p-value ( < 0 . 01 ) . Expression data were submitted to GeneExpressionOmnibus ( http://www . ncbi . nlm . nih . gov/geo/ ) under the accession number GSE76231 . 7-day-old maize seedlings of the variety Gaspé Flint ( originally provided by B . Burr , Brookhaven National Laboratories ) were infected with mixtures of U . maydis strains FB1Δros1-Ros1HA x FB2Δros1-Ros1HA expressing an HA tagged Ros1 protein or FB1Δros1-Ros1 x FB2Δros1-Ros1 expressing a non-tagged Ros1 protein as negative control . Leaf samples ( from 5 different plants ) were collected at 8 dpi and incubated in fixation buffer ( 50 mM HEPES pH 7 . 5 , 1% formaldehyde ) for 10 min under vacuum . Excess formaldehyde was quenched by addition of 2 M glycine . Samples were then ground to a fine powder in liquid nitrogen . For chromatin preparation , powder was resuspended in lysis buffer ( 50 mM HEPES pH 7 . 5 , 150 mM NaCl , 1 mM EDTA , 1% Triton X-100 , 5 mM benzamidine , 2 mM PMSF , 1 X Roche complete EDTA free protease inhibitors ) and further treated by sonication to lyse the remaining cells using a microtip sonifier ( Branson ) . Chromatin was then sheared in a bioruptor sonication bath ( Diagenode ) for 10 cycles ( 30 s on / 30 s off ) at high power setting . An aliquot was saved to serve as input control and the rest of the chromatin solution was incubated with ChIP grade Protein A/G magnetic beads ( Life technologies ) coupled to a monoclonal anti-HA antibody ( Sigma ) for 10 h at 4°C . Beads were washed three times in lysis buffer , two times in high salt buffer ( 50 mM HEPES pH 7 . 5 , 300 mM NaCl , 1 mM EDTA , 1% Triton X-100 , 1 X Roche complete EDTA free protease inhibitors ) and once in TE buffer ( 10 mM Tris-HCl , 1 mM EDTA , pH 7 . 5 ) . DNA / Ros1HA complexes were eluted from the beads in TE SDS buffer ( 10 mM Tris-HCl , 1 mM EDTA , 1% SDS , pH 7 . 5 ) by incubation for 15 min at 65°C . Samples and input controls were de-crosslinked for 8–10 h at 65°C in TE SDS buffer containing 200 mM NaCl and 0 . 65 μg/μL proteinase K . DNA was purified using the ChIP DNA clean and concentrator kit ( Zymo research ) . The experiment was done in three biological replicates which were sequenced separately . DNA libraries were prepared using the Illumina TruSeq RNA sample preparation Kit v2 starting from the end repair step of the protocol . Up to 100 ng ChIP DNA was used as starting material . After end repair and A-tailing , indexing adapters were ligated . The products were then purified and amplified for 18 PCR cycles to create the final libraries . After validation ( Agilent 2200 TapeStation ) and quantification using the KAPA Library Quantification Kit ( VWR ) and the Applied Biosystems 7900HT Sequence Detection System , equimolar amounts of library were pooled . The pool was sequenced using the Illumina TruSeq PE Cluster Kit v3 and the Illumina TruSeq SBS Kit v3-HS ( 101 x 7 x 101 Cycles ) on an Illumina HiSeq 2000 sequencer with a paired-end ( 101 x 7 x 101 cycles ) protocol . Sequencing data were mapped to U . maydis genome ( ftp://ftpmips . gsf . de/fungi/Ustilaginaceae/Ustilago_maydis_521/ ) and analyzed using the ChIP-seq tool of the CLC genomics workbench 7 . 5 software ( CLC bio ) . ChIP-peak discovery was based on read coverage enrichment ( p-value ) and shape of read distribution ( peak shape score ) . ChIP-seq data were submitted to GeneExpressionOmnibus ( http://www . ncbi . nlm . nih . gov/geo/ ) under the accession number GSE76231 . Gene expression analysis from infected plant material was performed as described in Brefort et al . ( 2014 ) [87] , with some modifications . Briefly , samples of U . maydis cells grown in YEPSL and of infected tissue were ground to powder on liquid nitrogen and RNA was extracted with TRIzol ( Life technologies ) . After extraction , the first-strand cDNA synthesis kit ( Life technologies ) was used to reverse transcribe 1–2 μg of total RNA with oligo ( dT ) Primers . The qPCR analysis was performed using the SYBR Green Supermix ( Life technologies ) and an iCycler ( Bio-Rad ) . Cycling conditions were 2 min 95°C , followed by 45 cycles of 30 s 95°C/30 s 62°C/30 s 72°C . The experiment was done in three biological and three technical replicates and gene expression levels were calculated relative to the expression levels of the constitutively expressed fungal gene encoding peptidyl prolyl isomerase ( ppi ) . Primers used were ppi_qF / R for the reference gene ppi and primer pairs ros1_qF / R , rum1_qF / qR , ust1_qF / qR , hgl1_qF / qR , tup1_qF / qR , 05550_qF / qR , 04503_qF / qR , 02212_qF / qR , 01070_qf/qR , pks1_qF / qR , 04101_qF/qR , biz1_qF / qR , rbf1_qF / qR , fox1_qF / qR , mig2-3_qF / qR , 04096_qF / qR , dik1_qF / qR , 02473_qF / qR , and 03046_qF / qR for ros1 , rum1 , ust1 , hgl1 , tup1 , UMAG_05550 , UMAG_04503 , UMAG_02212 , UMAG_01070 , pks1 , UMAG_04101 , biz1 , rbf1 , fox1 , mig2-3 , UMAG_04096 , dik1 , UMAG_02473 , UMAG_03046 respectively . All primer sequences are listed in S6 Table . Relative expression was determined using the ΔΔCt method [88] . t-tests were used to assess statistically relevant differences between expression levels at different time points ( p ≤ 0 . 05 ) . Quantification of relative fungal biomass in infected maize leaves was performed as described previously [87] , with some modifications . 2 cm long sections from 10 leaves with the most prominent symptoms were harvested from 10 different plants at the indicated time points . For genomic DNA extraction leaf material was frozen in liquid nitrogen , ground to a fine powder , and extracted using a phenol-based protocol modified from Hoffman and Winston ( 1987 ) [89] . The qPCR analysis was performed using the Platinum SYBR Green Supermix ( Life technologies ) in an iCycler ( Bio-Rad ) . Cycling conditions were 2 min 95°C , followed by 45 cycles of 30 s 95°C / 30 s 62°C / 30 s 72°C . U . maydis biomass was quantified with primers ppi_qF/R amplifying the fungal ppi gene . Maize glyceraldehyde 3-phosphate dehydrogenase was amplified with primers Gapdh_qF/R and served as reference gene for normalization . The experiment was done in three biological and three technical replicates . t-tests were used to assess statistically relevant differences among strains ( p ≤ 0 . 05 ) . Wheat germ agglutinin-Alexa Fluor 488 / propidium iodide staining of infected leaf material was performed as described previously [38] . For staining of the mucilaginous matrix , leaf tumors from maize seedlings infected with strains FB1 x FB2 or FB1∆ros1 x FB2∆ros1 were collected at 10 dpi . Samples were fixed in 4% glutaraldehyde and embedded in Epoxy resin . 1–2 μm thick sections were generated with a microtome and stained with methylene blue-azure II-basic fuchsin following the protocol described by Humphrey and Pittman ( 1974 ) [49] . To examine fungal colonization of leaf tissue , samples from infected plants were fixed in ethanol , transferred to 10% KOH , incubated at 85°C for 4 hours , washed twice with PBS buffer ( 140 mM NaCl , 16 mM Na2HPO4 , 2 mM KH2PO4 , 3 . 5 mM KCl , and 1 mM Na2-EDTA , pH 7 . 4 ) , and incubated under vacuum in staining solution ( 10 μg/mL propidium iodide and 10 μg/mL WGA Alexa Fluor 488 in PBS , pH 7 . 4 ) according to Doehlemann et al . ( 2008 ) [38] . WGA Alexa Fluor 488 was purchased from Life technologies . To visualize the septa in strains expressing ros1 in axenic culture , cell walls were stained with calcofluor ( Fluorescent brightener 28 ) . Nuclei were stained with 4' , 6-diamidino-2-phenylindole ( DAPI ) . For microscopy , an Axioplan II microscope ( Zeiss ) with differential interference contrast optics was used . Fluorescence of GFP , mCherry , calcofluor and DAPI was observed using GFP ( ET470/40BP , ET495LP , and ET525/50BP ) , rhodamine ( HC562/40BP , HC593LP , and HC624/40BP ) , and DAPI ( HC375/11BP , HC409BS , and HC447/60BP ) filter sets ( Semrock ) . Pictures were taken with a CoolSNAP-HQ charge-coupled device camera ( Photometrics ) . Image processing was done with MetaMorph software ( Universal Imaging ) . Confocal microscopy was performed using a TCS-SP5 confocal microscope ( Leica Microsystems ) . GFP and wheat germ agglutinin-Alexa Fluor 488 were excited at 488 nm and emitted fluorescence was detected in the 495–530 nm range . Propidium iodide and mCherry were excited at 561 nm and emission was detected in the 580–630 nm range . Images were processed using LAS-AF software ( Leica Microsystems ) . To allow the recombinant expression of the Ros1WOPR domain comprising amino acids 1–321 of Ros1 , this domain was C-terminally fused to a His-tag ( Ros1WOPR-His ) . To this end ros1 was amplified with primers WOPR_NdeF / WOPR_XhoR , cloned in pET28 ( Novagen ) to generate pRos1WORP-His and transformed into E . coli strain BL21 ( DE3 ) pLysS ( Promega ) . Expression of Ros1WOPR-His was induced in exponentially growing cell cultures for 4h at 28°C in dYT medium supplemented with 0 . 15% glucose , 1 mM MgSO4 and 0 . 5 mM IPTG . To achieve cell lysis cell pellets were resuspended in BugBuster Master Mix reagent ( Merck Millipore ) supplemented with protease inhibitors ( complete EDTA-free tablet , Roche ) and incubated for 20 min at room temperature . The crude cell extract was then centrifuged ( 30000 x g; 30 min ) and the supernatant was loaded on a Ni-NTA column ( HisTrap FF Crude , GE Healthcare ) equilibrated in wash buffer ( 50 mM Na2HPO4 , 300 mM NaCl , 20 mM imidazole , pH 8 . 0 ) using an Äkta FPLC system ( GE Healthcare ) . After unbound protein was washed off the Ros1WOPR–His protein was eluted with elution buffer ( 50 mM Na2HPO4 , 300 mM NaCl , pH 8 . 0 , 135 mM imidazole ) . The Ros1WOPR-His containing fractions were pooled and loaded on a Superdex75 gel filtration column ( SuperdexTM 75 10/300GL , GE Healthcare ) equilibrated in gel filtration buffer ( 20 mM HEPES , 20 mM NaCl , 0 . 1 μM PMSF and 1 mM DTT ) . The Ros1WOPR-His containing fractions were pooled and incubated in batch with source 15Q anion exchange beads ( GE Healthcare ) equilibrated in gel filtration buffer . Ros1WOPR-His remained in the supernatant , which was concentrated via Amicon Ultra-4 centrifugation units with an Ultracel-3 membrane ( Merck Millipore ) . The protein was stored at 4°C for up to five days . For ros1 WT-probe and ORF-probes , pCros1 was used as template . For the mutated versions of the ros1 WT-probe , corresponding fragments were generated by annealing several overlapping oligonucleotides carrying the desired mutations and cloning them into p123 [81] between restriction NdeI and BamHI sites The inserts in the resulting plasmids pm1 , pm2 and pm1+2 were sequenced and plasmids were subsequently used as template for the respective PCR reactions using 5’ biotin labeled primers WT-probe_F/R for the WT-probe and the mutated versions m1 , m2 and m1+2 and primers ORF-probe_F/R for the ORF_probe . Wild type competitor DNA corresponding to the same sequence as WT-probe was PCR amplified using unlabeled WT-probe_F/R primers . Probes corresponding to promoter regions upstream of UMAG_02854 , UMAG_04040 , UMAG_02538 , cmu1 , UMAG_03046 , UMAG_03138 , UMAG_12258 , and UMAG_02775 were generated by PCR using 5’ biotin labeled primer pairs Probe-02854_F / R , Probe-04040_F / R , Probe-02538_F / R , Probe-cmu1_F / R , Probe-03046_F / R , Probe-03138_F / R , Probe-12258_F / R , Probe-02775_F1 / R1 and Probe-02775_F2 / R2 respectively . Competitors were generated by PCR using corresponding non labeled primers . All primer sequences are listed in S6 Table . EMSAs were performed using the Lightshift Chemiluminescent EMSA Kit ( Thermofischer ) . Binding reactions were carried out in 20 mM HEPES buffer , pH 8 . 0 supplemented with 1 mM DTT , 50 mM NaCl , 25 ng/μL poly dI/dC , 2 . 5% glycerol 0 . 05% NP40 , 5 mM MgCl2 and 1 μg/μL BSA . 10 fmol of biotin labeled dsDNA probe and 3 μg purified Ros1WOPR-His protein were used per binding reaction . For competition reactions , competitor fragment was added in a 500-fold molar excess . Binding reactions were incubated for 30 min at room temperature . In competition experiments Ros1WOPR was incubated for 20 min with the non-labeled competitor prior to addition of the probe . Binding reactions were separated on native 4% polyacrylamide 0 . 5 X TBE gels . Gels were transferred to a nylon membrane and biotin-labeled DNA fragments were detected using a streptavidin horseradish peroxidase conjugate and a highly sensitive chemiluminescent substrate as recommended by the manufacturer ( Thermofischer ) . ros1 ( UMAG_05853 ) : XM_011393913; UMAG_02775: XM_011390829; UMAG_01390: XM_011388973 | The fungus Ustilago maydis is a pathogen of maize which induces tumor formation in the infected tissue . In these tumors huge amounts of fungal spores develop . As a biotrophic pathogen , U . maydis establishes itself in the plant with the help of a large number of secreted effector proteins . Many effector proteins are important for virulence because they counteract plant defense reactions . In this manuscript we have identified and characterized Ros1 , a master regulator for the late stages of U . maydis development . This transcription factor is expressed late during infection and controls nuclear fusion , hyphal aggregation and late proliferation . ros1 mutants are still able to induce tumor formation but these are a dead end because they do not contain any spores . We show that Ros1 interferes with the early regulatory cascade controlled by a complex of two homeodomain proteins . In addition , Ros1 triggers a major switch in the effector repertoire , suggesting that different sets of effectors are needed for different stages of fungal development inside the plant . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"fungal",
"spores",
"ustilago",
"maydis",
"gene",
"regulation",
"regulatory",
"proteins",
"dna-binding",
"proteins",
"cereal",
"crops",
"fungi",
"regulator",
"genes",
"model",
"organisms",
"plant",
"science",
"transcription",
"factors",
"crops",
"gene",
"types",
"plan... | 2016 | The WOPR Protein Ros1 Is a Master Regulator of Sporogenesis and Late Effector Gene Expression in the Maize Pathogen Ustilago maydis |
Cells employ multiple levels of regulation , including transcriptional and translational regulation , that drive core biological processes and enable cells to respond to genetic and environmental changes . Small-molecule metabolites are one category of critical cellular intermediates that can influence as well as be a target of cellular regulations . Because metabolites represent the direct output of protein-mediated cellular processes , endogenous metabolite concentrations can closely reflect cellular physiological states , especially when integrated with other molecular-profiling data . Here we develop and apply a network reconstruction approach that simultaneously integrates six different types of data: endogenous metabolite concentration , RNA expression , DNA variation , DNA–protein binding , protein–metabolite interaction , and protein–protein interaction data , to construct probabilistic causal networks that elucidate the complexity of cell regulation in a segregating yeast population . Because many of the metabolites are found to be under strong genetic control , we were able to employ a causal regulator detection algorithm to identify causal regulators of the resulting network that elucidated the mechanisms by which variations in their sequence affect gene expression and metabolite concentrations . We examined all four expression quantitative trait loci ( eQTL ) hot spots with colocalized metabolite QTLs , two of which recapitulated known biological processes , while the other two elucidated novel putative biological mechanisms for the eQTL hot spots .
Cells are complex molecular machines that employ multiple levels of regulation that enable them to respond to genetic and environmental perturbations . Advances in biology over the past several years to elucidate the complexity of this regulation have been truly astonishing . However , despite transformative advances in technology , it remains difficult to assess where we are in our understanding of cell regulation , relative to a complete comprehension of such a process . One of the primary difficulties in our making such an assessment is that the suite of research tools available to us seldom provides insights into aspects of the overall picture of the system that are not directly measured . While different technologies provide information that our analytical tools , both algorithmic and intellectual , seek to combine into a coherent picture , one of the primary limitations of the majority of analytical tools in use today is a focus on single dimensions of data , rather than on maximally integrating data across many different dimensions simultaneously to view processes more completely , thereby achieving a greater understanding of these processes . The full suite of interacting parts in a cell over time , if they could be viewed collectively , would enable our achieving a more complete understanding of cellular processes , much in the same way we achieve understanding by watching a movie . The continuous flow of information in a movie enables our minds to exercise an array of priors that provide context and constrain the possible relationships ( structures ) , while our internal network reconstruction engine pieces all of the information together regarding the highly complex and nonlinear relationships represented in the movie , so that in the end we are able to achieve an understanding of what is depicted at a hierarchy of levels . If instead of viewing a movie as a continuous stream of frames of coherent pixels and sound , we viewed single dimensions of the information independently , understanding would be difficult if not impossible to achieve . For example , consider viewing a movie as independent , one dimensional slices through the frames of the movie , where each slice is viewed as pixel intensities across that one dimension changing over time ( like a dynamic mass spec trace ) . In this way it would be very difficult to understand the meaning of the movie by looking at all of the one dimensional traces independently . Despite the complexity of biological systems , even at the cellular level , research in the context of large-scale high dimensional -omics data has tended to focus on single data dimensions , whether constructing coexpression networks on the basis of gene expression data , carrying out genome-wide association analyses on the basis of DNA variation information , or constructing protein interaction networks on the basis of protein–protein interaction data . While we achieve some understanding in this way , progress is limited because none of the dimensions on their own provide a complete enough context within which to interpret results fully . This type of limitation has become apparent in genome-wide association studies ( GWAS ) , where many hundreds of highly replicated loci have been identified and highly replicated as associated with disease; but our understanding of disease is still limited because the genetic loci do not necessarily inform on the gene affected , on how gene function is altered , or more generally , how the biological processes involving a given gene are altered [1]–[4] . It is apparent that if different biological data dimensions could be formally considered simultaneously , we would achieve a more complete understanding of biological systems [2] , [3] , [5]–[7] . ( See the documentary film The New Biology at http://www . youtube . com/watch ? v=sjTQD6E3lH4 . ) Therefore , to form a more complete understanding of biological systems , we must not only evolve technologies to sample systems at ever higher rates and with ever greater breadth , we must innovate methods that consider many different dimensions of information to produce more descriptive models ( movies ) of the system . Methods are emerging that integrate pairs of data dimensions . For example , we recently developed methods that simultaneously integrate DNA variation and RNA expression data generated in a population context to identify coherent modules of interconnected gene expression traits driven by common genetic factors [2] , [8] . In addition , many groups have begun incorporating a time dimension in the context of high-dimensional molecular-profiling data to elucidate how networks can transform over time [9] , [10] . Here we develop and apply a network reconstruction approach that simultaneously integrates six different types of data: endogenous metabolite concentration , RNA expression , DNA variation , DNA–protein binding , protein–metabolite interaction , and protein–protein interaction data , to construct probabilistic causal networks that elucidate the complexity of cell regulation ( Figure 1 ) . The goals of our integrative analysis are not only to find causal regulators underlying expression quantitative trait loci ( eQTL ) hot spots , but to uncover mechanisms by which these predicted causal regulators affect genes and metabolites whose transcriptional profiles or metabolite profiles are linked to the eQTL hot spots . We leveraged a previously described cross between laboratory ( BY ) and wild ( RM ) yeast strains ( referred to here as the BXR cross ) for which DNA variation and RNA expression had been assessed [11] , [12] , to carry out a quantitative metabolite profiling using quantitative NMR ( qNMR ) under the same experimental conditions as the gene expression study [12]–[14] . We demonstrate that , like transcript and protein levels , concentrations of many metabolites are strongly linked to metabolite QTLs ( metQTLs ) . Several of the metQTLs are seen to colocalize with expression quantitative trait loci ( eQTLs ) previously identified in the same yeast population [13] , enabling us to infer causal relationships between metabolites and expression traits [13] , [14] . Then , by extending a previously described Bayesian network ( BN ) reconstruction algorithm [13] , we constructed a probabilistic causal network by integrating metabolite levels , genotype , gene expression , transcription factor ( TF ) binding , and protein–protein interaction data . The resulting network not only validates the functional importance of eQTL hot spots in the BXR cross , but elucidates the mechanisms by which variation in DNA at eQTL hot spots affect gene expression . By systematically using the networks to elucidate the regulators of these eQTL hot spots , we are not only able to recapitulate known regulatory mechanisms , we are able to provide a number of novel and experimentally supported causal relationships predicted by our network , including that cellular amino acid concentrations are related to both amino acid biosynthesis pathways and amino acid degradation pathways , with VPS9 predicted and prospectively validated as a key driver of a previously identified eQTL hot spot that could not previously be well characterized . In addition , we further experimentally demonstrated that PHM7 , a previously predicted and validated causal regulator for stress response genes whose expression variations are linked to the PHM7 locus on Chromosome XV , affected trehalose , a yeast metabolite product of the stress response pathway . These results combined not only help uncover the mechanisms by which gene expression profiles are regulated by metabolite profiles , but they also confirm the importance of gene expression in understanding system-wide variation linked to genetic perturbations .
Given the strong genetic signal detected in the metabolite data and the coincidence of metQTL and eQTL hot spot regions , we set out to explore an integrated network analysis strategy using the gene expression profiles [11] as well as the metabolite data described above . Gene expression and metabolite traits were treated equivalently as nodes in our BN reconstruction process . As such , we modified our previously reported BN reconstruction method [13] to accommodate metabolite data , in addition to genotype , gene expression , protein interaction , and TF–DNA binding data . The KEGG biochemical pathway database [37] was used to generate structure priors between metabolites and genes encoding enzymes known to be involved in biochemical reactions in canonical pathways . Intuitively , genes encoding enzymes that directly catalyze biochemical reactions for the metabolites were assigned stronger prior probabilities of being related during network reconstruction , whereas genes that encode enzymes catalyzing downstream or upstream biochemical reactions of the metabolites were assigned weaker priors ( see Methods for details ) . Differentially regulated genes and the structure priors for genotype , TF–DNA , and protein–protein interaction data were defined as previously described [13] . The 56 reliably quantified metabolites were included as input into the BN reconstruction program . From this probabilistic causal network we can identify subnetworks for all of the metabolites or any set of genes ( see Methods for details ) . To assess the predictive power of this network , we examined how metabolites and gene expression traits relate to one another at the four eQTL hot spots in Table 1 , providing for the possibility of elucidating regulatory mechanisms and generating testable hypotheses about novel regulatory relationships .
By integrating six different fundamental types of data , including RNA expression , DNA variation , DNA–protein binding , protein–metabolite interaction , and protein–protein interaction data , with metabolite data , we constructed a BN using an approach that simultaneously considers all of these data , with the resulting network providing a number of novel insights into the mechanisms of the eQTL hot spots in a segregating yeast population ( the BXR cross ) . Importantly , we validated the biological consequences of the transcriptional variation linked to each of the four eQTL hot spots identified in the BXR cross to which metabolite levels were also linked . Our results indicate that the incorporation of metabolite levels into the network reconstruction process significantly enhanced the utility of the network-based models [46] , [47] . While the integration of metabolite abundance and gene expression traits in a genetic context have been attempted in plants [48] and mouse [49] , the main distinguishing characteristic of our study is the de novo construction of a global molecular network that simultaneously incorporates many different types of information ( DNA , RNA , protein , and metabolite ) , along with known biochemical pathways as prior information . To aid in further understanding how we integrate these data to construct probabilistic causal networks , and to enhance the ability to repeat our results , we provide as Text S1 results of an in-depth description of the construction of the URA3 subnetwork ( Figure 4 ) , using different types of data to assess the contributions of different data types to the predictive power of the network and to the identification of key modulators of important biological processes . We examined in detail all 4 eQTL hot spots that coincided with metQTLs . Our findings for eQTL hot spots 1 and 2 recapitulated well-known biological processes , and for eQTL hot spots 3 and 4 our predictions implicated novel genes as modulators of established biological processes , which we subsequently validated prospectively . Among the many predictions made by our network , we uncovered novel insights into the biological processes that in the BXR cross are responsible for variations in amino acid levels . While amino acid concentrations are known to be regulated by multiple processes ( e . g . , synthesis , degradation , recycle , and storage ) , our approach objectively identified that variations in concentrations of a number of amino acids in the BXR cross were affected by both the amino acid biosynthesis and degradation pathways . We predicted and prospectively validated VPS9 as a major driver of amino acid concentrations via the amino acid degradation pathway . These results open novel and interesting questions about the mechanism by which sequence variation at this locus affects phenotype . VPS9 is involved in vesicle-mediated vacuolar protein transport , and in Saccharomyces cerevisiae , the vacuole is the main compartment for amino acid storage , recycling , and cytosolic amino acid concentration maintenance [50] . The cellular effects of variation in VPS9 are likely mediated by differential regulation of amino acid storage in the vacuole; we speculate that such storage changes may affect cytosolic amino acid pools that in turn have downstream consequences on transcript and protein levels of amino acid pathways , as has been shown for CHA1 [40] and GCV3 [51] . However , only with enhanced screening of all molecular states of the systems can we achieve a complete understanding of these processes . Thus , while the integrated BN elucidated some of the mechanistic underpinnings of the eQTL hot spots in the BXR cross , additional information will be required to more fully understand how processes perturbed in the BXR cross lead to phenotypic changes . Despite lacking an exhaustive assessment of all molecular traits in the BXR cross , it is of particular note that the strong correlations we observed between gene expression and metabolite data may help resolve an ongoing debate regarding the functional consequences of gene expression regulation . While some reports indicate that gene expression levels and protein abundances are not well correlated [52] , other reports indicate a high degree of correlation [53] . A recent proteomic study in the BXR cross demonstrated that a large number of protein levels are linked to eQTL hot spots [34] , two of which ( the eQTL hot spots 1 and 3 ) were highlighted in our present work . Metabolites are the final functional products of protein activity regulation . We showed that PHM7 not only alters expression levels of stress response genes linked to eQTL hot spot 4 , but also alters the abundance of trehalose , a metabolite product of the stress response genes . Our results demonstrate that gene expression and metabolite levels are not only strongly correlated , but that a significant proportion of that covariation can be explained by common genetic control . Given that variations in protein levels can result from sequence-specific transcriptional and translational regulation or from nonsequence-specific protein degradation , the integration of gene expression and metabolic traits can help dissect the complex processes that regulate protein levels . The yeast growth conditions for metabolite profiling were the same as previously used to generate the gene expression data in the BXR cross [12] . Both gene expression and metabolite abundances are under strong genetic regulation and are linked to common eQTL hot spots ( Table 1 ) . When metabolite data were integrated with gene expression data , our resulting integrated network was able to recapitulate the mechanism of multiple known biological processes that in turn explained the connection between genes linked to the LEU2 locus and genes with Leu3 binding sites , with the metabolite 2-isopropylmalate objectively identified as the key intermediate . These results also confirmed that changes in expression of stress response genes lead to changes in stress response metabolites such as trehalose . Therefore , the integration of the gene expression and metabolite data has provided new insight into common biological processes that are perturbed by genetic variation segregating in the BXR cross . Going forward , as more technologies emerge that can generate large-scale data in different dimensions for low cost , we will achieve a more complete understanding of biological systems only if we integrate all of the information together to consider all of the different cellular components and how they interact with one another at the population level . For example , comprehensive proteomic data and protein phosphorylation data are needed and should be further integrated with other high throughput genomic and genetic data . For metabolites , their cellular abundances are not only affected by specific enzymes in related biochemical reactions , but they are also affected by proteins that bind them or transport them into different cellular compartments . Further research on how to integrate these data into networks is needed . In addition , there is an abundance of existing knowledge , such as genetic interactions and regulatory cascades , which can be converted into prior information and integrated with other data and priors . Further efforts in developing methods to integrate these diverse data and information are warranted . In more complex systems , we will need to consider the fundamental building blocks of a cell in the context of cell–cell interactions that lead to tissue-based networks , the interactions of tissues that lead to organ-based networks , and the interactions of organs in a given system to understand the physiological states of that system associated with complex phenotypes of interest , given these phenotypes emerge from this complex web of interacting networks [54] . Only by taking the full complement of raw data available on living systems can we move from the accumulation of knowledge to actual understanding , and from understanding , wisdom .
Yeast parental strains BY4716 ( MATα lys2Δ0 ) and RM11-1a ( MATa leu2Δ0 ura3Δ0 HO:kan ) and 111 segregants of BXR cross [11] were provided by R . Brem . Auxotrophies , mating type , and G418 resistance were confirmed for all strains to be as previously reported [12] . Cells were grown under identical conditions as previously described [12] . Strains were freshly started from freezer stocks and stored at room temperature on synthetic complete medium plates for no longer than 1 wk before each experimental run . For each run , cells from the plates were precultured in 10 ml of synthetic complete media ( Table S8 ) at 30°C with shaking for 24 h . These cultures were then diluted into 25 ml fresh synthetic complete media to an optical density of 0 . 005 to 0 . 02 . This starting density was determined from previous growth rate measurements and empirical observations such that after overnight growth at 30°C , the cultures would be exponentially growing , i . e . , at a cell density of less than 2×107 cells/ml . Overnight cultures were diluted into 52 ml fresh synthetic complete medium to an optical density of 0 . 1 , and incubated with shaking for approximately 5 h at 30°C . Starting at 3 h after dilution , optical density was monitored every 60 min . Cell suspensions were counted in a hemocytometer to obtain cell count per OD values and an estimate of cell-doubling time . Since some of the yeast strains produced flocculent cultures under these growth conditions , all cultures were diluted 5× into 0 . 25 ml PBS and sonicated three times on ice for 45 s using a Misonix sonicator 3000 equipped with a microprobe before optical density was determined and/or cells were counted . At an optical density of approximately 1 . 0 , each exponentially growing culture was concentrated 10-fold by rapid centrifugation at room temperature and suspension of the cells in 5 ml of synthetic complete medium prewarmed to 30°C . These concentrated cell suspensions were then incubated at 30°C with shaking for 1 h . Metabolites were then immediately extracted from the cells in these concentrated suspensions . Yeast parental strain BY4742 ( MATα his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 ) and six deletion strains derived from it ( Δtsl1::kanMX , Δnup188::kanMX , Δcac2::kanMX , Δyml096w::kanMX , Δvps9::kanMX , and Δarg81::kanMX ) were provided by Elton Young's lab , Department of Biochemistry , University of Washington , from a copy of the Yeast Deletion Consortium knockout collection prepared in Stanley Fields' lab , Department of Genome Sciences , University of Washington . Cells were grown under identical conditions as the BXR cross strains in synthetic complete medium , and metabolite extracts were also obtained and further processed in identical fashion ( see below ) . Each experiment was repeated on three different days . Yeast parental strain BY4743 ( MATa/MATα his3Δ1/his3Δ1 leu2Δ0/leu2Δ0 lys2Δ0/+ met15Δ0/+ ura3Δ0/ura30 ) was obtained from ATCC ( Manassas , Virginia ) , and the derived PHM7 knockout strain 31775 ( phm7::KanMX/phm7::KanMX ) constructed by the Yeast Deletion Project [55] was obtained from Open Biosystems ( Huntsville , Alabama ) . Cells were grown under identical condition as the PHM7 knockout gene expression experiment [13] , and metabolites were extracted as described below . Each experiment was repeated on three different days . BY4742 and Δvps9 strains ( both MATα ) were grown as described above and harvested by centrifugation in crushed ice when cells reached optical density of approximately 1 . 0 . Total RNA was extracted using RNeasy mini-columns , transcribed with SMARTScribe Reverse Transcriptase ( Clontech ) from oligo ( dT ) , and diluted 1 , 000× . Real-time PCR was run for 17 genes ( including VPS9 ) associated with the Chromosome XIII eQTL hot spot subnetworks and ACT1 internal standard gene ( Table S6 ) on an ABI 7900HT instrument with 2× Sensimix dT ( Quantance ) , primers at 0 . 2 µmol/l , and SYBR Green reagent . Relative expression was calculated using the ΔΔCt method with ACT1 internal standards [56] . TAF9 was used to estimate the false positive rate as 0 . 033 . Intracellular metabolites were extracted using a modification of previously described methods [31] , [57] . First , all intracellular metabolic processes were rapidly quenched by pipetting each concentrated cell suspension into 20 ml of rapidly mixing 60% ( v/v ) methanol at −40°C . Cells were rapidly ( 5 min ) sedimented in a centrifuge precooled to −8°C and washed twice with 20 ml of the −40°C methanol . Metabolites were then extracted with boiling 75% ( v/v ) ethanol at 80°C and 0 . 25 ml dry volume of acid-washed glass beads ( Sigma G1277 ) , by vigorous vortexing for 30 s . The cell-glass bead slurry was incubated 3 min at 80°C , vortexed 30 s , and then placed on ice for 5 min . Large cellular debris and glass beads were removed by centrifugation at 2 , 000 g for 5 min . The resulting ethanolic extracts were clarified by three rounds of centrifugation at 14 , 000 g in a microcentrifuge . The clarified metabolite extracts were stored at −80°C until drying . Extracts were dried in a Savant Speed Vac under 150 mtorr vacuum in low retention microcentrifuge tubes . Dried metabolite extracts were stored at −80°C until preparation for NMR analysis . The process of NMR spectra acquisition and quantification follows the previously outlined procedure [29] . Dry metabolite extracts were dissolved in 0 . 7 ml deuterated 80 mM potassium phosphate buffer ( containing 2 mM DSS-d6 as an internal reference standard ) , and transferred to 100-mm 5-ml NMR tubes . NMR samples were stored in Varian 768AS auto-sampler at 8°C before and after NMR analyses . NMR data were acquired on the Varian 700 MHz NMR spectrometer at 25°C with one-dimensional proton pulse sequence . The water peak was suppressed by the WET pulse sequences [58] . For each sample , 512 acquisitions were acquired with 3 s of acquisition and 15 s of delay between pulses . Analyses of NMR spectra were carried out using DataChord Spectrum Miner ( One Moon Scientific , Inc . ) . Stacked NMR spectra were referenced to DSS-d6 as 0 ppm , and peaks of each endogenous metabolite were checked against their reference spectra ( about 700 common endogenous metabolites ) . Each metabolite usually displays multiple peaks , for example trehalose , shown in Figure S6 . Overlapping peaks were quantified by peak area correction according to stoichiometric peak ratios for each metabolite . For three correlated variables , , and , there are three groups of causal/reactive relationships among them as the following:For example , the three graphs in the group G1 , , describe the same set of condition independent relationship that and are independent conditioning on . The three graphs have the same probabilities and are called Markov equivalent . In an F2 cross , we can represent quantitative traits as and , and the genetic locus as . In an F2 cross experimental design , all F2 strains are under the same experimental condition . Therefore , the only source of variation in the quantitative traits and are genetic differences in , so that the relationships and are plausible . On the other hand , the genetic variation in is stable and does not change during an F2 cross experiment , so that and are not plausible . Thus in an F2 cross , only one graph in each of the three Markov equivalent groups above is plausible . We can simplify the above three groups as follows:where the genetic locus is the anchor in each causal/reactive relationship in a F2 cross . For two quantitative traits and linked to the same locus in the yeast cross , there are three basic relationships that are possible between the two traits relative to the DNA locus as described above . Either DNA variations at the locus lead to changes in trait that in turn lead to changes in trait , or variations at locus lead to changes in trait that in turn lead to changes in trait , or variations at locus independently lead to changes in traits and , as previously described [14] . Assuming standard Markov properties for these basic relationships , the joint probability distributions corresponding to these three models , respectively , are: where the final term on the right-hand side of equation M3 reflects that the correlation between and may be explained by other shared loci or common environmental influences , in addition to locus . We assume Markov equivalence between and for model M3 so that . is the genotype probability distribution for locus and is based on a previously described recombination model [59] . The random variables and are taken to be normally distributed about each genotypic mean at the common locus , so that the likelihoods corresponding to each of the joint probability distributions are then based on the normal probability density function , with mean and variance for each component given by: ( 1 ) for the mean and variance are and , ( 2 ) for the mean and variance are and , and ( 3 ) for the mean and variance are and , where represents the correlation between and , and and are the genotypic specific means for and , respectively . The mean and variance for follow similarly from that given for . From these component pieces , the likelihoods for each model are formed by multiplying the densities for each of the component pieces across all of the individuals in the population [14] . The likelihoods are then compared among the different models in order to infer the most likely of the three . Because the number of model parameters among the models differs , a penalized function of the likelihood was used to avoid the bias against parsimony . The model with the smallest value of the penalized statisticwas chosen . Here , is the maximum likelihood for the ith model , pi is the number of parameters in the ith model , and k is a constant . In this instance we took the penalized statistic to be the Bayesian Information Criteria ( BIC ) where k is set to , with n denoting the number of observations . BNs are directed acyclic graphs in which the edges of the graph are defined by conditional probabilities that characterize the distribution of states of each node given the state of its parents [60] . The network topology defines a partitioned joint probability distribution over all nodes in a network , such that the probability distribution of states of a node depends only on the states of its parent nodes: formally , a joint probability distribution on a set of nodes can be decomposed as , where represents the parent set of . In our networks , each node represents a quantitative trait that can be a gene or a metabolite . These conditional probabilities reflect not only relationships between genes , but also the stochastic nature of these relationships , as well as noise in the data used to reconstruct the network . Bayes formula allows us to determine the likelihood of a network model M given observed data D as a function of our prior belief that the model is correct and the probability of the observed data given the model: . The number of possible network structures grows super-exponentially with the number of nodes , so an exhaustive search of all possible structures to find the one best supported by the data is not feasible , even for a relatively small number of nodes . We employed Monte Carlo Markov Chain ( MCMC ) [61] simulation to identify potentially thousands of different plausible networks , which are then combined to obtain a consensus network ( see below ) . Each reconstruction begins with a null network . Small random changes are then made to the network by flipping , adding , or deleting individual edges , ultimately accepting those changes that lead to an overall improvement in the fit of the network to the data . We assess whether a change improves the network model using the BIC [62] , which avoids overfitting by imposing a cost on the addition of new parameters . This is equivalent to imposing a lower prior probability on models with larger numbers of parameters . Even though edges in BNs are directed , we can't in general infer causal relationships from the structure directly . For example , in a network with two nodes , and , the two models and have equal probability distributions as . Thus , with correlation data itself , we can't infer whether is causal for or vise versa . In the more general case , for a network with three nodes , , , and , there are multiple groups of structures that are mathematically equivalent . For example , the following three different models , , , and , are Markov equivalent ( which means that they all encode for the same conditional independent relationships ) . In the above case , all three structures encode the same conditional independent relationship , , and are independent conditioning on , and they are mathematically equalThus , we can't infer whether is causal for or visa-versa from these types of structures . However , there is a class of structures , V-shape structures ( e . g . , ) , which have no Markov equivalent structure . In this case , we can infer causal relationships . There are more parameters to estimate in the Mv model than M1 , M2 , or M3 , which means a large penalty in the BIC score for the Mv model . In practice , a large sample size is needed to differentiate the Mv model from the M1 , M2 , or M3 models . The same 3 , 662 informative genes used previously [13] and 56 metabolites were included in the network reconstruction process using a BN reconstruction software program based on a previously described algorithm [63] , [68] as outlined above . One thousand BNs were reconstructed using different random seeds to start the reconstruction process . From the resulting set of 1 , 000 networks generated by this process , edges that appeared in greater than 30% of the networks were used to define a consensus network . Our previous simulation study shows that the 30% inclusion threshold results in a stable structure and achieves the best tradeoff between precision and recall [68] . The histogram of percentage of occurrences of all potential edges shows that 30% is a reasonable cutoff threshold for inclusion ( Figure S8 ) . Edges in this consensus network were removed if ( 1 ) the edge was involved in a loop , and ( 2 ) the edge was the most weakly supported of all edges making up the loop . The genetic , TFBS , and PPI data were used to derive structure priors as previously described ( details described above in Methods ) [13] . Structure priors for metabolites and genes are derived from KEGG chemical reactions as described above . All data and software used to construct the BNs described herein are available at http://www . mssm . edu/research/institutes/genomics-institute/rimbanet . Subnetworks for sets of genes were constructed as follows . Genes in the input set were used as seeds and the direct neighbors of seeds were identified . Seeds and their direct neighbors define the nodes of a given subnetwork . Links between nodes in the subnetworks are the same as in the complete BN . | It is now possible to score variations in DNA across whole genomes , RNA levels and alternative isoforms , metabolite levels , protein levels and protein state information , protein–protein interactions , and protein–DNA interactions , in a comprehensive fashion in populations of individuals . Interactions among these molecular entities define the complex web of biological processes that give rise to all higher order phenotypes , including disease . The development of analytical approaches that simultaneously integrate different dimensions of data is essential if we are to extract the meaning from large-scale data to elucidate the complexity of living systems . Here , we use a novel Bayesian network reconstruction algorithm that simultaneously integrates DNA variation , RNA levels , metabolite levels , protein–protein interaction data , protein–DNA binding data , and protein–small-molecule interaction data to construct molecular networks in yeast . We demonstrate that these networks can be used to infer causal relationships among genes , enabling the identification of novel genes that modulate cellular regulation . We show that our network predictions either recapitulate known biology or can be prospectively validated , demonstrating a high degree of accuracy in the predicted network . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"systems",
"biology",
"genomics",
"biology",
"computational",
"biology",
"genetics",
"and",
"genomics"
] | 2012 | Stitching together Multiple Data Dimensions Reveals Interacting Metabolomic and Transcriptomic Networks That Modulate Cell Regulation |
Serological studies are the gold standard method to estimate influenza infection attack rates ( ARs ) in human populations . In a common protocol , blood samples are collected before and after the epidemic in a cohort of individuals; and a rise in haemagglutination-inhibition ( HI ) antibody titers during the epidemic is considered as a marker of infection . Because of inherent measurement errors , a 2-fold rise is usually considered as insufficient evidence for infection and seroconversion is therefore typically defined as a 4-fold rise or more . Here , we revisit this widely accepted 70-year old criterion . We develop a Markov chain Monte Carlo data augmentation model to quantify measurement errors and reconstruct the distribution of latent true serological status in a Vietnamese 3-year serological cohort , in which replicate measurements were available . We estimate that the 1-sided probability of a 2-fold error is 9 . 3% ( 95% Credible Interval , CI: 3 . 3% , 17 . 6% ) when antibody titer is below 10 but is 20 . 2% ( 95% CI: 15 . 9% , 24 . 0% ) otherwise . After correction for measurement errors , we find that the proportion of individuals with 2-fold rises in antibody titers was too large to be explained by measurement errors alone . Estimates of ARs vary greatly depending on whether those individuals are included in the definition of the infected population . A simulation study shows that our method is unbiased . The 4-fold rise case definition is relevant when aiming at a specific diagnostic for individual cases , but the justification is less obvious when the objective is to estimate ARs . In particular , it may lead to large underestimates of ARs . Determining which biological phenomenon contributes most to 2-fold rises in antibody titers is essential to assess bias with the traditional case definition and offer improved estimates of influenza ARs .
Each year , seasonal influenza is responsible for about three to five millions severe illnesses and about 250 , 000 to 500 , 000 deaths worldwide [1] . These epidemics can generate important economic losses due to high levels of worker absenteeism as well as a saturation of emergency services at the peak of the epidemic [1] . In addition , avian or swine influenza viruses occasionally adapt to humans and generate influenza pandemics like in 1918 , 1957 , 1968 and 2009 , sometimes with catastrophic consequences like in 1918 , when 20 to 50 million people died worldwide . Appropriate assessment of the epidemiological characteristics of the influenza virus is important to guide control policies . In particular , this requires being able to track the number of influenza cases with severe clinical outcomes ( i . e . the tip of the severity pyramid ) as well as the total number of people infected by an influenza virus ( i . e . the base of the severity pyramid ) . For example , the case fatality ratio ( proportion of influenza cases who die ) is a key measure of severity that informs decision making during influenza pandemics , and takes the number of influenza related death as numerator and the number of influenza cases as denominator . Estimates of infection attack rates are also essential for characterizing the spread of the virus in human populations in order to predict epidemic trajectory , the potential impact of control measures such as social distancing measures , and the likelihood and magnitude of subsequent epidemics arising from continued circulation of the same virus [2] , [3] . Although it is usually possible to estimate the number of severe influenza cases from sentinel surveillance ( e . g . based on data collected at medical practices , clinics or hospitals ) , it is much harder to estimate the total number of people infected by an influenza virus . First , a substantial proportion of influenza infections are asymptomatic [4] , [5] . Second , among those with symptoms , only a proportion seek healthcare; and this proportion may vary from season to season or even during the course of an epidemic . Last , Influenza-Like-Illness ( ILI ) symptoms are not specific to influenza . So , a substantial proportion of patients consulting for ILI may not have been infected by an influenza virus . Serological studies have become the gold standard approach for estimating influenza infection attack rates due to the difficulty of estimating infection rates by other means . Although cross-sectional serological surveys can provide valuable and timely information , paired blood samples collected before and after an epidemic in a cohort of individuals is the optimal approach for precisely assessing infection rates . The haemagglutination-inhibition ( HI ) assay remains the most commonly used approach for detecting serological evidence of recent influenza infection [6]–[12] . The assay detects the presence of antibodies that prevent the haemagglutinin protein of the influenza virus from agglutinating red blood cells [13] , [14] . For each serum sample , antibody titers are expressed as the reciprocal of the highest serum dilution that can still prevent a fixed concentration of virus from agglutinating red blood cells . A rise in antibody titers between the first and second blood is taken as a marker of infection . However , because the procedure is susceptible to measurement errors , a 2 fold rise ( that is a 1-dilution increase ) is usually considered as insufficient evidence for infection . Seroconversion is therefore typically defined as a 4-fold rise ( i . e . a 2-dilutions increase ) or more in antibody titers . This ad-hoc rule became established when these methods were first developed and is now widely adopted [15] , [16] . In the meantime , however , statistical methods for addressing measurement errors have made substantial progress . In particular , there is now an extensive body of literature on methods to ensure that the presence of measurement errors does not bias estimates of key parameters of interest . Given these developments , it is timely to revisit the way serological data are interpreted . Central to the traditional approach to analyzing serological data is the belief that data about 2-fold rises provide no information since such increases can be caused by frequent measurement errors . This concern about measurement errors is certainly relevant when trying to make specific diagnoses for individual cases . For example , one may be averse to the risk of false positives; but less so to the risk of false negatives . However , estimating infection attack rates at the population level is a very different aim from setting up a specific diagnostic tool , and may benefit from a different use of the data . First , it is important to note that estimating infection attack rates is not just a matter of specificity ( i . e . ensuring that subjects satisfying the diagnostic definition of infection were indeed infected by an influenza virus ) but also a matter of sensitivity ( i . e . ensuring that all subjects infected are diagnosed as such ) . An approach that favours specificity over sensitivity may lead to underestimating infection attack rates . A second important observation is that , even in a context of frequent 2-fold errors , data about 2-fold rises may still be informative . Consider for example a situation where all individuals exhibit a 2-fold rise during the season: such a pattern cannot be explained by measurement error alone since measurement errors are made both at baseline and post-epidemic and should be about equally distributed provided the sample size is sufficiently large . Here , we explore how modern statistics for the analysis of data with measurement errors can change and improve our interpretation of serology . We present a new method to quantify errors in the measurement of antibody titers and to estimate the true distribution of paired serological measurements corrected for measurement errors . The methodology is applied to data collected in a cohort study conducted in Vietnam between 2007 and 2009 .
We estimate that the 1-sided probability of a 2-fold error was 9 . 3% ( 95% CI: 3 . 3% , 17 . 6% ) when the true antibody titer was below detection levels , rising to 20 . 2% ( 95% CI: 15 . 9% , 24 . 0% ) otherwise ( posterior probability that latter larger than former: 98 . 7% ) . There was a satisfying fit of the model to replicate measurement data ( Figure 1 ) . The model where measurement errors were independent of true antibody titers failed to fit the data ( Figure S2 and Supplementary Material ) . Figure 2 summarizes the distribution of paired serology , corrected for measurement errors for the different seasons ( 2008 , Spring 2009 , Autumn 2009 ) and subtypes ( H1N1 , H3N2 and B ) . A range of observations can be made . The first observation concerns 2-fold rises in antibody titers between baseline and post serology ( yellow bars ) . Such increases are usually ignored in analyses because 2-fold errors are common . In some instances , like for example subtypes H3N2 and B in 2008 and H1N1pdm09 in Autumn 2009 , 2-fold rises appeared negligible and at levels that could be generated by measurement errors alone , since 0 was within the 95% CI of the estimated proportion of subjects having a 2-fold rise ( Figures 2B , 2C , 2G ) . In other instances , however , the proportion of individuals experiencing a 2-fold rise ranged from 20% to 33% with lower bounds of the 95% CIs above 0 ( range: 7%–23% ) , indicating that these rises cannot be solely explained by measurement errors . Assuming that most of these 2-fold rises were due to infection , our estimate of infection attack rates for H1N1 in 2008 and H1N1 , H3N2 and B in Spring 2009 would be dramatically higher than traditional estimate based on 4-fold rises or more ( Figure 3A ) . So , even if only a proportion of the 2-fold rises were due to influenza infections , the traditional estimate might still represent a substantial underestimate of the true infection attack rates The fact that and were very similar for H3N2 and B in 2008 and virtually identical for H1N1pdm09 in Autumn 2009 ( Figure 3A ) highlights important heterogeneities in the way antibody titers increase by season/subtype ( Figure 3B ) . For example , for H1N1pdm09 in Autumn 2009 , almost all those experiencing a rise in antibody titers exhibited a 4-fold rise or more; but for H1N1 in 2008 , most of those experiencing a rise only had a 2-fold increase . The absence of a simple linear relationship between and the proportion of 2-fold rises suggests that the standard approach of inflating by a fixed proportion ( generally equal to the proportion of PCR positive cases who do not seroconvert; around 10–20% ) to get corrected estimates of infection attack rates may be inappropriate . Rather , corrections might have to be applied on a season-to-season and subtype-to-subtype basis . The last notable observation is that decay in antibody titers is observed . For example , 30% ( 95% CI: 22 , 36 ) of individuals exhibited a decay for subtype H3N2 in 2008 . Figure 4 shows the observed rise in antibody titers for PCR positive cases . Twenty seven percent of these cases experienced no rise or only a 2-fold rise in titer during the season . This again suggests that the case definition of a 4-fold rise or more may underestimate attack rates by at least 27% . PCR positive cases with low baseline titers experienced an average increase significantly larger than those with higher baseline titers ( p = 0 . 026 ) ( Figure 4 ) [17] , [18] . Simulations were run to test the hypothesis of an absence of cross-reactivity between subtypes H1N1 , H3N2 and B in 2008 and Spring 2009 ( see Supplementary Material ) . We found that there was good adequacy between the data and patterns that would be obtained in the absence of cross-reactivity . The hypothesis of an absence of cross-reactivity could therefore not be rejected ( Figure S3 ) . Figure 5 compares the distribution of observed paired serology as observed in the data ( black point ) and as predicted by the model . Model fit was satisfactory . In a simulation study , we found that estimates of parameters characterizing measurement errors were unbiased ( Table 1 ) , as well as those characterizing the selection process ( Table S2 ) . We also found that estimates of the proportion of subjects with an antibody titer increase ( empirical absolute bias: 0 . 1% ) , of the proportion of subjects with an antibody titer decay ( empirical absolute bias: 0 . 0% ) and of the probabilities characterizing jointly baseline antibody titers and the change in antibody titers during a season ( empirical absolute bias: 0 . 0% ) were unbiased ( Figure 6 ) . Our statistical model describes the distribution of paired serology across all subjects . However , since we infer true paired serology for each individual , it is possible to reconstruct a posteriori the distribution of true paired serology for the different age groups . The age-specific distributions for true paired serology are presented in Figure S4 . Interesting differences can be noticed between age groups . For example and consistent with the literature , for H1N1pdm09 in Autumn 2009 , the proportion of 4-fold rises falls from 39% ( 95% CI: 37% , 39% ) in <18 y . o . to 15% ( 95% CI: 15% , 16% ) in 18–48 y . o . and 8% ( 95% CI: 7 , 9 ) in >48 y . o . For H3N2 in 2009 , the decay in antibody titers was more important among <18 y . o . ( 53%; 95% CI: 38% , 65% ) than among older age groups ( 25% , 95% CI 19% , 30% for 18–48 y . o . and 18% , 95% CI 12 , 22 for >48 y . o . ) . For H3N2 in Spring 2009 , although the proportions of 4-fold rises were similar across age groups , our analysis suggests that the proportion of 2-fold rises may have been higher among <18 y . o ( 43% , 95% CI: 23 , 58 ) than in other age groups ( 30% , 95% CI 17% , 41% for 18–48 y . o . and 27% , 95% CI 13 , 38 for >48 y . o . ) . We find that , for each age group , there is a satisfying adequacy between the observed distribution of paired serology and that predicted by the model ( Figure S5 ) .
In this paper , we have revisited the traditional interpretation of paired serological measurements of influenza antibody titers . Until now , data on 2-fold rises have been largely ignored because of the belief that measurement errors made them unreliable . Although this may be a valid concern if the aim is to get a specific diagnosis for individual cases , we argue that this is less so when the objective is to interpret antibody titer variations at the population level . We have shown that it is possible to quantify measurement errors , and to reconstruct the distribution of paired serology corrected for measurement errors . Our method gave unbiased estimates in a simulation study . After correction for measurement errors for the Vietnamese data examined here , we found that for some seasons and subtypes the proportions of individuals with 2-fold rises in antibody titers was too large to be explained by measurement errors alone . Estimates of infection attack rates varied greatly depending on whether or not 2-fold rises were included . It is therefore important to determine the biological phenomenon that could cause such increases , in particular whether they are caused by exposure to influenza viruses . A first hypothesis is that 2-fold titer increases are caused by infection by an influenza virus . In support of this hypothesis , it is clear that a proportion of virologically- or RT-PCR- confirmed influenza cases do not achieve a 4-fold rise in HI titer . This proportion was 27% in our dataset , similar to a large cohort of confirmed pandemic cases in the US [19] . However , past work has shown this proportion to be as high as 77% in people who have high pre-existing antibody titers [17] , or as low as 10% in patients seeking medical care for pandemic H1N1 infection in 2009 [20] . It is clear that antibody titer changes following infection vary between individuals and are affected by factors including pre-existing titer and timing of serum collection . In particular , since there is an upper limit to antibody concentrations , individuals with high pre-existing titers are limited in their ability to generate 4-fold rises and may produce only a 2-fold titer increase in response to infection [15] . However , the analysis performed here shows that 2 fold titer changes are common even among individuals with low pre-existing titers . Antibody concentrations reach a peak 4–7 weeks after infection and then decay over a period of around six months to a plateau that is maintained for several years [21] . Although the profile of HA antibody decay is not well characterised , the probability of detecting 2- or 4- fold rises will vary with the interval following infection . However , in our data the longest interval between the peak transmission period and blood sampling was in season 3 , when the proportion of 2-fold titer rises was lowest . A second hypothesis is that 2-fold rises correspond to infection which is attenuated by mucosal or serological antibodies to homologous or heterologous strains , or by innate or cell mediated immunity . Antibody responses to inactivated influenza vaccines clearly demonstrate the potential for antigenic stimulation without active infection and the phenomenon of boosting of immunity in exposed yet uninfected individuals is well documented for other viruses ( e . g . varicella zoster [22] ) . A third hypothesis is that 2-fold rises are an artefact unrelated to influenza infection or exposure . Seasonal variation in titres independent of infection might result from the presence of non-specific inhibitors of agglutinination . For example , this could happen if the circulation of other viruses boosted the immune system , leading to small increases in all antibody titers . In such a scenario , one might expect the effect to be similar on the different subtypes . However , in 2007 , a large proportion of individuals exhibited 2-fold increases for H1N1 but not for H3N2 or B , suggesting that this hypothesis is not strongly supported by the data . It is also important to understand why 2-fold titers changes were prominent during some seasonal influenza epidemics but not during the pandemic . One possibility may be that there was greater antigenic mismatch for some seasonal strains because of unrecognised co-circulation of different influenza strains from those used as antigens in the HI assay . In this situation , anti-HA antibodies generated by infection have lower avidity for the HA of the assay virus . Conversely , original antigenic sin , where an infection results in an anamnestic response and the generation of antibodies directed towards an earlier infecting strain , might also explain 2-fold titer rises in response to infection [17] . In all these scenarios however , 2-fold increases would still represent infection by an influenza virus . It is unlikely that 2-fold increases represent cross-reactivity of HI antibodies to strains of one subtype with strains of other subtypes . This is confirmed by our analysis that did not reject the hypothesis of an absence of cross-reactivity between subtypes . It is therefore important for future work to determine if 2-fold titer increases represent infection , antigenic stimulation ( attenuated infection ) , or artefact . If influenza infection rates are higher than currently recognised this might change our understanding of influenza transmission and of intra-host and inter-host immune mediated evolutionary pressures , and may have implications for the feasibility of control measures . In the dataset examined here , 2-fold increases exceeded 4-fold increases for H1N1 in 2008 and H1N1 , H3N2 and B in Spring 2009 . There was no clear pattern with respect to subtype or strain . The seasonal H1N1 strain circulating in 2008 ( A/Brisbane/59/2007 ) was antigenically distinct from those circulating previously ( A/Solomon Islands/03/2006 and A/New Caledonia/20/1999-like ) , but this strain continued to circulate in Spring 2009 . The seasonal H3N2 strain circulating in Spring 2009 ( A/Perth/16/2009 ) was antigenically distinct from the 2007/8 strain ( A/Brisbane/10/2007 ) . H3N2 A/Perth/16/2009-like viruses have been difficult to propagate and we had difficulty propagating sufficient virus for the HI assays using A/Perth/16/2009-like viruses isolated from the cohort during the Spring 2009 season . We therefore used a virus isolated from a patient in Hanoi by the National Influenza Center , and propagated in eggs followed by MDCK cells ( TX265M2E1 ) for undertaking HI testing of sera collected in Spring 2009 . It is possible that the propagation in eggs this virus underwent might have resulted in some antigenic change , resulting in lower titers in the HI assay . National influenza surveillance data indicates that both influenza B lineages - Yamagata and Victoria- co-circulated during the study period , with the Yamagata lineage dominating in 2007 and 2008 and the Victoria lineage in 2009 . For all HI assays , we used the same influenza B virus , which was isolated in 2008 and was characterized antigenically as Yamagata lineage-like , as with all influenza B viruses isolated from the cohort in 2008 . While Yamagata viruses dominated the influenza B samples we collected in 2007 and 2008 , the Victoria lineage was predominant in 2009 . This may be a factor explaining the lower influenza B titer increases seen in that year . If heterogeneities in the proportion of 2-fold titer rises are largely attributable to a poor match between assay antigen and infecting virus , future seroprevalence and seroincidence surveys will need to use a greater diversity of antigens than typically used currently . There are often strong age-related patterns in influenza serology . Ideally , we would therefore like to fit our statistical model independently for each age group . However , simulation studies indicate that the relatively small number of observations per age group would lead to relatively inaccurate estimates . We have therefore opted for an intermediate estimation strategy . Our statistical model fits a single distribution of true paired serology to all subjects; but since we infer true paired serology for each individual , we can reconstruct a posteriori the distribution of true paired serology for the different age groups . Even with such a conservative approach ( i . e . it favours scenarios where the different age groups exhibit similar distributions ) , we were able to detect clear age-related patterns . In particular , it indicated that age may be another factor that influences the occurrence of a 2-fold rise . Larger sample sizes will be needed to investigate this possibility further . The presence of relatively large proportions of individuals experiencing a 2-fold increase in antibody titers is not a peculiarity of the Vietnamese data examined here . Similar shifts were observed on data gathered by Cowling et al , with micro-neutralization assays for 2009 H1N1pdm09 influenza and on HI assays for seasonal influenza [23] ( Figure S6 ) . It is well known that there may be substantial within- and between- laboratory variability in HI assays as well as in other serological assays such as virus neutralisation ( VN ) [24] . The level of intra-laboratory variations may depend on both the laboratory and the type of assay used [24] . Here , we have introduced an approach that allows controlling for within-laboratory variations . The only additional data needed compared with standard serological surveys is that replicate measurements are performed for a subset of subjects . These replicate measurements allow within-laboratory quantification of variation in assay performance . With this information , it is then possible to reconstruct the distribution of paired serology that is corrected for the estimated level of within-laboratory variations . Although our approach gives a better control on within-laboratory variation , it does not address the problem of between-laboratory variation . The use of standards in bioassays is critical for minimising the impact of the latter problem [24] . To conclude , while a 4-fold titer increase may be a highly specific diagnostic of infection by an influenza virus for individual cases , this criterion is less justifiable when the objective is to estimate community ARs . Our work shows that requiring a 4-fold titer increase may lead to ARs being substantially underestimated . More research is needed to determine what proportion of 2-fold rises are causally linked to exposure to influenza , and what proportion may be caused by other mechanisms . It will be important to determine whether the high proportion of 2-fold titer increases seen in the settings of Vietnam and Hong Kong [23] are also observed in other ( e . g . temperate climate ) settings .
Samples were collected from a household-based cohort of 940 participants in 270 households in a single community in semi-rural northern Vietnam as previously described [5] . None of the participants had ever received influenza immunisation . Participants were under weekly active surveillance by village health workers for influenza-like-illness ( ILI ) and in the event of an ILI were asked to provide a nose and throat swab for detection of influenza RNA by reverse-transcription polymerase chain reaction . Participants were also asked to provide serial blood samples at times when national influenza surveillance data indicated that influenza circulation was minimal . The samples described here were collected over a period of three consecutive influenza seasons , from December 2007 through April 2010 . The bleeding times were 1st–7th December 2007 ( bleed 1 ) , 9th–15th December 2008 ( bleed 2 ) , 2nd–4th June 2009 ( bleed 3 ) , and on the 3rd April 2010 ( bleed 4 ) . This provided three sets of paired samples either side of an influenza transmission season: 548 paired samples for season 1 ( 2008 ) , 501 paired samples for season 2 ( Spring 2009 ) , and 540 paired samples for season 3 ( Autumn 2009 ) . In season 1 , the influenza A virus strains detected in the cohort through ILI surveillance were A/H1N1/Brisbane/59/2007-like and A/H3N2/Brisbane/10/2007-like; in season 2 , they were A/H1N1/Brisbane/59/2007-like and A/H3N2/Perth/16/2009-like; and in season 3 , it was A/H1N1/California/7/2009-like . There was co-circulation of influenza B Yamagata lineage and Victoria lineage in both season 1 and season 2 , with a predominance of Yamagata lineage in season 1 and Victoria lineage in season 2 . Nasal and oropharangeal swabs were assessed by real-time reverse-transcriptase polymerase chain reaction ( RT-PCR ) , according to WHO/USCDC protocols [25] . Influenza hemagglutination inhibition ( HI ) assays were performed according to standard protocols [WHO 2011 manual] . The seasonal influenza A viruses used were isolated from participants' swabs or from swabs taken from patients presenting in Ha Noi in the same season and propagated in embryonated hen's eggs or in MDCK cells . A reference antigen supplied by WHO ( A/H1N1/California/7/2009-like ) was used to assess season 3/pandemic sera . A single influenza B virus isolated from a participant during 2008 was used to assess serum for both the first and second seasons . The virus had a titer of 320 with B/Wisconsin/1/2010 ( Yamagata ) reference antisera and of <10 with B/Brisbane/60/2008 ( Victoria ) antisera . Each virus was first assessed for haemagglutination of erythrocytes from chickens , guinea pigs and turkeys then titrated with optimal erythrocytes . Serum was treated with receptor destroying enzyme ( Denka Seiken , Japan ) then heat inactivated and adsorbed against packed erythrocytes . Eight 2-fold dilutions of serum were made starting from 1∶10 and incubated with 4 HA units/25 µl of virus . Appropriate erythrocytes were added and plates read when control cells had settled . Virus , serum and positive controls were included in each assay . Pre- and post-season sera were tested in pairs . Each serum was tested in a single dilution series . The HI titre was read as the reciprocal of the highest serum dilution causing complete inhibition of RBC agglutination , partial agglutination was not scored as inhibition of agglutination . If there was no inhibition of HI at the highest serum concentration ( 1∶10 dilution ) the titer was designated as 5 . Only one sample had a titer >1280 and this was not adjusted . Replicate HI assay measurements were performed on a subset of samples from patients that seroconverted ( i . e . 4-fold rise in titer ) as well as some others that had titers ≥20 in both pre and post-season sera . A less technical description of statistical methods is given for non-specialists in Box 1 and Figure 7 .
The research was approved by the institutional review board of the National Institute of Hygiene and Epidemiology , Vietnam; the Oxford Tropical Research Ethics Committee , University of Oxford , UK; and the Ethics Committee of the London School of Hygiene and Tropical Medicine , UK . All participants provided written informed consent . | Each year , seasonal influenza is responsible for about three to five million severe illnesses and about 250 , 000 to 500 , 000 deaths worldwide . In order to assess the burden of disease and guide control policies , it is important to quantify the proportion of people infected by an influenza virus each year . Since infection usually leaves a “signature” in the blood of infected individuals ( namely a rise in antibodies ) , a standard protocol consists in collecting blood samples in a cohort of subjects and determining the proportion of those who experienced such rise . However , because of inherent measurement errors , only large rises are accounted for in the standard 4-fold rise case definition . Here , we revisit this 70 year old and widely accepted and applied criterion . We present innovative statistical techniques to better capture the impact of measurement errors and improve our interpretation of the data . Our analysis suggests that the number of people infected by an influenza virus each year might be substantially larger than previously thought , with important implications for our understanding of the transmission and evolution of influenza – and the nature of infection . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods",
"Ethics",
"statement"
] | [
"medicine",
"public",
"health",
"and",
"epidemiology",
"epidemiology",
"infectious",
"disease",
"epidemiology"
] | 2012 | Influenza Infection Rates, Measurement Errors and the Interpretation of Paired Serology |
Apoptosis is an important mechanism by which virus-infected cells are eliminated from the host . Accordingly , many viruses have evolved strategies to prevent or delay apoptosis in order to provide a window of opportunity in which virus replication , assembly and egress can take place . Interfering with apoptosis may also be important for establishment and/or maintenance of persistent infections . Whereas large DNA viruses have the luxury of encoding accessory proteins whose primary function is to undermine programmed cell death pathways , it is generally thought that most RNA viruses do not encode these types of proteins . Here we report that the multifunctional capsid protein of Rubella virus is a potent inhibitor of apoptosis . The main mechanism of action was specific for Bax as capsid bound Bax and prevented Bax-induced apoptosis but did not bind Bak nor inhibit Bak-induced apoptosis . Intriguingly , interaction with capsid protein resulted in activation of Bax in the absence of apoptotic stimuli , however , release of cytochrome c from mitochondria and concomitant activation of caspase 3 did not occur . Accordingly , we propose that binding of capsid to Bax induces the formation of hetero-oligomers that are incompetent for pore formation . Importantly , data from reverse genetic studies are consistent with a scenario in which the anti-apoptotic activity of capsid protein is important for virus replication . If so , this would be among the first demonstrations showing that blocking apoptosis is important for replication of an RNA virus . Finally , it is tempting to speculate that other slowly replicating RNA viruses employ similar mechanisms to avoid killing infected cells .
Rubella virus ( RV ) is an enveloped positive strand RNA virus in the family Togaviridae and is the sole member of the genus Rubivirus ( reviewed in [1] ) . Humans are the only natural host for RV and in most cases the virus causes a systemic infection the symptoms of which include maculopapular rash , lymphadenopathy , low-grade fever , conjunctivitis and sore throat . However , RV infections can be complicated by the appearance of acute or chronic arthralgia , arthritis , thrombocytopenia and encephalopathy . In utero infection during the first trimester of pregnancy often results in a characteristic series of birth defects known as congenital Rubella syndrome . Worldwide , RV is thought to cause more birth defects that any other infectious agent yet , very little is known about molecular aspects of viral pathogenesis . A number of studies suggest that viral persistence may underlie some of the most serious aspects of infection including congenital Rubella syndrome and arthritis [2] , [3] , [4] , [5] , [6] . Among the togavirus family , RV is unique in that its replication is associated with mitochondria . The link between RV infection and this organelle first became apparent when analysis of purified virions revealed that cardiolipin , a phospholipid that is only found in mitochondria , is a significant component of the RV envelope [7] . Subsequently , it was discovered that RV infected cells exhibit striking mitochondrial defects . Virus infection induces clustering of mitochondria in the perinuclear region as well as formation of electron-dense plaques between apposing mitochondrial cisternae: structures that have been termed confronting membranes [8] , [9] . The function of these structures is not known but expression of capsid protein in the absence of other RV proteins is sufficient to induce their formation [10] . A large pool of the capsid protein localizes to the surface of mitochondria [11] and the inter-mitochondrial plaques [12] but given that assembly of RV virions occurs primarily on Golgi membranes , the targeting of the capsid to this organelle likely reflects a nonstructural function of this protein . The studies described above underscore the close link between the capsid protein and mitochondria in RV biology and form the basis for our central hypothesis; that association of the RV capsid protein with mitochondria is important for virus replication . All viruses must contend with host cell anti-viral mechanisms and large DNA viruses have the luxury of harboring in many cases , multiple genes devoted to thwarting host cell defenses ( reviewed in [13] ) . In contrast , simple RNA viruses express a very limited number of proteins , most of which are directly involved in replication and virus assembly . Accordingly , it is beneficial if not essential that these viral proteins have multiple functions . It is well documented that togavirus infection often results in apoptotic death of mammalian cells ( reviewed in [14] , [15] ) and to our knowledge , there are no published studies showing that members of this virus family inhibit programmed cell death pathways . With the exception of RV , togavirus replication and virus egress from vertebrate cells occurs within 4–6 hours followed by extensive death by apoptosis within 24 hours . Accordingly , for most togaviruses , preventing apoptosis is likely not required in order for efficient replication to occur . In contrast , the replication cycle for RV is unusually slow; the eclipse period is at least 12 hours and viral titers peak virion secretion occurs between 48–72 ( reviewed in [16] ) . RV-induced apoptosis in mammalian cells has been reported but generally , extensive cytopathic effect is not a hallmark of RV infection . When apoptosis does occur , it is not until 5–7 days post-infection that maximum levels are reached [17] and this is well past the peak virus production phase . In the present study , we report that the capsid protein blocks apoptosis in RV infected cells most likely to allow sufficient time for virus replication . This process occurs at the level of mitochondria through a Bax-dependent pathway .
We reasoned that in order for RV replication and virion secretion to increase through 48 hours and beyond , programmed cell death must be inhibited during this period . Accordingly , we compared the levels of apoptosis in RV and mock-infected A549 cells by indirect immunofluorescence using an antibody specific for activated caspase 3 . Interestingly , less than 5% of infected cells exhibited signs of apoptosis 48 hours post-infection ( Figure 1A and B ) . Moreover , when challenged with the kinase inhibitor staurosporine , a potent inducer of apoptosis [18] , RV infected cells were significantly more resistant to apoptosis than mock infected cells . Specifically , the percentage of caspase 3-positive cells was almost three fold lower in the infected samples . This was not due to detachment of infected cells as data in Figure 1C show that treatment with staurosporine did not cause significant loss of infected cells . Finally , Figure 1D shows that even after 72 hours , RV infection does not significantly affect the percentage of viable A549 cells . Together , these data indicate that RV-infected A549 cells are resistant to programmed cell death . Next , we sought to determine which viral protein ( s ) was primarily responsible for protecting infected cells against apoptosis . Previous studies have indicated that expression of the nonstructural proteins p150 and p90 are cytotoxic [19] , [20] and therefore , we focused our attention on the virus structural proteins . Plasmids encoding glycoproteins E2 and E1 or capsid , were transiently tranfected into A549 cells and at 40 hours post-transfection , cells were induced to undergo apoptosis by treatment with anti-Fas . Samples were processed for indirect immunofluorescence ( Figure 2A ) and the numbers of active caspase 3-positive transfectants were determined . Data in Figure 2B show that the levels of apoptosis were similar in cells expressing the viral glycoproteins E2 and E1 and the negative control protein eGFP . In contrast , expression of the RV capsid protein was just as protective against anti-Fas as the well-characterized anti-apoptotic protein Bcl-XL [21] . Compared to eGFP or E2E1 transfectants that were treated with anti-Fas , the percentage of apoptotic cells among capsid transfectants was three fold lower ( Figure 2B ) . We also examined whether RV infection and/or capsid expression protects Vero cells from apoptosis . This cell line is used extensively to study RV replication and similar to what was observed with A549 , infection of Vero cells with RV , or transient expression of capsid protein conferred protection from staurosporine-induced apoptosis ( Figure S1A , B , arrowheads ) . Because Vero cells do not respond to anti-Fas treatment , it was not possible to determine the effect of capsid expression on death receptor pathways . These data appear to be at odds with a previous study which reported that the RV capsid was pro-apoptotic in RK-13 cells [22]; a cell line that is exquisitely sensitive to RV-induced apoptosis [23] . Accordingly , we assayed intrinsic and extrinsic apoptotic pathways by staurosporine and anti-Fas treatment of RK-13 cells at 48 and 72 hours post-transfection . In both cases , expression of the capsid protein conferred resistance to apoptosis similar to Bcl-XL ( Figure S2 ) . Together , these data indicate that the RV capsid is an anti-apoptotic protein that protects cells from multiple apoptotic stimuli . We next endeavored to identify what step in apoptotic signaling was blocked by capsid protein . For these experiments , lentiviral transduction was used to create A549 cells that stably express capsid protein under the control of a doxycycline-regulated promoter . Results from indirect immunofluorescence showed that less than 50% of the polyclonal population of transduced cells expressed RV capsid following doxycyline treatment ( Figure 3A ) . Similar to results shown in Figure 2 , induction of capsid expression protected the stably transduced A549 cells against staurosporine- and Fas-mediated activation of caspase 3 ( Figure S3 ) . To further confirm that apoptotic stimuli do not activate caspases in these cells , we measured the appearance of the downstream caspase 3 substrate , cleaved Poly ( ADP-ribose ) polymerase ( PARP ) . Figure 3B shows that expression of capsid protein results in decreased anti-Fas-induced cleavage of PARP compared to luciferase-expressing cells . These data indicate that capsid protects A549 cells from staurosporine and anti-Fas treatment by blocking caspase activation . We next determined where upstream of caspase 3 activation , that capsid protein acted . Both staurosporine- and anti-Fas- can trigger apoptosis through the mitochondrial pathway , so we tested the ability of capsid protein to block depolarization of mitochondrial membranes in response to apoptotic stimuli . Doxycycline-treated A549 cells expressing capsid protein or luciferase were challenged with staurosporine or anti-Fas and then stained with the membrane potential sensitive dye TMRM . Samples were analyzed by FACS , after which the relative specific cell death levels for each sample were calculated ( Figure 3C ) . Data in Figure 3D show that compared to luciferase , expression of capsid protein reduced the relative specific death induced through intrinsic ( staurosporine ) or death receptor-dependent pathways ( Fas ) by 20–35% . However , because less than 50% of the lentivirus-transduced cells express detectable levels of capsid protein , these numbers likely underestimate the true level of protection afforded by stable expression of the RV capsid protein . Bax and Bak are two key apoptotic molecules that form oligomers on mitochondria [24] , [25] , [26] and apoptosis occurs when the mitochondrial outer membrane is permeabilized by these pore-forming molecules [27] . Accordingly , we next focused our efforts on these Bcl-2 family members starting with Bax . Normally , Bax is an inactive monomer found in the cytosol or loosely bound to the mitochondrial outer membrane of healthy cells [28] , [29] . In response to apoptotic stimuli , Bax activation is characterized by a multi-step process whereby it undergoes a conformational change [30] , [31] , integrates into the mitochondrial membrane [28] , [32] where it forms higher order oligomers [33] . It is the large Bax oligomers that are linked to the formation of membrane pores that facilitate release of mitochondrial cytochrome c and downstream caspase activation [33] , [34] . Of these multiple steps , Bax conformational change can be detected by immunoreactivity with a conformation-specific antibody , 6A7 [35] , [36] . We observed that RV infection induces Bax conformational change , however cytochrome c remained associated with mitochondria ( Figure 4A , arrows ) . Moreover , Bax conformational change as detected by 6A7 staining was evident in the majority ( 76% ) of cells expressing capsid protein ( Figure 4B arrows ) . In contrast , among cells expressing the viral glycoproteins E2 and E1 , only 6% contained activated Bax . Despite initial stimulation of Bax , similar to infected cells , no loss of cytochrome c from mitochondria was observed in capsid-expressing cells . Because capsid protein stimulates Bax in a manner that does not produce functional pores that mediate efflux of cytochrome c , we initially thought that capsid protein blocks oligomerization of Bax . However , data in Figure 5A indicate that this is not the case . Rather , our results suggest that capsid protein and Bax form mixed large hetero-oligomers even in the absence of apoptotic stimuli . Indeed , reciprocal co-immunoprecipitation experiments confirmed that capsid forms a stable complex with Bax ( Figure 5B ) . Staurosporine treatment enhanced the formation of the capsid:Bax hetero-oligomers but evidently did not facilitate the assembly of functional Bax pores as the cells were not apoptotic . Interestingly , we found no evidence that capsid protein binds to Bak ( Figure 5C ) suggesting the interaction of this viral protein with Bcl-2 family proteins is highly specific . Together , these data suggest that capsid protein and Bax form mixed oligomers that do not function as pores . Since capsid protein forms a complex with Bax , we next tested whether its expression could inhibit Bax-mediated apoptosis . Over-expression of either Bax or Bak induces cell death in the absence of other apoptotic stimuli [37] . A549 cells were co-transfected with plasmids encoding GFP-Bax and capsid , Bcl-XL ( positive control ) or vector alone ( negative control ) and at 24 hour post-transfection , samples were stained with the membrane-potential specific dye TMRM and then subjected to flow cytometric analyses ( Figure 6A ) . As a second control , we transfected cells with a plasmid encoding a capsid deletion construct ( CapNT ) that is not targeted to mitochondria ( see below ) . Loss of TMRM staining as a result of depolarization of mitochondrial membranes was used as the measure of apoptotic cell death . Quantitation of the data ( Figure 6C ) revealed that expression of capsid protein reduced the level of Bax-induced cell death by more than 60% compared to CapNT or vector alone . Similar results were observed for cells expressing Bcl-XL , a protein which has previously been shown to block the effects of Bax over-expression [38] . Data in Figure S4 show that capsid expression also protects primary human embryonic fibroblast ( HEL-18 ) cells [17] from Bax-mediated apoptosis . The anti-apoptotic activity of capsid protein was specific to Bax as evidenced the fact that it did not attenuate Bak-mediated apoptosis ( Figure 6B , C ) . To further understand how capsid functions to block apoptosis , we determined whether expression of this viral protein inhibits Bax-induced release of cytochrome c . A549 cells were co-transfected with plasmids encoding GFP-Bax and capsid or empty vector . Localization of cytochrome c was monitored by fluorescence microscopy at 24 hours post-transfection . As expected , in cells expressing GFP-Bax and vector alone , there was marked loss of cytochrome c from mitochondria ( Figure 7A , asterisks ) . In contrast , in cells that expressed both capsid protein and GFP-Bax , cytochrome c remained associated with this organelle ( Figure 7A , arrows ) . However , consistent with data shown in Figures 5 and 6 , capsid did not block GFP-Bak-induced loss of cytochrome c from mitochondria ( Figure 7B arrows ) . Based on the assumption that association of capsid protein with mitochondria is critical for its anti-apoptotic function , we next mapped the region of capsid protein that is required for targeting to this organelle . Analyses of the RV capsid protein sequence with web-based algorithms such as PSORT II Prediction ( http://psort . nibb . ac . jp/form2 . html ) indicated that conventional mitochondrial targeting signals are absent . We therefore constructed a series of capsid deletion mutants whose localizations were determined by expression in A549 cells ( Figure 8A ) . From the indirect immunofluorescence data shown in Figure 8B , it can be seen that the 23 amino acid residue E2 signal peptide which forms the hydrophobic carboxyl-terminus of capsid protein , is required for association with mitochondria . Moreover , the observation that a pool of CapCT overlaps with cytochrome c indicates that the carboxyl-terminal region of capsid protein contains information that is sufficient for targeting to mitochondria . Intriguingly , expression of the CapCT construct caused extreme compaction of the mitochondrial network to the perinuclear region , much more so than in cells expressing full-length capsid protein . We next determined whether association of capsid with mitochondria correlated with its ability to block apoptosis . Transfected cells expressing the various capsid constructs were challenged with staurosporine or anti-Fas , and then apoptosis induction was assessed using the activated caspase 3 assay . The amino-terminal capsid construct ( CapNT ) neither associates with mitochondria nor protects against apoptosis ( Figures 8 , 9 ) . Conversely , CapCT , a pool of which is targeted to mitochondria , protects as well as full-length capsid protein against staurosporine and anti-Fas challenge . CapΔRSP , which lacks the hydrophobic E2 signal peptide and a membrane proximal arginine-rich ( R ) motif , is not targeted to mitochondria and does not block staurosporine or anti-Fas-mediated induced activation of caspase 3 . Interestingly , although CapΔSP does not localize to mitochondria , it did confer resistance to both Fas- and staurosporine-induced apoptosis ( Figure 9 ) . This observation suggests that the membrane-proximal R motif is important for the anti-apoptotic function of capsid . Table 1 summarizes the localization and anti-apoptotic properties of the capsid deletion mutants . To investigate if the arginine residues in the membrane-proximal R motif were important for the anti-apoptitic function of capsid protein , we created a point mutant ( CapCR5A ) in which five arginines in this motif were changed to alanine residues ( Figure 10A ) . This capsid mutant was targeted to mitochondria where it activated Bax and stimulated cytochrome c release in the absence of apoptotic stimuli ( Figure 10B and C asterisks ) ; indicating that the arginine residues within the R domain are critical for the anti-apoptotic activity of capsid protein . Moreover , it would appear that mutation of these arginine residues unmasks an intrinsic pro-apoptotic activity of capsid protein , which may explain why it alone can stimulate Bax conformational change and membrane insertion . Next , we compared the Bax-binding ability of the CR5A mutant relative to wild type capsid and capsid deletion constructs . The observation that more CapΔSP is recovered in anti-Bax coimmunoprecipitations than CapΔRSP ( Figure 11A ) suggests that the R domain is important for interaction with Bax . However , ablation of the arginine residues in the R domain did not affect binding to Bax indicating that the arginine residues per se in this motif are not essential for interaction with Bax ( Figure 11A ) . Binding between Bax and CapCT or CapNT was not detected in our assays ( Figure 11B ) . Indirect immunofluorescence analyses revealed that unlike wild type capsid and CapCR5A , neither CapNT , CapCT , CapΔSP nor CapΔRSP induced the 6A7-specific conformation change in Bax ( data not shown ) . Together , these results suggest that capsid protein employs a multi-step mechanism to block apoptosis . Specifically , binding to Bax through the R domain and/or the carboxyl terminus stimulates a conformational change in Bax; but pore formation and/or functionality is blocked by the arginines in the R motif of capsid protein . We introduced the CR5A mutations into the capsid gene of a RV infectious clone in order to determine if the membrane-proximal arginine-rich ( R ) domain in capsid protein is required for blocking apoptosis during infection . Our hypothesis was that early onset apoptosis would result in decreased replication and virus particle production . A549 cells were infected with wild type or CR5A strains of RV and virus replication and apoptosis induction were analyzed . Data in Figure 12A show that cells infected with CR5A virus were significantly more susceptible to Fas-dependent apoptosis . Moreover , in non-treated ( control ) samples , the level of virus-induced apoptosis was four fold higher in cells infected with the CR5A mutant . Similar results were obtained with infected Vero cells ( data not shown ) . Next , we compared the levels of RV proteins in CR5A and wild type ( WT ) RV infected cells as a function of time . Figure 12B shows that in cells that were infected with WT RV , the level of virus nonstructural ( p150 ) and structural proteins ( capsid ) peaked at 72 hours . In contrast to p150 levels which were only moderately lower , steady state levels of capsid protein were dramatically lower in CR5A infected cells at all time points . To control for the possibility that CR5A capsid was unstable in the infected cells , we also determined the relative levels of another structural protein , E1 . Similar to capsid protein levels in CR5A infected cells , levels of E1 were much lower than in WT virus infected cells; suggesting a defect in synthesis of structural proteins in CR5A infected cells . Consistent with this theory , data in Figure 12C show that secretion of CR5A virions is severly impaired . This was not because the CR5A capsid is misfolded as data in Figure S5 show that this mutant capsid protein functions as well as wild type capsid in driving assembly and secretion of Rubella virus-like particles . Nonstructural proteins are translated directly from the 40S genomic RNA whereas capsid and other structural proteins are made from a subgenomic RNA . Accordingly , it is possible that virus transcription and replication are impaired in the CR5A mutant . Quantitative RT-PCR with p90 specific primers was used to determine the relative levels of genomic RNA in the WT and CR5A infected samples ( Figure 12D ) . From these data , it can be seen that replication of viral RNA was severely affected in CR5A infected cells . This was not due to decreased infection efficiency because at six hours post-infection , there was on average >50% more genomic RNA in CR5A infected cells ( Table 2 ) . Moreover , as demonstrated by plaque assays , cells infected with CR5A virus did release infectious virus ( Figure 12E ) . Interestingly , the CR5A plaques were larger and had a spotty appearance compared to wild type virus-produced plaques which were smaller and clearer . Although data in Figure S5 indicate that Cap5RA is not misfolded , without additional investigation , we could not completely rule out the possibility that the replication defects associated with the CR5A strain virus were due to other inherent defects of the mutant capsid protein . Therefore , we attempted to artificially block apoptosis by over-expression of Bcl-XL or adding the caspase inhibitor Z-VAD-FMK to CR5A infected cells . Over-expression of Bcl-XL did not rescue the CR5A replication but this result was non-informative as further investigation revealed that this anti-apoptotic protein was unable to protect mitochondria from the effects of CapC5RA in transfected cells ( data not shown ) . In contrast , addition of Z-VAD-FMK did have a modest effect on production of viral proteins in CR5A infected cells ( Figure 13A ) . The effect was most pronounced at 72 hrs post-infection where levels of p150 and capsid were considerably higher in Z-VAD-FMK treated cells . In contrast , blocking caspase activity in cells that were infected with wild type RV did not appreciably alter the steady state levels of viral proteins . Finally , it can be seen from the data in Figure 13B that Z-VAD-FMK treatment had a modest effect on production of CR5A virus . Compared to CR5A-infected Vero cells treated with DMSO alone , addition of Z-VAD-FMK resulted in a modest increase in viral titers as evidenced by increased clearing of RK-13 monolayers . Together , these data are consistent with our hypothesis that the anti-apoptotic function of capsid is important for virus replication .
Apoptosis is a common defense mechanism used by host cells to limit the spread of viral infections and consequently , a number of viruses have developed mechanisms to disrupt programmed cell death pathways . With few exceptions , all known viral apoptosis inhibitors are accessory proteins that are encoded by DNA viruses and therefore , a great deal of effort has focused on these proteins ( reviewed in [39] ) . Ironically , even though RNA viruses cause the vast majority of viral diseases in humans , comparatively little is known about if or how these types of viruses interfere with apoptotic signaling . Among the exceptions are picornaviruses , a number of which encode “security” proteins ( leader protein and 2BC ) that can block apoptosis [40] , [41] . The mechanisms by which these proteins block apoptosis are not known and interestingly , caspase activation can still occur normally . These proteins may not actually prevent cell death per se , but rather , shift the balance toward necrotic cell death as opposed to apoptosis . Hepatitis C virus ( HCV ) is known to modulate apoptotic signaling but unlike picornaviruses , this virus is not cytolytic and readily establishes persistent infections in vivo . HCV encodes a number of proteins that reportedly exhibit anti-apoptotic activity . For example , the nonstructural proteins NS2 and NS5A interfere with programmed cell death by different mechanisms [42] , [43] . The functions of HCV structural proteins in apoptotic signaling events are less clear; in particular , the core/capsid protein . The majority of data suggest that this protein acts to induce apoptosis although a number of published studies suggest otherwise ( reviewed in [44] ) . Similarly , with one exception [45] , expression of HCV E2 glycoprotein reportedly acts in a pro-apoptotic manner [46] , [47] , [48] . By and large , these studies involved plasmid-based expression of individual HCV proteins and indeed the data provide much to ponder with respect how this virus interfaces with apoptotic pathways . However , it is still not clear how individual HCV proteins or those of any other RNA virus affect cell death during infection . Multiple laboratories have reported that RV infection induces programmed cell death in a variety of cultured cell lines [17] , [19] , [23] , [49] , [50] but it is worth noting that in virtually all cases , maximum synthesis of viral macromolecules and release of virions occur well before extensive apoptosis is observed . For example , in Vero cells , robust expression of structural proteins is first detected at 16 hours post-infection and secretion of infectious virions peaks 32 hours later [51] . Conversely , late apoptotic events such as DNA fragmentation and expression of pro-apoptotic proteins p53 and p21 does not peak until 5–7 days post-infection [17] . This indicates that that the majority of programmed cell death occurs long after the peak of virus production . Consistent with these observations , we show that RV infected cells are in fact , resistant to apoptosis for at least 48 hours post-infection . Here , we provide evidence that in addition to functioning in virus assembly , the RV capsid protein is a potent inhibitor of apoptosis . With the possible exception of HCV capsid and E2 , structural proteins of RNA viruses have been found to cause apoptosis rather than prevent cell death ( reviewed in [39] , [44] , [52] ) . As far as we are aware , this is the first example of a structural protein from an RNA virus that functions to block cell death pathways through interactions with Bax . Mapping studies suggest that expression of the virus nonstructural proteins is the cause of RV-induced cell death [19] , [20] . Accordingly , counteracting apoptotic pathways that become activated by expression of these early proteins may be essential for efficient replication; a theory that is supported by data from reverse genetic experiments with the CR5A mutant . Our data appear to be in disagreement with previously published data showing that capsid protein expression induces apoptosis [22] . An important distinction between the previous study and the present work is that we assayed the ability of capsid protein to protect against various apoptotic stimuli rather than assaying whether or not capsid expression induces apoptosis in the absence of stimuli . In addition , we found that capsid protein blocks apoptosis in multiple cell lines ( including a primary human cell line ) whereas in the afore-mentioned study , capsid protein was reported to be pro-apoptotic in RK-13 but not other cell lines . Data in the present study are also consistent with the fact that stable cell lines that express high levels of RV structural proteins ( including capsid protein ) are readily established in a variety of cells types [19] , [53] , [54] . Importantly , the results of the reverse genetic experiments suggest that the anti-apoptotic function of the capsid protein is critical for RV replication . Although we cannot absolutely rule out the possibility that mutations in the R domain of capsid directly affect its functions in replication and assembly this seems unlikely . First , the CapCR5A mutant was able to drive particle assembly and CR5A virions efficiently delivered viral RNA to host cells . Second , the region of capsid that complements p150 function in virus replication is in the amino-terminal one third of the protein [55] . Accordingly , the most logical conclusion is that the anti-apoptotic role of capsid protein is necessary to promote survival of the host cell during the long replication cycle . To our knowledge , this has never been demonstrated before for an RNA virus but it is tempting to speculate that other slowly replicating RNA viruses employ similar mechanisms to avoid killing infected cells . Although capsid protein may interfere with apoptosis by more than one mechanism , because the Bax-dependent pathway is a critical feature of mitochondrial apoptosis in most human cell types , interfering with the pore-forming ability of this protein is likely the key anti-apoptotic function of capsid protein . Binding of capsid protein to Bax induces a major conformational change , which interestingly , seems to promote activation and oligomerization of Bax . It is not clear if this phenomenon is related to the anti-apoptotic activity of capsid or if it is an inconsequential effect of complex formation with Bax . Figure 14 depicts a model in which capsid protein interferes with formation of functional Bax pores . In some critical aspects , the RV capsid protein may function analogously to the cytomegalovirus accessory protein vMIA , a putative Bcl-2 homolog that forms mixed oligomers with Bax [56] . However , confirmation of this theory is dependent upon determining the structure of the RV capsid protein . Mapping studies localized the anti-apoptotic activity to the carboxyl-terminal region of capsid protein . The E2 signal peptide , which is required for membrane association of capsid protein [57] , [58] , is also essential for targeting to mitochondria but not for blocking apoptosis . Conversely , while the membrane proximal arginine-rich ( R ) motif in capsid is dispensable for targeting to mitochondria , it is required for protection from intrinsic and extrinsic apoptotic stimuli ( Table 1 ) . The R motif ( RSARHPWRIR ) of RV capsid bears remarkable similarity to the Bax-binding motif ( RRHRFLWQRR ) in vMIA [59] which is critical for blocking Bax- but not Bak-dependent apoptosis [60] . Despite the apparent similarities between vMIA and RV capsid protein , one difference worth noting is that the arginine-rich motif in capsid is not required for binding to Bax . As mentioned above , it is possible that capsid protein blocks apoptosis through other mechanisms , at least one of which does not involve Bax . For example , CapCT , which neither binds to nor activates Bax , protects cells from staurosporine and Fas-dependent apoptosis . However , unlike full-length capsid , CapCT does not protect cells from Bax over-expression . Capsid binds two other pro-apoptotic proteins p32 and Par-4 [61] and through sequestration into non-functional complexes , it is possible that the functions of these proteins in apoptotic signaling are mitigated . Although we have no direct evidence to support this theory , binding to membrane-associated capsid protein may prevent Par-4 from engaging in pro-apoptotic complexes in the nucleus or cell surface [62] , [63] , [64] . Finally , we showed that translocation of the capsid-binding pro-apoptotic protein p32 into mitochondria is inhibited by RV infection [12] . Because targeting of p32 to mitochondria is critical for its function in programmed cell death pathways [65] , [66] , reducing the levels of mitochondrial p32 would be expected to reduce apoptosis . To summarize , we describe a novel mechanism by which a viral capsid protein potently blocks apoptosis . Our data suggest that this function of capsid is important for virus replication and it is also tempting to speculate that establishing and/or maintaining persistent infections in vivo also requires this activity . RV is known to persistently infect lymphocytes and capsid-dependent inhibition of Fas-dependent apoptosis may allow the virus to disseminate through the body using apoptosis-resistant lymphocytes as conduits . It will be of interest to examine proteins from other slowly replicating RNA viruses to determine if capsids or other proteins double as inhibitors of apoptosis .
The following reagents were purchased from the respective suppliers: Protein A and G Sepharose from GE Healthcare Bio-Sciences Corp ( Princeton , NJ ) ; General lab chemicals from Sigma Chemical Co . ( St . Louis , MO ) ; Media and fetal bovine serum ( FBS ) for cell culture from Life Technologies-Invitrogen , Inc . ( Carlsbad , CA ) ; A549 , HEK 293T , Vero , and RK-13 cells from the American Type Culture Collection ( Manassas , VA . ) . Hel-18 cells [17] were obtained from Dr . Eva Gonczol , ( National Center for Epidemiology , Budapest , Hungary ) . A549 and HEK 293T cells were cultured in Dulbecco's minimal essential medium ( high glucose ) containing 10% FBS , 2 mM glutamine , 1 mM HEPES , and antibiotics . Vero cells were cultured in Dulbecco's minimal essential medium ( high glucose ) containing 5% FBS , 2 mM glutamine , 1 mM HEPES , and antibiotics . RK-13 cells were cultured in minimum essential medium containing 10% FBS , 2 mM glutamine , 1 mM HEPES , 0 . 1 mM non-essential amino acids , and antibiotics . Hel-18 cells were cultured in RPMI-1640 medium containing 10% FBS , 2 mM glutamine , 1 mM HEPES , 0 . 1 mM non-essential amino acids , and antibiotics . Cells were incubated at 37°C in a humidified atmosphere with 5% CO2 . RV stocks were diluted with cell culture media and then added to cells that had been washed with PBS . Cells were incubated with the virus ( 1 ml/35 mm dish ) for 4 hours at 35°C after which time the inoculum was replaced with normal growth media . Infected cultures were kept at 35°C until experimental analyses . To investigate the effect of blocking apoptosis on virus replication , Vero cells were infected with M33 ( wild type ) or CR5A strains of RV ( MOI: 1 ) in the presence of 50 µM pan-caspase inhibitor Z-VAD-FMK ( Promega , Madison , WI ) which was initially made as a 20 mM stock solution in dimethyl sulfoxide ( DMSO ) . Z-VAD-FMK or control vehicle ( DMSO ) was added to infected cells every 24 hrs . Samples were processed at the indicated time points and virus titers were determined by plaque assay [67] . Plasmids for expression of pCMV5-Capsid , pCMV5-CapCT , pCMV5-CapΔSP and pCMV5-E2E1 have been described previously [58] , [67] , [68] . An expression plasmid encoding amino acid residues 1–152 of capsid ( CapsidNT ) was constructed by polymerase chain reaction ( PCR ) using a forward primer with a EcoRI site and a Kozak consensus ribosome-binding site ( 5′-CGGAATTCGCCACCATGGCTTCCACTACCCCCATCACC-3′′ ) and a reverse primer with a BamHI site and stop codon ( 5′-CGCGCGGATCCCTAGGCCTCAGTGGGTGC-3′ ) . The restriction sites are underlined in the primer sequences . The CapΔRSP cDNA which encodes amino acid residues 1–267 of capsid , was constructed by PCR using the forward EcoRI and Kozak site-containing forward primer ( 5′-CGGAATTCGCCACCATGGCTTCCACTACCCCCATCACC-3′′ ) and a reverse primer with a BamHI site and stop codon ( 5′-CGCGCGGATCCCTACTCGGTGGTGTGAGGG-3′ ) . The template cDNA for the CapNT and CapΔRSP PCR reactions was pCMV5-capsid . The CapC5RA expression plasmid was prepared by PCR using a forward primer containing an EcoRI site and a Kozak site ( 5′-TCACGGAATTC-3′ ) and a reverse primer containing a BamHI site ( 5′-TCAGGATCCCTAGGCGCGCGCGGTGC-3′ ) . The template DNA was pBRM33-CR5A . The CapsidNT , CapsidΔRSP , and CapsidC5RA cDNAs were resulting products were sublconed into the EcoRI and BamHI sites of the mammalian expression vector pCMV5 [69] to produce pCMV5-CapNT , pCMV5-CapΔRSP and pCMV5-CapCR5A respectively . For establishing capsid and luciferase expressing stable cell lines , the Lenti-X-tet-On advanced inducible expression system ( Clontech Laboratories , La Jolla , CA ) was utilized . A cDNA encoding full-length capsid was amplified by PCR using a forward primer with a BamHI site and Kozak consensus ribosome binding site ( 5′-TAGGATCCGCCACCATGGCTTCCACTACCCCCATCACC-3′ ) and a reverse primer with a EcoRI site ( 5′-GGCCGAATTCCTAGGCGCGCGCGGTGC-3′ ) respectively , where the restriction sites are underlined . The DNA used as a template was pCMV5-capsid . The PCR product was digested with BamHI and EcoRI , and subcloned into the pLVX-Tight-Puro vector . Lentivirus-production in HEK 293T and transduction of A549 cells were performed as per the manufacturer's instructions . At 48 hours post-transduction , cells were split 1∶2 into medium containing G418 ( 500 µg/ml ) and puromycin ( 0 . 5 µg/ml ) . Surviving cells were tested for inducible expression of capsid by indirect immunofluorescence and immunoblot analyses . The resulting polyclonal stable cell lines A549-Luciferase and A549-Capsid were maintained in media containing G418 ( 250 µg/ml ) and puromycin ( 0 . 25 µg/ml ) . To induce capsid or luciferase gene expression doxycycline ( 1 µg/ml ) was added to the culture medium . Codons for arginine-to-alanine mutations in the C-terminus of capsid were introduced into the RV M33 infectious clone ( pBRM33 ) [70] by a two step cloning procedure . A 421 base pair synthetic fragment ( Epoch BioLabs Inc , Sugarland , TX ) containing five arginine-alanine substitutions ( R264 , 268 , 271 , 275 , 277 ) was used to replace the analogous region in pCMV5-24S [68] . The resulting plasmid was named pCMV5-24S-CR5A . Next , the BsrGI/BamHI fragment from pCMV5-24S CR5A was used to replace the analogous region ( BsrGI/BamHI ) in pBRM33 resulting in the infectious clone pBRM33-CR5A . Total RNA samples were isolated with TRI Reagent ( Ambion ) from Vero cells infected with M33 ( wild type ) or CR5A strains of RV ( MOI: 1 ) according to the manufacturer's instructions . Prior to the RT-PCR reaction , 1 µg of total RNA was treated with 2 U of amplification grade DNase I ( Invitrogen ) as per manufacturer's recommendations . The DNase I-treated RNA samples were reverse transcribed to single-stranded cDNA using qScript Flex cDNA synthesis kit and Oligo ( dT ) 20 primer ( Quanta Biosciences , Gaithersburg , MD ) as per manufacture's instructions . Quantitative PCR reactions were conducted on a MX3005P thermocycler ( Stratagene , La Jolla , CA ) using a PerfecCTa SYBR green supermix low Rox real-time PCR kit ( Quanta Biosciences ) . Reactions were carried out by triplicate in a total volume of 25 µl containing 5 µl of cDNA and 0 . 2 µM of each oligonucleotide primer . Primers used to amplify RV nucleotides 5520–5706 from the RV p90 non-structural protein coding region of the RV genome were as follow: RV-F ( 5′-AGGTCATGTCTCCGCATTTC-3′ ) and RV-R ( 5′-GTCCCGAGTAGCAAGGGTCT-3′ ) . The amplification cycles for p90 samples consisted of an initial denaturating cycle at 95°C for 3 min , followed by 40 cycles of 15 s at 95°C , 30 s at 58°C , and 20 s at 72°C . Fluorescence was quantified during the 58°C annealing step , and the product formation was confirmed by melting curve analysis ( 57°C to 95°C ) . As an internal control , levels of the house keeping gene product cyclophilin A determined . Amplification was performed using the following primers , CYP-F ( 5′- TCCAAAGACAGCAGAAAACTTTCG-3′ ) and CYP-R ( 5′-TCTTCTTGCTGGTCTTGCCATTCC-3′ ) . The amplification cycles for Cyclophilin A consisted of an initial denaturating cycle at 95°C for 3 min , followed by 40 cycles of 15 s at 95°C , 20 s at 60°C , and 40 s at 72°C . Fluorescence was quantified during the 60°C annealing step , and the product formation was confirmed by melting curve analysis ( 57°C to 95°C ) . Quantification of the samples was performed using the two standard curves method [71] , and the relative amount of RV genomic RNA was normalized to the relative amount of Cyclophilin A mRNA . Three independent PCR analyses were performed for each sample . A549 cells ( 1×105 ) in 35 mm culture dishes were infected with the M33 strain of RV ( MOI = 2 ) and then incubated for 48 hours at 35°C prior to lysis . Alternatively , A549 cells were transiently transfected with pCMV5-capsid , pCMV5-CapNT , pCMV5-CapCT , pCMV5-CapΔSP , pCMV5-CapΔRSP or pCMV5-CapCR5A using Lipofectamine 2000 ( Invitrogen ) . Cells were lysed in 1% ( wt/vol ) CHAPS , 150 mM NaCl , 50 mM Tris , pH 8 . 0 containing Complete EDTA-free protease inhibitors ( Roche ) or 1% NP-40 , 150 mM NaCl , 2 mM EDTA , 50 mM Tris , pH 7 . 4 containing protease inhibitors . Cell lysates were clarified by centrifugation at 10 , 000× g for 10 minutes at 4°C . Immunoprecipitation was performed with clarified lysates and 1 µg/ml of mouse monoclonal anti-Bax6A7 ( Sigma ) , or 1∶1000 of rabbit polyclonal anti-capsid serum ( 7W7 ) , or 2 µg/ml of rabbit anti-Bak ( Millipore ) antibodies overnight at 4°C with rotation . Fifteen microliters of protein A or protein G sepharose ( 50% suspension ) were added and then samples were rotated for 1 hour at 4°C before washing; twice with lysis buffer and once with PBS . Proteins were eluted from the beads by boiling in protein gel sample buffer , separated by SDS-PAGE , and then transferred to polyvinylidene fluoride ( PVDF ) membranes ( Immobilon-P Millipore , Bedford , MA ) . Membranes were incubated for 1 hour at room temperature with the following antibodies and dilutions: 1∶1000 rabbit anti-RV capsid ( 7W7 ) [61] , 1∶1000 mouse anti-capsid ( H15C22 ) , 1∶1000 goat anti-RV ( Meridian Life Science Inc , Saco , Maine ) , 1∶1000 rabbit anti-Bak ( Cell Signaling ) , 1∶1000 rabbit anti-Bax ( Abcam ) or 1∶5000 mouse anti-Bax ( YTH-2D2 , Trevigen Inc , Gaithersburg , MD ) . To detect E1 glycoprotein by immunoblotting , it was necessary to perform SDS-PAGE under non-reducing conditions . E1 was detected using a 1∶1000 dilution of a mouse monoclonal antibody ( H2C213 ) obtained from Abbott Labs ( Abbott Park , IL ) . After three washes with Tris-Buffered-Saline-Tween ( TBS-T ) , the membranes were incubated with either goat anti-rabbit , goat anti-mouse or rabbit anti-goat horseradish peroxidase-conjugated IgG ( Bio-Rad Hercules , CA ) for 1 hour . Membranes were washed four times with TBS-T and immunoreactive proteins were detected using Supersignal West Pico chemiluminescent substrate ( Pierce Biotechnology , Rockford , IL ) followed by exposure to X-ray film ( Fuji Photo Film Co , LTD , Tokyo , Japan ) . A549-Capsid or A549-Luciferase cells were cultured in the presence of doxycycline and after 36 hours , anti-human Fas activating clone CH11 antibody ( 0 . 12 µg/ml ) ( Millipore , Temecula , CA ) and cycloheximide ( 10 µg/ml ) were added to the cultures for 6 hours . Cells were then lysed in 1% NP-40 buffer containing a cocktail of protease inhibitors . The lysates were centrifugated at 10 , 000× g for 10 min at 4°C , and protein concentrations were determined by BCA protein assay ( Pierce Biotechnology , Rockford , IL ) using bovine serum albumin as a standard . Equivalent amounts of protein ( 60 µg ) from each lysate were resolved in 8% SDS-PAGE and transferred to PVDF membranes followed by immunoblotting with mouse monoclonal anti-cleaved PARP ( Asp214 ) clone 19F4 antibody ( Cell Signaling ) . A549 and Vero cells cultured on glass coverslips were infected with RV ( MOI = 1 ) or transiently transfected with pCMV5-capsid or pCMV5-CapNT or pCMV5-CapCT or pCMV5-CapΔSP or pCMV5-CapΔRSP or pCMV5-E2E1 or pcDNA3 . 1-Bcl-XL , and peGFP-Bax or peGFP-Bax ( Gift of Dr . Michele Barry , University of Alberta ) using Lipofectamine 2000 ( Invitrogen ) . After 24 or 48 hours post-infection or post-transfection as indicated , cells were fixed in 4% paraformaldehyde for 20 min , followed by quenching with PBS containing 50 mM ammonium chloride . Cell membranes were permeabilized with PBS containing 0 . 2% Triton-X-100 for 5 min before incubation with primary and secondary antibodies . All the washes were done in PBS containing 0 . 1 mM CaCl2 and 1 mM MgCl2 . RV proteins were detected with rabbit anti-capsid ( 7W7 ) , mouse anti-capsid ( H15C22 ) , mouse-anti E1 ( H2C213 ) , goat anti-RV , or human anti-RV ( GB ) which has been described previously . Mitochondria were detected using rabbit anti-cytochrome c ( from Dr . L . Berthiaume , University of Alberta ) or with a mouse anti-complex II monoclonal antibody ( Mitosciences , Eugene , OR ) . Activated isoforms of Bax and capsase 3 were detected with a Bax-specific mouse monoclonal antibody 6A7 ( Abcam ) or caspase 3-specific rabbit monoclonal antibody ( BD Pharmingen ) respectively . Primary antibodies were detected with Alexa Fluor 594 chicken anti-mouse , Alexa Fluor 488 donkey anti-rabbit , Alexa Fluor 488 donkey anti-mouse , Alexa fluor 637 Donkey anti-rabbit and/or Alexa 594 goat anti-rabbit ( Molecular Probes , Invitrogen , Carlsbad , CA ) . Coverslips were mounted onto microscope slides using ProLong Gold antifade reagent with 4'-6-Diamidino-2-phenylindole ( Molecular probes , Invitrogen ) . Samples were then examined using Zeiss 510 confocal microscope or a Zeiss Axioskop2 microscope equipped with a CoolSNAP HQ digital camera ( Photometrics ) . A549-Cap or A549-Luciferase cells were cultured for 36 hours in the presence of doxycycline , followed by incubation with 1 µM staurosporine ( Sigma-Aldrich ) or anti-Fas antibody ( 0 . 12 µg/ml ) and cycloheximide ( 10 µg/ml ) for 6 hours . Cells were then homogenized in ice-cold mitochondria isolation buffer containing 200 mM mannitol , 70 mM sucrose , 10 mM Hepes , and 1 mM EGTA ( pH: 7 . 5 ) using a dounce homogenizer with a loose fitting pestle . Unbroken cells and nuclei were pelleted by centrifugation at 500× g for 10 min at 4°C . The supernatants were then centrifuged at 10 , 000× g for 20 min at 4°C to obtain crude mitochondrial pellets that were cross-linked with 10 mM bis ( maleimido ) hexane ( BMH; Thermo Scientific ) for 30 min at room temperature . Samples then were separated on 4–12% polyacrylamide gels and then processed for immunobloting with rabbit polyclonal antibodies to Bax antibody ( Abcam , Cambridge , MA ) and capsid ( 7W7 ) . Expression of capsid or luciferase in A549-cap or A549-luc cells respectively was induced with doxycycline for 36 hours , followed by incubation with staurosporine ( 1 µM ) or anti-Fas antibody ( 0 . 12 µg/ml ) to induce apoptosis . Cells then were stained with 0 . 2 µM Tetramethylrhodamine methyl ester ( TMRM ) ( Invitrogen , Molecular probes ) for 30 min at 37°C before analyses by flow cytometry ( FACScan , Becton Dickinson ) . For each sample , 10 , 000 events were acquired . Data were analyzed using CellQuest software . The percentage of killing was calculated as the number of TMRM-negative cells divided by the total number of cells , and standard deviations were determined from three independent experiments . For Bax or Bak killing assays , A549 or Hel-18 cells were transfected with peGFP-Bax or peGFP-Bak together with pCMV5 , pCMV5-CapNT , or pCMV5-capsid using Lipofectamine 2000 or Lipofectamine LTX respectively ( Invitrogen ) . After 24 hours , cells were stained with 0 . 2 µM TMRM for 30 min at 37°C prior to analyses by two-color flow cytometry . TMRM fluorescence was detected through the FL-2 channel equipped with a 585-nm filter and eGFP fluorescence was measured using the FL-1 channel equipped with a 489-nm filter . Data were acquired on 10 , 000 eGFP-positive cells per sample , and analysis was performed using CellQuest software . The relative specific cell death was calculated as the number of eGFP-positive TMRM-negative cells divided by the total number of eGFP positive cells . Standard deviations were generated from three independent experiments . Vero cells ( 1×105/35 mm dish ) were transiently transfected with pCMV5-capsid or pCMV5-CapCR5A and pCMV5-E2E1 using Lipofectamine 2000 . Assembly and secretion of RV-like particles was assayed after 48 hours of transfection as described elsewhere [58] . Data from FACS and indirect immunofluorescence-based apoptosis assays were subjected to statistical analyses ( student's t test or one-way analysis of variance ( One-way ANOVA ) ) using Predictive Analytics Software ( version 17 . 0 . 3 ) ( SPSS Inc , Chicago , Il ) . | Among the variety of defense systems employed by mammalian cells to combat virus infection , apoptosis or programmed cell death is the most drastic response . Some large DNA viruses encode proteins whose sole function is to block apoptosis . Conversely , very little is known about whether RNA viruses encode analogous proteins . In many cases , RNA viruses are able to replicate before cell death occurs , which may be one reason why so little thought has been given to this topic . However , a number of RNA viruses , some of which are important human pathogens , have slow replication cycles and it stands to reason that they must block apoptosis during this time period . Here we show that the multifunctional capsid protein of Rubella virus is a potent inhibitor of apoptosis . Data from reverse genetic experiments suggest that the anti-apoptotic function of a virus-encoded protein is important for replication of an RNA virus . We anticipate that other slowly replicating RNA viruses may employ similar mechanisms and , as such , these studies have implications for development of novel anti-virals and vaccines . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"virology/host",
"antiviral",
"responses",
"virology",
"virology/viral",
"replication",
"and",
"gene",
"regulation",
"cell",
"biology/cellular",
"death",
"and",
"stress",
"responses"
] | 2011 | The Rubella Virus Capsid Is an Anti-Apoptotic Protein that Attenuates the Pore-Forming Ability of Bax |
Extensive experimental information supports the formation of ligand-specific conformations of G protein-coupled receptors ( GPCRs ) as a possible molecular basis for their functional selectivity for signaling pathways . Taking advantage of the recently published inactive and active crystal structures of GPCRs , we have implemented an all-atom computational strategy that combines different adaptive biasing techniques to identify ligand-specific conformations along pre-determined activation pathways . Using the prototypic GPCR β2-adrenergic receptor as a suitable test case for validation , we show that ligands with different efficacies ( either inverse agonists , neutral antagonists , or agonists ) modulate the free-energy landscape of the receptor by shifting the conformational equilibrium towards active or inactive conformations depending on their elicited physiological response . Notably , we provide for the first time a quantitative description of the thermodynamics of the receptor in an explicit atomistic environment , which accounts for the receptor basal activity and the stabilization of different active-like states by differently potent agonists . Structural inspection of these metastable states reveals unique conformations of the receptor that may have been difficult to retrieve experimentally .
G-protein coupled receptors ( GPCRs ) are versatile signaling proteins that functionally couple a host of extracellular stimuli to intracellular effectors , thus mediating several vital cellular responses . The majority of marketed drugs act as agonists , inverse agonists , or antagonists at these receptors depending on whether they increase , reduce , or have no effect on the so-called ‘basal activity’ that characterizes unliganded GPCRs for diffusible ligands . Not only can a specific GPCR activate different G-protein or arrestin isoforms [1] , but a single ligand can display different efficacy for different signaling pathways , an observation that has been dubbed “functional selectivity” , “agonist trafficking” , “biased agonism” , “differential engagement” , or “protean agonism” in the literature [2]–[6] . At the molecular level , a simple explanation for this phenomenon is that ligands with varied efficacies can shift the conformational equilibrium of a GPCR towards different conformations of the receptor , which in turn can activate one or another intracellular protein . Although several spectroscopy studies ( e . g . , for the β2-adrenergic receptor , herein referred to as B2AR , see [7]–[9] ) have been instrumental in showing that ligands with different efficacies stabilize GPCR conformational states that are structurally and kinetically distinguishable , perhaps the most direct evidence of ligand-induced conformational specificity comes from the recent high-resolution crystallographic structures of several different ligand-bound GPCRs . In the majority of cases , these structures were obtained in the presence of an inverse agonist , and therefore in an inactive state . Only very recently have high-resolution crystal structures of agonist-bound GPCRs started to appear in the literature [10]–[15] . However , possibly restrained by crystallization conditions , not all these agonist-bound structures present the features that are usually attributed to an active GPCR conformation , most typically: the large outward movement of transmembrane helix 6 ( TM6 ) with respect to the center of the receptor helical bundle , which is accompanied by the disruption of an important salt bridge between the conserved D/E3 . 49-R3 . 50 pair and E6 . 30 , commonly referred to as the “ionic lock” . Residue numbering here and throughout the text follows the Ballesteros-Weinstein notation [16] . According to this notation , each residue is indicated by a two-number identifier N1 . N2 where N1 is the number of the transmembrane helix , and N2 is the residue number on that helix relative to its most conserved position , which is designated N2 = 50 . We direct the reader elsewhere ( e . g . , [17] , [18] ) for recent reviews of all the relevant structural changes that have been attributed by various biophysical techniques to active forms of GPCRs . A different extent of structural rearrangement was noted at the binding site of high-resolution crystal structures of GPCRs depending on the type of ligand to which they were bound . For instance , only minor local structural changes were noted between the high-resolution crystal structures of the B2AR in the presence of inverse agonists such as carazolol [19] , timolol [20] , ICI-118 , 551 [21] , or a compound deriving from virtual screening [21] and the neutral antagonist alprenolol [21] . Slightly more pronounced differences were noted by comparing these inverse agonist/antagonist-bound binding pockets with those stabilized by full agonists ( i . e . , either the covalently-bound ligand FAUC50 [11] or BI-167107 [10] ) . Among them , the most notable differences were the hydrogen bonding contacts that only agonists formed with S5 . 42 and S5 . 46 on TM5 . Similar interactions helped discriminate between inverse agonist-bound crystal structures of the β1-adrenergic receptor ( B1AR ) and structures obtained in the presence of full agonists ( e . g . , isoprenaline or carmoterol ) [13] . Notably , only one of these two hydrogen bonds involving TM5 , specifically the interaction with S5 . 42 , was also present in structures stabilized by the partial agonists salbutamol or dobutamine , suggesting a distinguishable binding mode between full and partial agonist structures [13] . Analogous to the cases of the B1AR and B2AR , where specific residues ( i . e . , S5 . 46 ) are found to bind uniquely to agonists , key residues ( S7 . 42 and H7 . 43 ) that bind agonists ( either adenosine or NECA ) but not antagonists ( ZM241385 ) were revealed by the very recent crystal structures of a thermostabilized construct of the adenosine A2A receptor [15] . Unlike another recent crystal structure of this receptor stabilized by both T4-lysozyme and the conformationally selective agonist UK-432097 [12] , these agonist-bound structures did not display changes at the cytoplasmic side that resemble those of an active state of a GPCR . In addition to the crystal structure of the adenosine A2A receptor bound to UK-432097 [12] , these more marked changes at the cytoplasmic side have so far only been observed in the high-resolution crystal structures of opsin [22] , [23] , Meta II rhodopsin [14] , and the nanobody-stabilized B2AR [10] . Despite these recent remarkable achievements in structural biology of GPCRs , the majority of pharmacologically relevant ligands of these receptors do not appear to be ideally suited for the stabilization and crystallization of these receptors , most likely because of their low affinity , slow off-rate , and poor solubility . Not only might this prevent the identification of physiologically relevant conformational states of a given GPCR , but it is considered a limiting bottleneck for the characterization of different structures of these receptors . Molecular dynamics ( MD ) simulations can help to fill this information gap by enabling an atomic-level characterization of ligand-specific conformations that are impossible or difficult to retrieve experimentally . Moreover , these simulations allow extension of static structural data into dynamic representations , thus laying the basis for a mechanistic understanding of the selective activation of GPCR-mediated signaling pathways . To enable characterization of large conformational changes within the limited timescales commonly accessible to MD simulations , and to evaluate the extent to which ligands with different efficacies affect the free-energy landscape of GPCRs , we implemented a computational strategy employing a combination of different adaptive biasing techniques . Specifically , we used well-tempered metadynamics [24] to identify metastable states of a GPCR along putative activation pathways between inactive and active crystallographic states determined by adiabatic biased MD . We tested the accuracy of this strategy in reproducing crystallographic [19] , [21] and/or spectroscopic [7]–[9] data available for the B2AR in its interaction with either a full agonist ( i . e . , epinephrine ) , a weak partial agonist ( i . e . , dopamine ) , a very weak partial agonist ( i . e . , catechol ) , two inverse agonists ( i . e . ICI-118 , 551 and carazolol ) , or one neutral antagonist ( i . e . , alprenolol ) . The results show a clear ligand-induced modulation of the free-energy landscape of the receptor with shifts in the conformational equilibrium towards inactive or active conformations depending on the physiological response elicited by the simulated ligand .
A model of the B2AR ( Figure S1 was prepared starting from one of the available crystal structures of this receptor ( PDB ID: 2RH1 ) , removing the lysozyme insertion , and modeling the missing intracellular loop 3 ( IL3 ) with the Rosetta ab-initio loop modeling protocol [25] . The intracellular loop 2 ( IL2 ) , which is probably misfolded [26] , [27] in the inactive structure of the B2AR ( 2RH1 ) , but in a helical conformation in the active nanobody-stabilized crystal ( 3P0G ) of the receptor , was also replaced by the lowest-energy Rosetta model with a helical fold . The resulting receptor model was embedded into an explicit 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine ( POPC ) /10% cholesterol membrane bilayer using a pre-equilibrated 8×8×10 nm patch hydrated with SPC/E water molecules , and the procedure described in [28] . As found in the crystal structure [19] , one palmitoyl group was covalently attached to a C-terminal residue ( Cys 341 ) of the receptor before insertion in the membrane . The system was then hydrated with SPC/E water molecules [29] and Na+ and Cl– ions were added to ensure charge neutrality . The resulting system of ∼50 , 000 total atoms was equilibrated with unbiased MD simulations for 20 nanoseconds ( ns ) using the Optimized Potentials for Liquid Simulations all-atom ( OPLS-AA ) force field [30] for the receptor and united-atoms Berger parameters for the lipids [31] . The Gromacs 4 . 0 . 7 [32] package enhanced with the Plumed plug-in [33] was used for all simulations . Specifically , NPT simulations were carried out under periodic boundary conditions , using the Parrinello-Rahman algorithm [34] with a time constant of 1 . 0 ps and a reference pressure of 1 bar to control pressure , and the Nose-Hoover [35] algorithm with a time constant of 1 . 0 ps to maintain a constant temperature of 300 K . Prior to production runs ( summarized in Table S1 ) , the system was equilibrated by a series of three 0 . 2 ns runs with progressively weaker restraints on the protein backbone followed by a 3 . 0 ns unconstrained equilibration . We used the standard Gromacs leap-frog [32] algorithm with a time step of 2 . 0 fs , LINCS algorithm [36] to preserve the bond lengths , and SETTLE algorithm [37] to maintain the geometry of the water molecules . Lennard-Jones interactions were treated with a twin-range cutoff of 0 . 9∶1 . 4 nm and an integration time step of 2 . 0 fs; the neighbor list was updated every 10 steps . Electrostatic interactions were described using the particle-mesh Ewald method [38] , with a cutoff of 0 . 9 nm for real-space interactions , and a 0 . 12-nm grid with fourth-order B-spline interpolation for reciprocal-space interactions . The starting equilibrated unliganded conformation of the B2AR within the lipid bilayer was subjected to ten independent Adiabatic Biased MD ( ABMD ) simulations [39] , [40] to obtain transition pathways of the receptor from an inactive to an active conformation , built using the coordinates of the nanobody-stabilized crystal structure ( PDB code: 3P0G ) . Briefly , this method biases the system towards a given value χ0 of a predefined order parameter χ ( R ) , where R represents the coordinates of the atoms in the system . A harmonic bias acts only when the distance of χ ( R ) from the target χ0 is bigger than its minimum value previously reached during the simulation ( i . e . if χ ( R ( t ) ) - χ0 > mins<t χ ( R ( s ) ) - χ0 ) , according to the following equation: ( 1 ) The order parameter χ measures the distance from the putative activated conformation of the receptor , and is defined as the Cα root mean square deviation ( RMSD ) from the active conformation of the B2AR ( all residues were included except the long flexible IL3 ) . In order to obtain activated final states , the simulation was run with χ0 = 0 . After carrying out 10 independent ABMD runs with an elastic constant of k = 10 kcal/ ( mol⋅nm2 ) , the trajectories were pooled and clustered using an average linkage agglomerative algorithm and the same dissimilarity measure used to run ABMD . Bonded and van der Waals interactions for the ligands were assigned manually choosing the appropriate OPLS-AA atom types [30] for each atom in the molecule . Coulomb point charges were obtained according to the RESP approach [41] from quantum chemical calculations ( i . e . , geometry optimization using Gaussian 03 [42] and restricted Hartree-Fock calculations with the 6-31G* basis set ) . Ligands for which an experimental crystal structure in complex with the B2AR is available , i . e . 2RH1 for carazolol [19] , 3NY8 for ICI-118 , 551 [21] , and 3NYA for alprenolol [21] , were positioned in the binding pocket accordingly . The other ligands , i . e . the full agonist epinephrine and the partial agonists dopamine and catechol , were docked into the initial inactive model of B2AR , using a standard Autodock 4 . 0 protocol [43] , [44] . Inferences from agonist-bound crystal structures of B2AR [10] , [11] and B1AR [13] were taken into account when selecting the most accurate initial binding poses of these ligands for free-energy calculations . Notably , simulations of initial conformations comprising slightly different binding poses produced similar free-energy profiles ( data not shown ) . The free-energy profiles of liganded and unliganded systems were estimated using metadynamics [45]–[47] . Briefly , this technique enables an efficient reconstruction of the free-energy as a function of a set of k predetermined order parameters , referred to as collective variables si ( R ) , 1≤i≤k . A history-dependent bias potential is added to the force-field driving the system dynamics so as to discourage the re-visiting of regions of the si phase space that have already been explored . Specifically , the bias potential is ( 2 ) where t′ is a multiple of a deposition time τ and the values of wt′ and σi regulate the shape and size of the Gaussian bias contributions . In the original metadynamics algorithm , wt′ = w is constant , and the free-energy profile can be estimated up to an insignificant additive time-dependent constant as W ( R ) = − limt→∞ V ( R , t ) . Here , we used well-tempered metadynamics [24] , a variant of the original metadynamics algorithm that enables assessment of simulation convergence while keeping the computational effort focused on physically relevant regions of the conformational space . In this variant of the method , the value of wt′ depends on the bias accumulated up to t′ according to the equation: ( 3 ) where ΔT is a constant with the dimension of a temperature , kB is the Boltzmann constant , and w is a constant energy representing the maximum height of the Gaussian biases . Since in the regions where the bias is higher the exponential factor reduces the rate of the bias update , the bias potential smoothly converges to a constant value in time , and the underlying free-energy can be derived by ( 4 ) where T is the temperature at which the simulation is performed . To efficiently sample the conformational space along the activation pathway , reference states from the clustered ABMD runs were selected by cutting the agglomerative tree at 30 clusters , and selecting from them n = 10 clusters homogeneously covering the pathway . The reference states Rj ( 1≤j≤n ) were numbered assigning j = 1 to the cluster closer to the inactive state ( Cα RMSD from 2RH1 ∼0 . 4 Å ) and j = 10 to the one closer to the active state ( Cα RMSD from 3P0G ∼0 . 3 Å ) . Two path collective variables describing the position along ( s ) and the distance from ( z ) the pathway were defined [48] as follows: ( 5 ) ( 6 ) where d ( R , Rj ) is the squared Cα RMSD ( excluding IL3 ) with respect to the reference structure Rj , and Z = ∑1≤j≤n exp ( − γ d ( R , Rj ) ) . The simulations were performed choosing γ = 1/0 . 25 Å-2 , σs = 0 . 1 , and σz = 1 Å2 , and well-tempered metadynamics was used with a bias factor ΔT = 10 T , an initial value of w = 0 . 4 kcal/mol , and a deposition interval τ = 8 ps . Metadynamics simulations were run for 300 ns , time at which the reconstructed free-energy difference between the metastable states converged to 0 . 2 kcal/mol . Since the trajectory was generated adding the metadynamics bias , the resulting conformations cannot be used to obtain statistical information on order parameters other than the collective variables . However , it is possible to unbias the distribution of any given function of the system coordinates using the algorithm described in [49] . This so-called reweighting technique was used to estimate the free-energy surface of the complexes as a function of three important descriptors of receptor activation , namely the distance between R3 . 50 and E6 . 30 ( the “ionic lock” ) , the rotamer of residue W6 . 48 ( the so-called “toggle switch” ) , and the outward displacement of the intracellular segment of TM6 . Three order parameters were defined to monitor the behavior of these changes upon activation . For the ionic lock , we defined dIL = ||〈R3 . 50〉 − 〈R6 . 30〉|| , where 〈R3 . 50〉 and 〈R6 . 30〉 represent the center of mass of the η-nitrogens of R3 . 50 and the δ-oxygens of E6 . 30 , respectively . For the toggle switch , we monitored the first dihedral angle χTS of the side chain of W6 . 48 . Finally , the movement of TM6 was measured by aligning the receptor to the inactive crystal structure ( 2RH1 ) and calculating the distance dTM6 = ||M − 〈R6 . 35〉|| ( angled brackets indicate the centroid of all the atoms of the residues ) between the midpoint of an imaginary line connecting residues K6 . 35 and Y2 . 41 in the inactive structure , M = ½[〈R2 . 41〉 + 〈R6 . 35〉] ( located roughly at the center of the intracellular exposed surface of the receptor ) , and residue K6 . 35 . The outward movement is described by the difference in dTM6 values between any given conformation and the reference inactive crystal structure , i . e . by ΔdTM6 = dTM6–dTM6 ( 2RH1 ) . Representative conformational states of the metastable energy basins identified by metadynamics were selected and their structural stability analyzed . Specifically , standard , unbiased , NPT molecular dynamics simulations of these conformational states were initiated by randomizing new initial starting velocities with the Maxwell distributions at 300 K , and were run for ∼50 ns using the same simulation parameters described above .
An activation pathway from the inactive to the active B2AR crystal structures ( PDB codes 2RH1 and 3P0G , respectively ) was obtained by ABMD following the protocol described in the Materials and Methods section . This pathway was used to define the s and z collective variables ( see the Materials and Methods section for corresponding equations ) that were employed for the metadynamics simulations . Panel A of Figure 1 illustrates the free-energy ? G of the unliganded receptor as a function of the position s along the activation pathway following integration of the dependence on z . Specifically , s = 0 . 0 and s = 1 . 0 indicate the inactive and fully activated extreme conformations of the pathway , respectively . This free-energy profile shows two minima , one at s∼0 . 2 that is close to the inactive state and the other at s∼0 . 6 that is shifted towards the active state . The two states are separated by a barrier of ∼2 . 5 kcal/mol , but they have a similar overall stability ( ΔG<kBT ) , and are therefore equally populated at equilibrium . Inspection of the entire two-dimensional free-energy profile ΔG ( s , z ) reported in the supplementary material ( see panel A of Figure S2 ) shows that these states correspond to conformations along the activation pathway with values of z close to 0 . Visual inspection of a representative structure of the s∼0 . 2 energy basin confirms that the corresponding transmembrane bundle is very close to the inactive B2AR crystal structure ( Cα RMSD excluding IL3 ∼0 . 6 Å ) , as substantiated by the very small outward movement of TM6 ( ΔdTM6 ∼ 0 . 4 Å ) with respect to the inactive crystal ( see panel B of Figure S2 ) . In contrast , a representative structure of the second energy basin at s∼0 . 6 ( RMSD ∼1 . 6 Å and ∼1 . 1 Å from the inactive and active crystal structures , respectively ) displays a more pronounced outward movement of TM6 ( ΔdTM6 ∼2 . 5 Å in Figure S2 ) . Figure 1B shows the free-energy of the unliganded B2AR as a function of order parameters that monitor molecular switches which have traditionally been reported as descriptors of GPCR activation . Specifically , these molecular switches are: 1 ) the ionic lock between TM3 and TM6 , herein monitored using the distance dIL between R3 . 50 and E6 . 30 and 2 ) the W6 . 48 rotamer toggle switch , herein monitored using the first dihedral angle χTS of the residue side chain . Whilst the latter has not been observed in recent activated crystal structures of GPCRs , compelling spectroscopic data exist supporting a rotamer change of the W6 . 48 side chain upon activation [50] . Two different energy basins can be identified in the plot of Figure 1B: a more stable one , labeled a , in which both molecular switches are in their inactive conformation ( dIL∼3 Å and χTS∼163° ) , and a second basin , labeled c , where both switches are in their activated conformation ( dIL∼12 Å and χTS∼55° ) . The two basins are separated by a barrier of ∼3 . 0 kcal/mol . A transition state at χTS∼65° and dIL∼5 Å ( labeled b on the free-energy map ) suggests a preferential rotamer toggle switch prior disruption of the ionic lock . The neutral antagonist alprenolol , consisting of an “aromatic head” ( a 2-allyl-pheniloxyl moiety ) and an “aliphatic tail” ( oxy-propanol-amine ) ( see chemical structure in Figure 2A ) , was docked in accordance to the binding mode assumed by the ligand in the crystal structure of the corresponding ligand-bound receptor [21] . The results of the simulations for the alprenolol-bound receptor are illustrated in panels A-C of Figure 2 . As shown in Figure 2A , the overall shape of the free-energy profile of the alprenolol-bound B2AR as a function of the position ( s ) along the activation pathway is qualitatively similar to the profile obtained for the unliganded receptor , and reported in Figure 1A . A similarity is also noted between the two-dimensional energy surfaces of the alprenolol-bound ( Figure S3A ) and the unliganded B2AR ( Figure S2A ) . In spite of these qualitative similarities , the inactive state at s∼0 . 2 is more stable ( ∼1 kcal/mol ) than the intermediate state at s∼0 . 6 for the alprenolol-bound receptor compared to the unliganded one . Given the relatively higher stability of the alprenolol-bound receptor conformation with no significant outward movement of TM6 ( ΔdTM6 ∼0 . 4 Å at s∼0 . 2 in Figure S3B ) , these results suggest an energy profile that is more suitable for a very weak inverse agonist rather than a neutral antagonist . Notably , data are available in the literature in support of an inverse agonist [51] , [52] ( or even a partial agonist [53] ) role for alprenolol . Figure 2B shows a representative conformation of the lowest energy basin identified for the alprenolol-bound receptor . In this conformation , and similar to the corresponding crystal structure [21] , the alprenolol charged moiety in its aliphatic tail forms interactions with polar residues D3 . 32 and N7 . 39 , while the ligand aromatic head interacts with residues V3 . 33 , V3 . 36 , F6 . 51 , N6 . 55 , Y5 . 38 , and S5 . 42 , which define a cleft formed by TM3 , TM5 and TM6 . Figure 2C shows that the energetically most stable alprenolol-bound inactive state is characterized by inactive molecular switches ( χTS∼160° and dIL∼5 Å ) . This state , labeled a in Figure 2C , is separated by an energy barrier of ∼3 kcal/mol from the second most stable energetic minimum at χTS∼50° and dIL∼12 Å ( c in Figure 2C ) , with a transition state ( b in Figure 2C ) at χTS∼85° and dIL∼5 Å . Thus , the presence of alprenolol in the binding pocket does not appear to disrupt the free-energy profile seen in the unliganded receptor , further confirming a possible rotamer toggle switch of the W6 . 48 residue prior breaking of the ionic lock . The stability of alprenolol in a representative state of the ligand-receptor complex extracted from the most stable energy basin at s∼0 . 2 was confirmed by carrying out ∼50 ns unbiased MD simulations . The evolution of the ligand and the protein RMSD during these simulations is reported in Figure S4 . We assessed the effect of two different B2AR inverse agonists , namely ICI-118 , 551 and carazolol , on the free-energy landscape of the receptor during transition from inactive to activated experimental states . Carazolol and ICI-118 , 551 share important structural features with alprenolol , e . g . , they both have an “aliphatic tail” ( oxy-propanol-amine for carazolol and oxy-butanol-amine for ICI-118 , 551 ) and an “aromatic head” . The results of the simulations for the carazolol-bound and the ICI-118 , 551-bound receptor are illustrated in panels A-C and D-F of Figure 3 , respectively . In the presence of either carazolol or ICI-118 , 551 , the B2AR free-energy profiles ( Figure 3A and 3D , respectively ) show a single lowest-energy basin at s∼0 . 18 close to the inactive state of the receptor . These much more stable energy basins are also present in the two-dimensional energy surfaces of the carazolol-bound ( Figure S5A ) and the ICI-118 , 551-bound ( Figure S5C ) B2AR , and comprise inactive conformations as further illustrated by the lower energy values for states characterized by the absence of outward movement of TM6 ( ΔdTM6 ∼ 0 . 4 Å in Figures S5B and S5D ) . Representative conformations extracted from the lowest energy basins of either the carazolol-bound ( Figure 3B ) or the ICI-118 , 551-bound ( Figure 3E ) receptors show that the energy-optimized binding poses of these ligands are very similar to their positions in the corresponding crystal structures [19] , [21] . Similar to the binding mode of alprenolol , the charged moieties contained in the aliphatic tails of these ligands interact with polar residues D3 . 32 , and N7 . 39 , while their aromatic heads are oriented toward TM3 , TM5 , and TM6 , thus directly interacting with residues in these helices ( see Figures 3B and 3E ) . To assess the stability of the ligands in these representative conformations , we performed standard , unbiased MD simulations . As shown in Figures S6A-D , which report the time evolutions of the RMSD of the protein , as well as those of the heavy atoms of carazolol and ICI-118 , 551 , after superposition of the receptor Cα atoms , the receptor conformations are stable and the binding modes of the ligands are conserved over a simulation time of ∼50 ns . The intermediate state at s∼0 . 6 that was significantly populated in the unliganded and neutral antagonist-bound receptor is much less stable at ΔG∼4 . 0 kcal/mol in the case of the carazolol-bound or ICI-118 , 551-bound receptors ( see Figures 3A and 3D , respectively ) . However , these are still metastable states , as judged by the presence of shallow minima at s∼0 . 6 in both the free-energy profiles , and are separated from the inactive states by multiple barriers . In terms of modulation of the toggle switch and the ionic lock , the free-energy as a function of χTS and dIL ( Figures 3C and 3F for the carazolol-bound and ICI-118 , 551-bound complexes , respectively ) features only one minimum in the inactive region of these molecular switches ( χTS∼160° and dIL∼3 Å ) . To study the effects of full agonists on the free-energy landscape of B2AR , we docked epinephrine into the receptor , and performed metadynamics calculations . Figure 4A shows the free-energy profile of the epinephrine-bound B2AR with the lowest energy state ( s∼0 . 9 ) likely to correspond to an activated conformation . The same observation is possible by inspection of the two-dimensional free-energy surface ( Figure S7A ) as well as the TM6 outward movement ( ΔdTM6 ∼5 . 9 Å in Figure S7B ) as a function of the position s along the activation pathway . However , a second low-energy metastable state is present in these free-energy profiles , close to the inactive state ( s∼0 . 2 ) , and with a free-energy difference of only ∼1 kcal/mol with respect to the most stable activated state . As illustrated in Figure 4B , our proposed binding mode of epinephrine within a fully activated B2AR ( energy basin at s∼0 . 9 ) is consistent with the binding poses displayed by full agonists in the B2AR [10] and B1AR [13] crystallographic structures . Specifically , the ligand amino group forms hydrogen bonds with D3 . 32 and N7 . 39 of B2AR , the ligand β-hydroxyl group interacts with D3 . 32 , and the ligand catecholamine hydroxyl groups interact through hydrogen bonding with the side chains of both S5 . 42 and S5 . 46 . In this state , the B2AR helix bundle is structurally very similar to the corresponding nanobody-activated crystal structure of the receptor ( C ? RMSD from 3P0G is ∼1 . 6 Å ) . The stability of the epinephrine binding pose and the specific receptor conformation were verified by carrying out ∼50 ns standard MD simulations ( see corresponding time evolutions of RMSD in Figure S8 ) . On the other hand , representative structures of the energy basin at s∼0 . 2 ( data not shown ) corresponded to conformations of the helix bundle very similar to the inactive crystal structure of B2AR ( Cα RMSD from 2RH1 is ∼1 . 0 Å ) . Two energy basins ( labeled a and c ) were identified from the free-energy as a function of the order parameters describing the ionic lock and rotamer toggle switches ( Figure 4C ) . Specifically , the basin comprising conformations in which both the ionic lock and rotamer toggle switches are in the ‘active’ ( dIL∼16 Å and χTS∼50° ) positions appear to be more stable than the basin with receptor conformations with ‘inactive’ ( dIL∼3 Å and χTS∼160° ) molecular switches . Also in this case , the minimum free-energy path between these two energy basins suggests activation of the toggle switch prior breaking of the ionic lock interaction along the path to full receptor activation . The weak and very weak partial agonists , dopamine and catechol , were also simulated in the context of the B2AR activation pathway . Figures 4D and 4G illustrate the free-energy profiles of the catechol-bound and dopamine-bound receptors , respectively . In both cases the receptor is most stabilized in an intermediate state ( s∼0 . 6 ) along the pathway to activation . Inspection of the free-energy as a function of the position ( s ) along and the distance ( z ) from the activation pathway ( see Figures S9A and S9C for the catechol-bound and dopamine-bound receptors , respectively ) confirms that these two ligands stabilize a state different from the inactive or fully activated ones as judged by the lowest energy values at z∼2 in Figure S9A for catechol , and at s∼0 . 6 , z∼0 . 0 in Figure S9C for dopamine . This difference is also evident from the free-energy surfaces as a function of the TM6 outward movement and the position along the activation pathway ( see Figures S9B and S9D , respectively ) , as well as from the structural superpositions shown in Figure 5 . Specifically , Figure 5 illustrates the structural differences between the TM regions of the predicted inverse agonist- and partial agonist-specific conformations ( Figure 5A ) , the inverse agonist- and full agonist-specific conformations ( Figure 5B ) , and the partial agonist- and full agonist-specific conformations ( Figure 5C ) . Figures 4E and 4H show the binding modes of catechol and dopamine , respectively . These binding poses were proven to be stable during ∼50 ns of unconstrained MD simulations ( see Figures S10A and S10B for the time evolution of the RMSD of catechol and dopamine , respectively , and Figures S10C and S10D for the time evolution of the RMSD of the corresponding protein Cα atoms ) . In agreement with inferences from recent B1AR structures co-crystallized with either full or partial agonists , these two B2AR partial agonists formed stable hydrogen bonds ( through the catechol moiety ) with the side chain of S5 . 42 , but do not with S5 . 46 . In terms of the ligand-induced modulation of the molecular switches , the catechol-bound B2AR state with a broken ionic lock ( located at χTS∼50° and dIL∼16 Å in Figure 4F ) is relatively less stable than the corresponding larger basin identified in the presence of dopamine ( see Figure 4I ) , consistent with spectroscopy data suggesting that catechol is unable to disrupt the ionic lock [9] .
Understanding the molecular mechanisms underlying GPCR functional selectivity is extremely important in modern drug discovery , since it provides a unique opportunity for the identification or rational design of ‘biased’ ligands as novel more effective therapeutics . Epitomizing an emerging paradigm in current drug discovery [54] , native states of GPCRs can be assumed in a dynamic equilibrium between different conformational sub-states [11] , [18] , [55] , which correspond to the valleys of an energy landscape , the barriers of which reflect the timescales of the conformational exchange . The relative populations of these sub-states follow statistical thermodynamics distributions and are shifted towards specific conformations as a consequence of ligand binding and/or other allosteric events such as those induced by protein-protein interactions . Thus , ligands with varied efficacies are believed to modulate the free-energy landscape of a GPCR , shifting the conformational equilibrium towards active or inactive conformations of the receptor , depending on their pharmacological action . A reliable characterization of the specific conformations that inverse agonists , agonists ( both full and partial ) , or antagonists can stabilize in a given GPCR is highly desirable for the structure-based discovery of novel ligands eliciting selected functional responses . This is difficult to achieve by X-ray crystallography for the majority of GPCRs due to their intrinsic structural instability , and the realization that the majority of pharmacologically active ligands are not ideal compounds for receptor stabilization that is suitable for crystallization . The enhanced sampling approach we describe here provides atomic-resolution information of receptor conformations along pre-determined activation pathways that are differentially stabilized by ligands with different efficacies . Our approach also provides a quantitative description of the thermodynamics of the B2AR basal activity , with the unliganded receptor being able to sample both an inactive state and an intermediate state that is shifted towards the activated conformation . This latter state is structurally different from the fully active state of B2AR captured by the nanobody-stabilized crystal structure . Although it exhibits a broken ionic lock and a cytoplasmic opening that is able to accommodate the camelid antibody , a few clashes are produced by the much smaller outward movement of TM6 ( ∼2 . 5 Å compared to the ∼5 . 9 Å that can be achieved by a full agonist ) . Given the small free-energy difference between the two lowest energy minima identified for the unliganded B2AR , these two states are almost equally populated at equilibrium , in agreement with the high basal activity of the B2AR . Moreover , the relatively low energy barrier between the two states is consistent with the flexible nature of the unliganded B2AR , and the consequent difficulty in obtaining crystals of the native receptor . We observed a more or less pronounced perturbation of the free-energy profile of the unliganded B2AR depending on the ligand considered for binding . Although alprenolol has often been described as a neutral antagonist of B2AR , its presence in the B2AR binding pocket slightly modifies the free-energy profile of the receptor , making the inactive state more stable in spite of the small difference in free-energy ( ∼kBT ) . This result is not completely surprising in light of the evidence existing in the literature for a role of alprenolol as an inverse agonist or even a weak agonist , depending on the assay used 56 , 57 . Our results show that the selection of a single conformational state is particularly effective in the case of inverse agonists . The docking of either carazolol or ICI-118 , 551 in the receptor dramatically changes the free-energy landscape of B2AR and reduces it to a funneled profile with a single major basin corresponding to the inactive conformation . This result is consistent with the greater availability of crystals of B2AR in an inactive conformation stabilized by potent inverse agonists in the binding pocket , and with the observation that the structural features of the inactive states of the various receptors obtained so far are similar . The situation is different when we study the free-energy landscape in the presence of agonists . The computational experiment with epinephrine shows that a full agonist is capable of stabilizing a state of B2AR presenting structural features that have been found in the nanobody-stabilized agonist-bound crystal structure of B2AR . However , in addition to this active state , we obtain a relatively stable agonist-bound inactive state that is structurally similar to the inverse agonist-bound crystal structure of B2AR . This is not surprising , given the absence of TM6 outward movements noted in both the B2AR crystal structure with a covalently-bound agonist [11] , and the agonist-bound B1AR crystal structures [13] . Moreover , the relatively small difference in free-energy between the fully active and the inactive agonist-bound conformations is probably due to the lack of the G-protein in the simulation setup , in line with the observation deriving from the two recent agonist-bound B2AR crystal structures [10] , [11] that a ligand alone is not sufficient to stabilize a fully active crystallographic state of the receptor , but a G-protein mimicking nanobody is necessary to trap this conformation . Different from the crystallographic information , but in line with experimental evidence from fluorescence spectroscopy [9] , we find that metastable states corresponding to fully ( and partial ) activated conformations of the receptor favor the rotamer change of the W6 . 48 side chain . The partial agonism elicited by dopamine and catechol shifts the conformational equilibrium towards states that are different from that stabilized by the full agonist , and captured in the nanobody-stabilized crystal structure . In particular , the two ligands affect the free-energy landscape in different ways . While the intermediate dopamine-bound state always features a broken ionic lock , the receptor samples conformations that have different ionic lock states when catechol is in the binding pocket . Notably , experimental evidence from fluorescence spectroscopy [9] also suggested that the very weak partial agonist catechol is not able to completely disrupt the interaction between the charged residues at the cytoplasmic end of TM3 and TM6 . Structurally , the two conformations stabilized by catechol and dopamine are different in the degree of separation between the extracellular ends of TM5 and TM6 and between the intercellular ends of TM3 and TM6 . Consistent with the hypothesis that global structural features of the receptor , such as the tilt of the extracellular half of TM5 , can optimize the binding to agonists [58] , we see a larger TM5 tilt in the presence of dopamine ( as well as for epinephrine ) and a smaller one in the presence of catechol . Owing to the greater ability of catechol to stabilize a state with a formed ionic lock , the intracellular ends of TM3 and TM6 also appear slightly closer ( by ∼1 Å ) together . In summary , we have designed a strategy using a combination of different adaptive biasing techniques that enables characterization of reliable ligand-specific conformations as demonstrated here in the case of B2AR . The strategy is completely general and may be of practical use for the structure-based design of ‘biased’ ligands that selectively activate signaling pathways , and may therefore exhibit improved therapeutic properties . | G protein-coupled receptors ( GPCRs ) constitute one of the most important classes of cellular targets owing to their known response to a host of extracellular stimuli , and consequent involvement in numerous vital biological processes . Compelling evidence herein referred to as ‘functional selectivity’ shows that ligands with varied efficacies can stabilize different GPCR conformations that may selectively interact with different intracellular proteins , and therefore induce different biological responses . Understanding how this selectivity is achieved may lead to the discovery of drugs with improved therapeutic properties . We propose here a computational strategy that enables identification of the specific conformations assumed by a GPCR when interacting with ligands that elicit different physiological responses . Not only can these computational models help bridge the information gap in structural biology of GPCRs , but they can be used for virtual screening , and possibly lead to the structure-based rational discovery of novel ‘biased’ ligands that are capable of selectively activating one cellular signaling pathway over another . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"biology",
"computational",
"biology",
"biophysics"
] | 2011 | Ligand-Induced Modulation of the Free-Energy Landscape of G Protein-Coupled Receptors Explored by Adaptive Biasing Techniques |
Gene duplication is supposed to be the major source for genetic innovations . However , how a new duplicate gene acquires functions by integrating into a pathway and results in adaptively important phenotypes has remained largely unknown . Here , we investigated the biological roles and the underlying molecular mechanism of the young kep1 gene family in the Drosophila melanogaster species subgroup to understand the origin and evolution of new genes with new functions . Sequence and expression analysis demonstrates that one of the new duplicates , nsr ( novel spermatogenesis regulator ) , exhibits positive selection signals and novel subcellular localization pattern . Targeted mutagenesis and whole-transcriptome sequencing analysis provide evidence that nsr is required for male reproduction associated with sperm individualization , coiling , and structural integrity of the sperm axoneme via regulation of several Y chromosome fertility genes post-transcriptionally . The absence of nsr-like expression pattern and the presence of the corresponding cis-regulatory elements of the parental gene kep1 in the pre-duplication species Drosophila yakuba indicate that kep1 might not be ancestrally required for male functions and that nsr possibly has experienced the neofunctionalization process , facilitated by changes of trans-regulatory repertories . These findings not only present a comprehensive picture about the evolution of a new duplicate gene but also show that recently originated duplicate genes can acquire multiple biological roles and establish novel functional pathways by regulating essential genes .
Gene duplication is a fundamental evolutionary process and provides a major source for genetic novelties [1]–[3] . The usual fate of a gene duplicate is pseudogenization , but some duplicates can fortuitously survive through neofunctionalization , in which one copy retains its ancestral function while the other copy acquires a novel function , or subfunctionalization , in which the duplicate and the ancestral copies subdivide the ancestral functions [4] , [5] . The two processes , especially neofunctionalization , should have contributed greatly to the biological diversity by providing genetic innovations . However , how a new duplicate gene acquires functions by integrating into a pathway and results in adaptively important phenotypes has remained largely unknown . Studying the recently originated young genes could be a very informative way to illustrate these processes , as genes at the early stage of evolution should have retained their original features well , which could have changed with time [3] . Currently , a number of young duplicate genes with potential biological functions have been reported [6]–[13] . Among them , three young Drosophila duplicate genes , arisen by retroposition , were reported to have male-related functions: K81 was proposed to be a testes-expressed paternal effect gene [6] , mojoless is required for male germline survival [7] , and sphinx is an RNA-coding gene responsible for male courtship behavior [8] , [14] . Nevertheless , little is known about how these young duplicate genes have been integrated into the molecular pathways and thereby have realized their functions in the host species . In this study , we systematically characterized a young Drosophila gene of the kep1 gene family , which originated recently in the Drosophila melanogaster ( D . melanogaster ) species complex ( including D . melanogaster , D . simulans , D . mauritiana , and D . sechellia ) about 5 . 4–12 . 8 million years ago through the duplications of the kep1 gene locus , mediated by the transposon DNAREP1_DM [15] . We performed a comprehensive investigation of its functions within an evolutionary context and successfully revealed its biological roles as well as the underlying molecular mechanism . The results shed novel light on the functional origin of new genes at the pathway level .
There are 7 members in the kep1 gene family , and their phylogenetic distributions are illustrated in Figure 1A . The parental gene kep1 is present in all Drosophila species . Through the duplications of the kep1 gene locus , the new genes nsr ( novel spermatogenesis regulator , CG3875 ) , CG3927 , CR9337 , and CG4021 originated in the common ancestor of the D . melanogaster species complex , and CR9337-r and CR33318 occurred after the sibling species in the complex diverged [15] . In this study , we focused on the intact new duplicates nsr , CG3927 , and CG4021 in D . melanogaster , in which the genetic manipulations are feasible . The kep1 family copies are located dispersedly on the second chromosome . D . melanogaster kep1 is a pre-mRNA splicing factor , influencing female fertility , eye development , and immune responses to bacterial infection [16] . Consistent with that , the coding sequences of kep1 are conserved throughout the Drosophila phylogeny ( Table S1 ) . Multiple alignments of the protein sequences of kep1 family members show that the three intact new genes have a well-retained KH RNA-binding domain but possess highly diverged C-termini ( Figure 1B ) . By sliding window analysis , the ratio of nonsynonymous changes ( dN ) over synonymous changes ( dS ) for each kep1-new gene pair was estimated and tested for selection . For all gene pairs , significant purifying selection signals are enriched in the KH domain region ( Figure 1C and Figure S1A ) , revealing functional constraint on the new genes . Most interestingly , the C-termini between the kep1-nsr pair shows significant positive selection signal ( dN/dS = 6 . 11 , p-value <0 . 05 ) ( Figure 1C ) , which probably arose from accelerated evolution in the nsr as a result of adaptive evolution . We analyzed the evolutionary patterns along the phylogenetic branches for nsr ( Figure 1D ) , CG3927 , and CG4021 ( Figure S1B ) , based on the maximum likelihood estimates of ω values ( dN/dS ) [17] . If we assume that the duplication events happened when D . melanogaster and D . yakuba diverged 7 . 4 million years ago [18] , even using the most conservative estimate of the synonymous substitution rate for Drosophila [19]–[21] , 24 . 3 , 17 . 9 , and 22 . 6 synonymous substitutions are expected to occur in the ancestral lineage of the D . melanogaster species complex for nsr , CG3927 , and CG4021 , respectively . These numbers are far beyond our observations , which are 2 . 6 for nsr , 0 for CG3927 , and 9 . 3 for CG4021 ( Figure 1D and Figure S1B ) . Therefore , the three new duplicate copies must have originated very late in the ancestral lineage , probably close to the split point of the sibling species in the D . melanogaster species complex . In the ancestral lineage , there are many nonsynonymous substitutions in the new genes , and the estimated ω values are 3 . 192 for nsr ( Figure 1D ) , infinite for CG3927 ( there are no synonymous mutations ) , and 1 . 149 for CG4021 ( Figure S1B ) , in which the ones for nsr and CG3927 are significantly larger than the neutral expectation ( Table S2 ) , indicating that positive selection should have shaped the two new genes , especially nsr . On the branches leading to individual species , the ω values decline , possibly because the new genes might have evolved functions that are under selective constraint . In D . melanogaster , the kep1 copy is ubiquitously expressed [22] , but the new duplicate copies display a male-specific expression pattern , according to our RT-PCR results ( Figure S2A ) . To provide clues for the biological functions of new kep1 family genes , GFP was fused to the coding sequences of each gene to designate their detailed expression patterns in D . melanogaster ( Figure S2B ) . Since the uniform male-specific expression pattern for all of the new duplicate genes is more likely a consequence of a shared regulatory region rather than independently evolved genetic mutations , we used the homologous upstream regulatory sequences of all kep1 family genes as the driving promoter ( Figure S2D ) . As expected , the shared regulatory region is sufficient to drive similar male-specific expression for each of the GFP-tagged kep1 family proteins , which are unexceptionally enriched in the primary spermatocytes of testes ( Figure 2A–2D ) . Previous large-scale profiling of gene expression patterns in D . melanogaster testes demonstrated that all kep1 family genes showed a high level of mitosis and meiosis expression , followed by much-reduced post-meiosis expression [23] . This result is consistent with our observation and also suggests that the kep1 family genes may be expressed in the spermatogonial stage as well . In the primary spermatocytes , kep1 family proteins are localized in a specked nuclear pattern ( Figure 2E–2H ) , a highly diagnostic feature for spliceosomal components [24] , [25] . Considering that D . melanogaster kep1 is a splicing factor responsible for the alternative splicing of the Drosophila caspase molecule dredd [16] , the observation above led us to speculate that new kep1 family genes might regulate the pre-mRNA processing of genes required for spermatogenesis and sperm function . Evolution of novel subcellular localization after duplication is thought to be an important evolutionary mechanism for the origins of genes with novel functions [26] . Though both are distributed in punctuate nuclear structures of primary spermatocytes , the localization of Nsr protein is much broader than the Kep1 protein ( Figure 2I ) . RNase A treatment of testes could lead to the ectopic accumulation and dispersal of GFP-tagged Nsr protein ( Figure S2E , S2F , S2G , 2H ) , indicating that the Nsr protein is localized in an RNA-dependent manner , and its expanded nuclear localization might imply a novel RNA-binding property . CG4021 protein is localized , completely overlapping with the Kep1 protein , in primary spermatocyte nuclei ( Figure 2J ) , and CG3927 protein was found to have a lack of a significant fluorescent signal for the comparison . To comprehensively understand the biological functions of the kep1 family genes , we have generated null mutants for all four intact gene copies in D . melanogaster by either gene targeting knockout [27] or imprecise P-element excision [28] ( Figure 3A and 3B ) . The wild-type ( WT ) control flies of the mutants are WT recombinants created by targeted mutagenesis or precisely excised strains of P-element excision , for the sake of an identical genetic background between the mutant and the WT flies . The null males of nsr display significantly reduced fecundity when compared to the WT males ( p-value <0 . 001 , Mann-Whitney U test ) ( Figure 3C ) . This phenotype can be fully restored by introducing the genomic sequences of nsr back into the genome ( Figure 3C ) . Heterozygous flies of nsr mutants are equally fertile as the WT flies ( Figure 3C ) . We found that the sperm storage tissue ( seminal vesicle ) of nsr male mutants was empty or contained little sperm , if any ( Figure 4A and 4B ) . During D . melanogaster spermatogenesis , germ cells from gonial precursors differentiate into cysts of 64 syncytial spermatids , which will undergo an actin-based individualization process , in which a bulk of unneeded cytoplasm is eliminated from the spermatids through remodeling of the cyst membrane . Extrusion of the cytoplasm along sperm bundles can form visible cystic bulges , which will migrate to the distal ends and are detached as waste bags . An actin structure , termed the “investment cone ( IC ) , ” is formed at the site where each spermatid develops its own membrane [29] , [30] . We labeled the sperm bundles together with the cystic bulges and waste bags with GFP under control of the don juan ( dj ) gene promoter [31] , and the ICs are visualized by FITC-conjugated phalloidin . The testis of nsr mutant male contains comparable amounts of spermatids as their WT controls; however , the structures of cystic bulges and waste bags are largely absent ( Figure 4C and 4D ) . In WT flies , ICs in the same cyst move coordinately in clusters ( Figure 4E ) , while they are scattered along the sperm bundles in the nsr mutants ( Figure 4F ) . The phenotypes above are typical features of an impaired individualization process [30] . Electron microscopy examination further confirmed that the spermatids of nsr mutants are unindividualized , with substantial amounts of residual cytoplasm ( Figure 4G and 4H ) . As the final step of spermatogenesis , the spermatids are assembled by coiling at the base of the testis to facilitate their transport into seminal vesicles [29] . Under a phase-contrast microscope , the sperm bundles of nsr mutants are twisted at the distal ends of testis , instead of regular coiling ( Figure 4I and 4J ) . Therefore , nsr is functionally involved in both sperm individualization and coiling . In contrast , though kep1 is required for female fertility in D . melanogaster [16] , no significant difference in male fertility was detected between kep1 mutant males and their WT controls ( Figure 3D ) . Also , we did not observe reduced fertility ( Figure 3D ) or other obvious defects for the CG3927 and CG4021 mutants . Considering that only nsr exhibits a robust signature of positive selection , this result may not be surprising . Either CG3927 and CG4021 have not acquired new functions or their phenotypic effects are not strong enough to be detected in our phenotyping assay . Microarray comparison of the transcription profiles between nsr WT and mutant testes only identified 14 genes that exhibited at least a 2-fold difference at the expression level , but none of them seemed to be male fertility-related ( Table S3 ) . Considering that the background hybridization noise and lack of probes for some genes might limit the power of microarray , we further implemented whole transcriptome shortgun sequencing ( RNA-Seq ) , which is regarded as a more precise way for measurements of transcript levels [32] . Using the Illumina paired-end sequencing platform , we generated 16 . 3 million reads ( 75-bp ) for WT testes and 9 . 6 million for nsr mutant testes . Based on these transcriptome data , we identified 10 genes that were significantly differentially expressed ( >5-fold ) between WT and mutants . Among them , kl-2 , kl-3 , and kl-5 are known male fertility genes , and the others are either not correlated with male fertility or functionally unknown ( Table S4 ) . The kl-2 , kl-3 , and kl-5 genes are 12 . 4-fold , 10 . 0-fold , and 6 . 84-fold down-regulated in the mutants , respectively ( Table S4 ) , and their sharp reductions in expression were validated by real-time PCR ( Figure 5A ) . Interestingly , the three genes were located adjacently on the Y chromosome , and all encode dynein heavy chain polypeptides of the sperm axoneme [33]–[35] . The phenotypic defect associated with the sterility of kl-2 mutants is not very clear [36] , [37] , while kl-3 or kl-5 mutations by P-element insertions result in loss of the outer dynein arm of the sperm axoneme and irregular coiling of spermatid tails , and complete deletion of either locus causes defects in sperm individualization [37]–[39] . Electron microscope examination of the spermatid flagellum showed that the outer dynein arms of sperm axonemes were also missing in the nsr mutants ( Figure 5B–5F ) . The deficiencies of nsr mutants , including sperm individualization , coiling , and axonemal structures , fit well with the phenotypes of the kl-3 and kl-5 mutants . This substantial agreement of the loss-of-function phenotypes between the Y-linked genes kl-3 , kl-5 , and nsr indicates that nsr is involved in male functions by regulating kl-3 , kl-5 , and , possibly , kl-2 as well . Moreover , it is very likely that nsr regulates the kl-2 , kl-3 , and kl-5 genes post-transcriptionally , because their primary transcript levels are largely unaltered between the mutants and WT flies , as shown by real-time PCR results ( Figure 5A ) . This is also in accordance with the conserved RNA-binding domain ( Figure 1B and 1C ) and the splicing factor-like distribution pattern ( Figure 2F ) of the Nsr protein . More importantly , our co-immunoprecipitation experiment demonstrated that the pre-mRNA cleavage stimulatory factor CstF-64 [40] can be specifically immunoprecipitated by TAP-tagged Nsr protein from testis extracts ( Figure S3A and S3B ) . This result fortifies the idea that nsr might function as an RNA processing factor , although future studies are needed to explore how nsr and CstF-64 collaboratively process the primary transcripts of these male genes . We traced the functional status of kep1 in the pre-duplication species D . yakuba by detecting its expression pattern using Kep1 antibody ( Figure S3C ) . Surprisingly , immunocytochemistry with Kep1 antibody showed only background staining of D . yakuba testis ( Figure 6D ) , whereas it is capable of yielding a robust staining pattern in the primary spermatocytes of D . melanogaster ( Figure 6B and 6C ) , exactly as revealed by transgenic GFP localization ( Figure 2A ) . The antibody worked well in detecting Kep1 proteins in ovary extracts from both D . yakuba and D . melanogaster by Western blot ( Figure S3D ) , ruling out the possibility that the antibody sensitivity is not equally sufficient for detecting Kep1 protein of D . yakuba . Absence of Kep1 protein in D . yakuba testis suggests that the kep1 gene should not be ancestrally required for male fertility , and it also raises the questions of when and how the novel testicular expression patterns of the kep1 family in D . melanogaster has been evolved . The immunofluorescent signals of Kep1 proteins in the sibling species of D . melanogaster , D . simulans ( Figure 6E ) and D . sechellia ( Figure 6F ) , suggest that this novel pattern has been established in the common ancestor of the D . melanogaster species complex . This interspecies difference of expression pattern between D . yakuba and D . melanogaster may arise from either cis-acting or trans-acting regulatory changes . The two genetic factors can be distinguished by testing the transcriptional activity of D . yakuba's cis-elements of kep1 in D . melanogaster . Controlled by D . yakuba's cis-elements of kep1 , GFP was also found to accumulate in the primary spermatocytes in D . melanogaster ( Figure 6G ) with the same subcellular localization as with the control of the cis-elements of D . melanogaster kep1 ( Figure S3E ) . This means that the activity of the cis-elements has not been differentiated between D . yakuba and D . melanogaster , and it is the changes in trans-regulatory repertoires that most likely have enabled all kep1 family genes to obtain novel testicular expression patterns .
There are two possible scenarios to explain the current functional roles of nsr in D . melanogaster: neofunctionalization and subfunctionalization [4] , [5] . Our results tend to support the neofunctionalization scenario , although we cannot completely exclude the possibility of subfunctionalization . Several pieces of evidence support the neofunctionalization scenario . Firstly , the parental gene kep1 is under strict purifying selection across the Drosophila phylogeny ( Table S1 ) . The significant conservation of kep1 and its inessentiality for male fertility in the pre-duplication species D . yakuba is consistent with the reported functions of kep1 in female fertility , eye development , and immune response [16] but not male fertility ( Figure 3D ) in D . melanogaster . These results suggest that kep1 possibly has retained its ancestral functions without evolving novel male functions after the duplication events , and nsr is free to evolve new functions . Secondly , nsr shows a robust signal of positive selection ( Figure 1C and 1D ) , especially in the C-termini ( Figure 1C ) . As we know , RNA recognition is a complex biological process that may need the collaboration of multiple factors; the RNA-binding domain alone possibly does not contain sufficient information for specific targeting [41] , [42] . Thus , the rapidly evolving C-termini of nsr could have contributed to novel RNA-binding ability by mediating co-option with different cofactors , and this idea is further strengthened by the specific immunoprecipitation of the pre-mRNA cleavage stimulatory factor CstF-64 by the Nsr protein ( Figure S3B ) . The subcellular localization pattern of the Nsr protein is also different from the Kep1 protein by displaying a larger localization range in the nuclei of primary spermatocytes ( Figure 2I ) , and cell type-specific expression or subcellular localization is regarded as one of the strategies for RNA-binding proteins to regulate specific splicing events [42] . Although it is still not clear what is the concrete molecular process that the novel distribution pattern of nsr has contributed to its roles in spermatogenesis , it is possible that this novel distribution might allow the spatial-specific assembling between nsr and its cofactors , and the subsequent specific regulation of mRNA substrates . Thirdly , our antibody did not detect obvious expression of Kep1 protein in D . yakuba testis , and thus , the parental gene kep1 should not be ancestrally required for male fertility . After the split of D . yakuba , trans-regulatory changes possibly occurred prior to or accompanied by the duplications of kep1 , which enabled the kep1 family genes to obtain novel testicular expression patterns and thereby lend them an opportunity to evolve novel male functions , as nsr has done . Nevertheless , the alternative subfunctionalization scenario cannot be completely excluded if a recent “gain and loss” turnover of male functions for kep1 did happen or if kep1 has lost its male functions in the D . yakuba lineage for some reason . In the recent “gain and loss” turnover , the parental gene kep1 could have acquired an essential role in spermatogenesis after the split of D . melanogaster and D . yakuba but prior to the duplication events , whereas the new copy nsr has taken over the spermatogenesis role from kep1 after its origination . The new duplicate gene nsr displays tremendous divergence from kep1 at the levels of biological function and molecular pathway . The kep1 gene participates in female fertility by regulating the apoptosis molecule dredd [16] , whereas the new gene nsr is integrated into the spermatogenesis pathway by regulating Y-linked male fertility genes; thus , our findings also provide an unusual case , showing a functional transition in a new gene from a female role to male role . It is interesting that the newly originated genes are often expressed primarily in male reproductive tissues in diverse organisms [43]–[47] , and most of the new Drosophila genes with known functions [6]–[8] , together with nsr , are associated with male reproduction . This phenomenon pronounces that new genes may tend to be functionally male-biased and suggests a significant role of natural selection and sexual selection in the fixation of beneficial mutations for male reproductive success . Our study reveals that nsr has been integrated into fundamental developmental processes by regulating pre-existing essential genes . Interestingly , the sperm maturation aspects that nsr participates in are conserved during evolution [48] . For example , the failure of eliminating sperm cytoplasm and loss of the outer axonemal dynein arm can also cause many types of human infertility [49]–[51] . The functional mechanism of nsr indicates that new genes could contribute to the evolutionary turnover of molecular pathways governing essential and conserved developmental processes , which partially explains the phenomenon that the same developmental processes in different organisms are sometimes achieved by a different set of genes . The positive selection signal and biological functions of nsr together strongly suggest that nsr might have contributed to the adaptive evolution of male reproductive pathways in the D . melanogaster species complex .
Protein sequences of nsr , CG3927 , CG4021 , and kep1 in D . melanogaster are downloaded from FlyBase ( http://flybase . org ) and aligned by ClustalW ( http://www . ebi . ac . uk/Tools/clustalw ) . Orthologous coding sequences of kep1 family genes in other Drosophila species ( http://flybase . org ) were predicted using a combination of BLAT ( http://genome . ucsc . edu ) and GeneWise ( http://www . ebi . ac . uk/Tools/Wise2 ) and manually checked . Alignments of coding sequences mentioned below are performed by MEGA 3 . 2 [52] , considering the coding structures . To estimate the selective constraint on kep1 through the Drosophila phylogeny , alignments of kep1 coding sequences from different Drosophila species were tested for purifying selection by MEGA 3 . 2 pairwisely . To detect the selective pressure on the new genes of the kep1 family , alignments of the coding sequences between kep1 and each new gene were performed and calculated for the dN/dS ratio with 120-bp windows and 6-bp slides . For each window , the maximum likelihood method [53] was used to test if the dN/dS ratio was significantly different from one ( two-tailed Fisher's exact test ) . The ω ( dN/dS ) values in the phylogeny of new kep1 family genes were estimated using the maximum likelihood approach , implemented by the codeml free-ratio model in the PAML4 . 2 package ( http://abacus . gene . ucl . ac . uk/software/paml . html ) [17] . To test if the ω ratio in the ancestral lineage of the D . melanogaster species complex was significantly different from one , the likelihood of the two-ratio model with an estimated ω was compared to an alternative two-ratio model , with ω constrained to be one for this lineage . All Drosophila strains were maintained at 25°C using standard cornmeal medium . The transgenic strains were produced by microinjection of w1118 embryos following standard P-element-mediated germline transformation [54] . P-element insertion stocks DG20303 and KG07486 were ordered from Bloomington Stock Center . Strains for P-element excision ( Sp/CyO; Δ2-3 , Sb/TM6B and Sp/CyO; MKRS/TM6B ) are kindly provided by Dr . Yongqing Zhang . Strains for targeted mutagenesis ( 70FLP70I-SceI , 70FLP and 70I-CreI ) were previously described by Xie and Golic ( 2004 ) . For GFP-tagged vectors , the pH-Stinger plasmid [55] was modified by excision with SpeI/NheI and re-ligation to remove its Hsp70 promoter and nuclear GFP . Gene promoter sequences ( plus 5′ UTR ) and GFP coding sequences were then cloned into XbaI/EcoRI and EcoRI/KpnI sites of the modified plasmid . Coding sequences of each gene were added into EcoRI sites and selected for correct insertion orientation ( Figure S2B ) . TAP-tagged transgenic vectors were constructed similarly but had GFP replaced with a TAP tag , which consists of two IgG-binding domains of protein A ( ProtA ) and a calmodulin-binding peptide ( CBP ) separated by a TEV protease cleavage site [56] ( Figure S2C ) . For all the vectors above , a homologous upstream region of kep1 family genes ( including D . yakuba kep1 ) was adopted as the promoter sequence ( Figure S2D ) . A rescue construct of nsr was prepared by inserting a 2 . 8-kb DNA fragment , ranging from the end of the upstream gene to the start of the downstream gene , into the NotI site of the pW8 transformation vector ( FlyBase ) . The primer information is available in Table S5 . P-element excision: The fly strains DG20303 ( with a P-element at the 5′ UTR of kep1 ) and KG07486 ( with a P-element annotated to locate at the nsr locus but found to be inserted at the 5′ UTR of CG3927 after PCR validation ) were mobilized with Δ2-3 transposase by standard P-element excision , respectively [28] . Excision lines were screened by PCR , and the endpoints were determined by sequencing . Gene knockout by ends-in targeting: The targeting vectors were designed to create a deletion , spanning from 42-bp downstream of the transcriptional start site to a site within the 3′ UTR of nsr , and a deletion spanning from the start codon to a site within the 3′ UTR of CG4021 , respectively . Targeted mutagenesis was performed as previously described [27] . Donor flies bearing the targeting vector were generated and crossed with flies carrying heat shock-activated FLP recombinase and I-SceI endonuclease ( 70FLP70I-SceI ) . The 0-3 day old progeny were heat-shocked at 38 . 5°C for 1 hour , and the enclosed white-eye virgins were crossed with males constitutively expressing FLP recombinase ( 70FLP ) . In total , at least 1000 vials were screened for nonmosaic red-eye individuals with successful insertions of the targeted allele at the site of the endogenous allele . Stocks of the recombinant flies were established and crossed with flies carrying heat shock-activated I-CreI endonuclease ( 70I-CreI ) . We heat-shocked 0-3 day old progeny at 38 . 5°C for 1 hour and screened for white-eye adults with recombinant reduction events at the targeted site . The reduction events will lead to either removal of the allele or maintenance of the WT allele . Strains of both genotypes were established to serve as knockout and WT lines , respectively . For the male fertility test , an individual male of each genotype ( <1 day ) was placed with three w1118 virgin females , which were collected within 5 hours of enclosure and aged for 2 days . The progeny were counted on the 18th day after the mating and compared between the mutant and WT lines using Mann-Whitney U test . A polyclonal antibody was raised against the glutathione-S-transferase-Kep1 ( amino acids 233–313 ) recombinant protein in guinea pigs . Testis squashes and immunostaining were performed as previously described [57] . The primary antibodies used are guinea pig anti-Kep1 serum ( 1∶200 dilution ) for Kep1 protein and rabbit peroxidase-antiperoxidase complex ( PAP ) ( 1∶1000 dilution , Sigma ) for ProtA . The secondary antibodies are Alexa 555-conjugated anti-guinea pig and Alexa 594-conjugated anti-rabbit ( Molecular Probes ) . Testes were co-stained with Hoechst 33342 ( 1 µg/ml , Molecular Probes ) to visualize nuclear DNA if needed . FITC-conjugated phalloidin ( 1∶100 dilution ) was used for IC staining . RNase A treatment was performed as previously described [58] by a 10-min incubation of TBS with 50 µg/ml RNase A ( Fermentas ) , and the controls were incubated in the same buffer , but free of RNase A . For sample preparation , adult testes or ovaries from 0–5 day old flies were dissected in PBS , transferred to RIPA buffer , ground , and boiled at 95°C for 10 min for lysis . The primary antibodies used were PAP ( 1∶2000 dilution , Sigma ) , mouse anti-β-actin ( 1∶3000 dilution , Abcam ) , and guinea pig anti-Kep1 ( 1∶500 dilution ) . Peroxidase-conjugated secondary antibodies were used for signal detection ( 1∶10000 dilution , Santa Cruz ) . Six hundred testes of 0–3 day old flies carrying TAP-tagged Kep1 protein , TAP-tagged Nsr protein , or TAP-tagged CG4021 protein were used for affinity purification , respectively . Testes were ground in 100 µl RIPA buffer plus protease inhibitor cocktail ( Roche ) with the Sample Grinding Kit ( GE Healthcare ) . The cell suspension was centrifuged at 4°C for 5 min , the supernatant was pre-cleared by 5 µl protein G plus-agarose beads ( Santa Cruz ) , and incubated with 2 µl PAP at 4°C overnight . Then , 10 µl protein G plus-agarose beads were added to the mixture and incubated at 4°C for 1 hour . Complexes of TAP-tagged proteins were liberated from the beads by cleavage of TEV protease as previously described [56] , subjected to SDS-PAGE , and visualized by Coomassie blue staining . The protein band of interest was cut out and identified by MALDI-ToF mass spectrometry . The dissected testes from WT controls and nsr mutants were fixed in 2 . 5% glutaraldehyde , washed twice with PBS , post-fixed with OsO4 , and dehydrated in an ascending series of ethanol . The resultant specimens were embedded in Araldite , sliced into ultrathin sections ( 50–100 nm ) , stained with 1% uranyl acetate , and examined with a JEOL electron microscope . Total RNA was isolated from adult testes with Trizol reagent ( Invitrogen ) and treated with DNase I ( Fermentas ) . Reverse-transcription was performed using the RevertAid First Strand cDNA Synthesis kit ( Fermentas ) with a no-reverse-transcriptase reaction as the negative control . Real-time PCR was performed in triplicate with SYBR Green PCR Mix ( Bio-Rad ) and subjected to the ABI 7000 Sequence Detection System . Oligo-dT primer was used to synthesize the cDNA templates for detecting mature transcripts and random hexamer primer for primary transcripts . Information on the PCR primers is available in Table S5 . The relative concentration of genes was calculated by analyzing their dissociation curves using the constitutively expressed gene rp49 as the internal control . With Trizol reagent ( Invirtrogen ) , total RNA was extracted from testes of 0–2 day old nsr mutant and WT flies , respectively . After amplification , mRNA was fluorescently labeled by GeneChip One-Cycle Target Labeling ( Affymetrix ) and subjected to GeneChip Drosophila Genome 2 . 0 Arrays ( Affymetrix ) in duplicate . Image collection was performed in accordance with standard Affymetrix protocols . The significance of gene expression change was estimated using the Significance Analysis of Microarrays ( SAM ) algorithm , which assigns a score to each gene on the basis of change in gene expression relative to the standard deviation of repeated measurements [59] . The microarray data have been deposited in Gene Expression Omnibus ( GEO ) ( http://www . ncbi . nlm . nih . gov/geo ) under accession number GSE22289 . With Trizol reagent ( Invirtrogen ) , 5 µg total RNA was extracted from testes of 0-1 day old nsr mutant and WT flies , respectively . The first-strand cDNA was synthesized with oligo-dT primer by Superscripts II reverse transcriptase ( Invitrogen ) , and second strand cDNA synthesis was followed according to the standard protocol . Then , the double-stranded cDNA was purified with the Qiaquick PCR purification kit ( Qiagen ) and fragmented with a nebulizer ( Invirtrogen ) , resulting in an average size of 150–250-bp . Overhangs of resultant fragmented cDNAs were blunted with T4 DNA polymerase ( NEB ) and Klenow polymerase ( NEB ) and treated with 3′-5′ exonuclease-deficient Klenow polymerase ( NEB ) to generate 3′ overhangs . After that , cDNA was ligated to an Illumina PE adapter oligo mix by the Quick ligation kit ( Qiagen ) . The adapter-modified cDNA within 200-bp was isolated by agarose gel , extracted with the QIAquick Gel Extraction Kit ( NEB ) , and amplified by PCR reaction . Finally , the library products were sequenced using the Illumina GA2 sequencing machine . Sequence data from this study have been submitted to the NCBI Short Read Archive ( http://www . ncbi . nlm . nih . gov/Traces/sra/sra . cgi ) under accession number SRA020074 . The generated 75-bp raw reads were mapped to the genomic sequences of D . melanogaster ( Ensembl release 55: ftp://ftp . ensembl . org/pub/release-55/fasta/drosophila_melanogaster ) using SOAP2 software ( http://www . soapmaker . ca ) [60] . The count of covering reads for each annotated transcript ( Ensembl release 55: ftp://ftp . ensembl . org/pub/release-55/gtf/drosophila_melanogaster ) was calculated as the index of their expression level . The alteration of transcript level between nsr mutants and WT flies was estimated and normalized for the variation of the total data size of transcript reads . The significance of expression difference ( p-value ) for each gene ( the longest transcript ) was further computed according to the R package “DEGseq” using the MA-plot-based method with a random sampling model and followed by an adjustment with q-values for multiple testing corrections [61] . | Gene duplication has long been appreciated as a major source for new genes and new functions . Nevertheless , it is still a fascinating mystery how new duplicate genes are functionally integrated into the existing gene network and how they contribute to the novel functions of organisms at the pathway level . By studying the recently originated kep1 gene family in Drosophila melanogaster , we show that one of the young duplicate genes , nsr , has evolved important biological functions associated with male reproduction by regulating several essential fertility genes in the short evolutionary period after its birth . The evolutionary dynamics , biological roles , and the underlying molecular mechanism of nsr revealed in this study present a vivid and comprehensive example of how new genes acquire important biological functions and demonstrate that recently originated new genes can regulate pre-existing essential genes and create novel architectures of genetic pathways . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"evolutionary",
"biology",
"evolutionary",
"biology/morphogenesis",
"and",
"cell",
"biology",
"evolutionary",
"biology/developmental",
"evolution"
] | 2010 | A Young Drosophila Duplicate Gene Plays Essential Roles in Spermatogenesis by Regulating Several Y-Linked Male Fertility Genes |
To date , the contribution of disrupted potentially cis-regulatory conserved non-coding sequences ( CNCs ) to human disease is most likely underestimated , as no systematic screens for putative deleterious variations in CNCs have been conducted . As a model for monogenic disease we studied the involvement of genetic changes of CNCs in the cis-regulatory domain of FOXL2 in blepharophimosis syndrome ( BPES ) . Fifty-seven molecularly unsolved BPES patients underwent high-resolution copy number screening and targeted sequencing of CNCs . Apart from three larger distant deletions , a de novo deletion as small as 7 . 4 kb was found at 283 kb 5′ to FOXL2 . The deletion appeared to be triggered by an H-DNA-induced double-stranded break ( DSB ) . In addition , it disrupts a novel long non-coding RNA ( ncRNA ) PISRT1 and 8 CNCs . The regulatory potential of the deleted CNCs was substantiated by in vitro luciferase assays . Interestingly , Chromosome Conformation Capture ( 3C ) of a 625 kb region surrounding FOXL2 in expressing cellular systems revealed physical interactions of three upstream fragments and the FOXL2 core promoter . Importantly , one of these contains the 7 . 4 kb deleted fragment . Overall , this study revealed the smallest distant deletion causing monogenic disease and impacts upon the concept of mutation screening in human disease and developmental disorders in particular .
Many recent studies have provided insights into the biological relevance of the non protein-coding portion of the human genome , previously referred to as junk DNA . One of them is the ENCODE pilot study , which has revealed that the number of functional genomic elements is much higher than previously anticipated , and that the vast majority of elements regulating gene expression are contained in the non-protein coding portion of the genome . In addition , it shed light on the pervasively transcribed nature of the human genome [1] . Comparative analysis of genomes is a major tool for the identification of regulatory elements . In this context , several arbitrary criteria have been used to define evolutionarily conserved elements , such as conserved non-coding sequences ( CNCs ) that were originally defined as elements sharing ≥70% homology over ≥100 bp of ungapped alignment of human and mouse sequences [2]–[4] . A fraction of them ( i . e . the most conserved ones ) have been shown to function as cis-regulatory elements , predominantly controlling the spatiotemporal expression of developmental genes [5]–[7] . To date , the contribution of disrupted potentially regulatory CNCs to human genetic disease is most likely underestimated , as no systematic screens for putative deleterious variations in CNCs have been conducted in this respect . One of the reasons for this is the large extent of the regions to be investigated , as the regulatory domain of a gene can stretch beyond 1 Mb in both directions of its transcription unit . In addition , putative functional consequences of variations outside a transcription unit are difficult to assess . An example of a developmental gene with a strictly regulated spatiotemporal expression pattern is FOXL2 ( NM_023067 ) . It is known to be the disease-causing gene of blepharophimosis-ptosis-epicanthus inversus syndrome ( BPES ) [MIM 110100] , a rare autosomal dominant development disorder of the eyelids with ( BPES type I ) or without ( BPES type II ) premature ovarian failure ( POF ) [8] . Overall , sporadic and familial BPES can be explained by intragenic mutations and gene deletions in 71% and 11% of the patients respectively [9] . Interestingly , we identified microdeletions upstream and downstream of FOXL2 in 4% of BPES [9] , [10] . In addition , 3 translocation breakpoints upstream of FOXL2 have been described [8] , [11] , [12] . Until now , there is no evidence for genetic heterogeneity of this condition . From the 5 reported microdeletions outside FOXL2 , one is located 3′ to FOXL2 , while the others are located 5′ to FOXL2 and share a smallest region of deletion overlap ( SRO ) of 126 kb [10] . This SRO is located 230 kb upstream of FOXL2 , telomeric to the three previously characterized translocation breakpoints , and contains several CNCs , harbouring putative transcription factor binding sites . Moreover , the SRO contains the human orthologue of the Polled Intersex Syndrome ( PIS ) mutation in goat . The PIS goat is a natural animal model for BPES associating absence of horns ( polledness ) and intersexuality . The sex reversal exclusively affects female animals in a recessive manner , whereas the absence of horns is dominant in both sexes . The phenotype is caused by a regulatory 11 . 7 kb deletion located 280 kb upstream of goat FOXL2 . It was shown that the deletion alters the transcription of at least three genes: FOXL2 , the non-protein coding gene PISRT1 ( PIS-regulated transcript 1 ) ( AF404302 ) and PFOXic ( promoter FoxL2 inverse complementary ) ( AY648048 ) [13]–[15] . In agreement with the findings in the translocation patients and in the PIS goat , the distant microdeletions found in human BPES were hypothesized to disturb long-range transcriptional control of FOXL2 expression through the disruption of one or more cis-acting regulatory elements . These findings added to an increasing number of long-range genetic defects in human development conditions [16]–[19] . Apart from translocations and microdeletions/duplications of cis-regulatory elements , subtle copy number variations ( CNVs ) or sequence variations of cis-regulators can also be associated with a phenotype in humans . These have been found in the long-range limb-specific cis-regulatory element ZRS of the SHH gene ( NM_000193 ) , leading to preaxial polydactyly ( PPD ) ( PPD2 , MIM 174500 ) , isolated triphalangeal thumb ( MIM 174500 ) , and triphalangeal thumb-polysyndactyly ( TPTPS ) phenotypes ( MIM 174500 ) [20]–[23] . In addition , Benko et al . reported a heterozygous point mutation in a highly conserved non-coding conserved sequence located 1 . 44 Mb upstream of SOX9 in a patient with Pierre Robin sequence ( PRS , OMIM 261800 ) [19] . To date , the underlying molecular defect remains unknown in 12% of BPES patients [9] . Here , we focus on the contribution of previously unidentifiable and subtle deletions/duplications , and sequence variations in putative cis-regulatory elements surrounding FOXL2 in BPES . We developed a combined strategy consisting of microarray based comparative genome hybridization ( array CGH ) , high-resolution quantitative PCR ( qPCR ) and sequencing of CNCs located in the SRO 5′ to FOXL2 . Samples from 57 BPES patients who do not carry an intragenic FOXL2 mutation or gene deletion were studied , revealing a distant 7 . 4 kb deletion as the most prominent finding . The deletion harbours putative regulatory elements . Functional studies in cellular systems were performed to assess their regulatory potential . In addition , Chromosome Conformation Capture analysis ( 3C ) was conducted to provide insights into the spatial organisation and interaction patterns of a normal and a disrupted FOXL2 locus .
We identified a de novo distant 7 . 4-kb deletion that is causally related to BPES . To our knowledge , this is the smallest fully characterized distant deletion implicated in the causation of a human genetic condition ( Table S2 ) . This deletion disrupts a long ncRNA PISRT1 and 8 CNCs , 4 of which are conserved up to chicken . Functional assays suggest a cis-regulatory and tissue-specific potential of 3 of them . The biological relevance of these findings was corroborated by the 3C study of a normal and aberrant FOXL2 locus in expressing adult cellular systems respectively , demonstrating a close proximity of the 7 . 4 kb deleted fragment and two other conserved regions with the FOXL2 core promoter , and the necessity of the integrity of the regulatory domain for correct FOXL2 expression . Altogether , we identified and characterized a novel tissue-specific cis-regulatory domain of FOXL2 expression . As we demonstrated the consequences of its disruption , our findings impact mutation screening of strictly regulated developmental and other disease genes . Specifically , our study emphasizes the need for high-resolution copy number screening of their cis-regulatory domains . Genome-wide tools such as oligonucleotide or SNP arrays and next-generation sequencing will play a prominent role in this . In addition , a well-selected patient population is another requirement , as illustrated here: ( 1 ) we only included patients with a diagnosis of BPES , a clinically distinguishable but rare disorder , and ( 2 ) they all underwent a uniform pre-screening excluding intragenic FOXL2 mutations and gene deletions . Sequence variations within individual CNCs did not contribute to the molecular pathogenesis of BPES in our study . This can be explained by the fact that sequence changes within individual CNCs might result in a more subtle , different or even normal phenotype , as the cis-regulatory elements they represent might act in a tissue-specific and quantitative manner [5] , [6] , [19] , [33] . The most striking example of the latter is the differential phenotype caused by point mutations in SHH and in its limb-specific enhancer ZRS of SHH , resulting in holoprosencephaly type III ( HPE3 ) ( OMIM 142945 ) and PPD respectively [20] , [34] . Other mechanisms may explain the phenotype in the remaining 53 molecularly undefined BPES patients . Although there is no clear evidence for locus heterogeneity in BPES , mutations in other disease genes apart from FOXL2 cannot be excluded in some of the remaining molecularly unresolved cases . Another possibility is the occurrence of regulatory variants within the untranslated regions ( UTRs ) or the core promoter . A number of non-pathogenic sequence variants have been reported in the FOXL2 putative core promoter and untranslated regions ( UTRs ) up to now . However , a single basepair insertion in the FOXL2 3′UTR was found to co-segregate with BPES in a large Chinese type II BPES family , and was shown to be located in an AU rich repeat [35] . No functional studies were provided however to unequivocally prove a relationship between the insertion and the phenotype in this family . Interestingly , in the FOXP3 gene ( NM_014009 ) , encoding another forkhead transcription factor , a presumed disease-causing sequence change was found in the 3′UTR within the poly ( A ) signal , in affected members of a five-generation family with X-linked immune dysfunction , polyendocrinopathy , enteropathy ( IPEX ) ( MIM 304790 ) [36] . The occurrence of interesting pathogenic or modifying variants in 3′UTRs is in line with their important role in the regulation of gene expression at both pre-mRNA , mature mRNA and post-transcriptional level through cis-acting elements that interact with a variety of trans-acting factors [37] . This is highlighted by their many conserved sequence motifs , including microRNA ( miRNA ) targets [37] . It cannot be ruled out that changes in post-transcriptional regulation by altered miRNA targeting may result in BPES . A unique example of a variant that alters the gene expression level by modifying miRNA targeting activity is a 3′UTR SNP in human SLITRK1 ( NM_052910 ) , which is implicated in Tourette syndrome ( MIM 137580 ) [38] . Finally , this study considerably adds to the importance of an intact tissue-specific cis-regulatory domain in this and other developmental disorders . This impacts upon the concept of mutation screening of developmental disease in particular , and of human genetic disease in general . In the future , online databases such as Decipher and the Database of Genomic Variants which collect information on copy number changes , might help to interpret copy number changes affecting putative regulatory regions that might lead to disease [39] , [40] .
Genomic DNA ( gDNA ) from 57 consenting BPES patients without intragenic mutation or copy number change of the FOXL2 coding region was used in this study . Criteria described previously were used to accept a diagnosis of BPES [9] . The study was conducted following the tenets of Helsinki and was approved by the local Ethics Committee of the Ghent University Hospital . In order to detect hemizygous regions outside FOXL2 , microsatellite analysis was performed as described previously [10] . Microsatellite analysis was conducted for 19 molecularly unresolved patients for whom parental DNA was available . In order to detect copy number changes outside the transcription unit of FOXL2 , a new purpose-built bacterial artificial chromosome ( BAC ) array , consisting of 132 unique genomic clones covering a region of 3 Mb around FOXL2 and 95 control BACs ( 3 on each chromosome and 26 on the X chromosome ) , was designed in-house as previously described [41] , [42] . In total , 500 ng of DNA was labelled by a random prime labelling system ( BioPrime ArrayCGH genomic labelling system , Invitrogen ) using Cy3 and Cy5 labelled dCTPs ( Amersham Biosciences ) . Hybridizations were performed automatically using the HS400 hybridization station ( Tecan ) for 21 molecularly unresolved patients , of which 13 were previously screened by microsatellite analysis . The scan images were processed with Imagene software ( Biodiscovery ) and further analysed with arrayCGHbase [43] . Quantitative qPCR ( qPCR-3q23 ) was performed as described for a second group of patients as an alternative to arrayCGH , in order to detect copy number changes encompassing the initial SRO [44] . First , 3 qPCR amplicons located within the SRO 5′ to FOXL2 were designed and used to identify possible extragenic deletions overlapping the SRO in 24 molecularly unresolved patients , not previously screened by array CGH . Second , 10 additional in-house designed amplicons were used to further delineate 3 new extragenic deletions . All 13 amplicons were designed in silico as described ( primer sequences available upon request ) [44] . qPCR was carried out using the qPCR Core kit for SYBR Green I ( Eurogentec ) on the LightCycler 480 ( Roche ) . Calculation of the gene copy number was performed with qBase software [45] . Two reference genes , ZNF80 ( NM_007136 ) and GPR15 ( NM_005290 ) , were used for normalization of the relative quantities . A comparative analysis of the SRO region ( delineated by SNP rs10935309 and rs4894405 ) was performed by pairwise comparison of the human and mouse genomes . More specifically , the GALA genome browser implemented with hg16 build was used to identify all non-coding sequences of ≥100 bp and sharing ≥70% identity with the mouse [46] . The analysis resulted in the identification of 25 CNCs that are reproducibly mapped when implementing the hg17 build . Subsequently , using the multiZ alignment track in the UCSC genome browser , the conservation of all identified CNCs was examined in the genomes of placental mammals , chicken and pufferfish . In addition , the overlap of all the identified CNCs with previously reported PhastCons sequences was evaluated using the PhastCons conservation in the UCSC Genome Browser [47] , [48] . In order to detect subtle copy number changes within or nearby the identified CNCs specifically , 36 qPCR amplicons were designed within the initial SRO: 19/36 map within CNCs ( no successful assays were obtained for CNC10 , 14 , 15 , 21 , 23 and 25 ) while the 15 additional assays map within some long flanking regions . The latter amplicons were designed following CNC copy-number analysis in order to increase the screen resolution or to verify the mapping of putative copy-number variants . SYBR Green I qPCR-CNC was performed in 53 selected patients as described [49] . All amplicons were designed in silico using PrimerExpress ( Applied Biosystems ) ( primers available upon request ) and validated as described [49] . For the patient with deletion D long-range PCR was performed using the qPCR primers delineating the deletion . For long-range PCR the iProof high-fidelity PCR kit ( Biorad ) was used according to the manufacturer's instructions . In order to determine the junctions at base pair resolution , direct sequencing was performed on the 5 kb product using 8 internal sequencing primers ( available upon request ) ( ABI 3730xl Applied Biosystems ) . We used several web-based tools to unravel the mechanism by which the 7 . 4 kb deletion occurred . Genomic sequences of several sizes and centered on the breakpoints were obtained from the UCSC genome browser . First , CLUSTALW was used to align the junction sequence ( 70 bp ) with the reference genomic sequence from both the proximal and the distal breakpoint region [50] . Second , BLAST2 was run under default conditions to perform a pairwise sequence comparison of the 2 kb proximal and distal breakpoint regions [51] . Third , several programs ( RepeatMasker , Mreps , Palindrome , and Censor ) were employed to screen for repetitive elements/structures , low-complexity sequences , tandem and palindromic inverted repeats [52]–[54] . For sequence analysis with RepeatMasker and Censor we used the 2 kb breakpoint regions and for analysis with Mreps and Palindrome 300 bp regions . In addition , the fractional GC content of the breakpoint regions was calculated using GEECEE . DNA Pattern Find was applied to locate specific sequence motifs within the 70 bp breakpoint regions and the junction fragment [55] . The investigated specific sequences are known to be implicated in DNA rearrangements elsewhere [56] . Several tracks within the UCSC genome browser ( Genes and Gene Prediction Tracks , mRNA and EST Tracks and Regulation Tracks ) were used to screen the reduced SRO ( chr3:140 , 431 , 841-140 , 439 , 199 ) . In addition , the Ensembl regulatory features track was used to gain information about possible DNaseI hypersensitivity sites and CCCTC-binding factor ( CTCF ) binding sites . Several RNA databases ( RNAdb , miRDB , miRNAMap , miRBase and NONCODE v2 . 0 ) were consulted in order to extract possible non-coding RNA sequences [57]–[61] . Finally , BLASTn was run under default conditions to define the human orthologue of caprine PISRT1 ( AF404302 ) within this reduced SRO . In order to define the location of the human 7 . 4 kb deletion with respect to the deletion in the PIS goat , BLAST2 was performed for goat BAC 376H9 and a 100 kb extract from human chromosome 3 NT_005612 . 15 containing the reduced SRO ( 45 . 400 . 000–45 . 500 . 000 ) . Relative PISRT1 expression levels were determined in several human cell lines/tissues using real-time quantitative RT-PCR with newly designed primers ( available upon request ) . Primers were designed as described [44] . cDNA prepared from fibroblasts from a control individual and from human granulosa KGN cells ( Riken Institute ) and cDNA from testis ( human testis Marathon-Ready cDNA , Clontech ) were used for PISRT1 expression analysis . . RNA was isolated from fibroblasts and KGN cells as described ( RNeasy , Qiagen ) , and treated with RNase-free DNAse ( Promega ) , followed by cDNA synthesis as described ( iScript cDNA synthesis kit , Bio-Rad ) ; qPCR was carried out using the qPCR Core kit for SYBR Green I ( Eurogentec ) on the LightCycler 480 ( Roche ) as described above . PISRT1 expression levels were normalized using 3 housekeeping genes ( HPRT1 , GAPDH and YWHAZ ) ( NM_000194 , NM_002046 and NM_145690 ) . The obtained data were analyzed using qBase plus [45] . To characterize the full-length human PISRT1 transcript , 5′ rapid amplification of the cDNA ends ( 5′ RACE , Clontech ) was performed according to the manufacturer's protocol , using the Advantage cDNA PCR Kit and human testis Marathon-Ready cDNA ( Clontech ) as a template ( primers available upon request ) . For our novel human PISRT1 transcript , an accession number was requested at the GenBank ( accession number FJ617010 ) . Primers surrounding each of the 25 CNCs ( ±50 bp of the core CNC ) were designed with Primer3 ( primers available upon request ) [62] . A specific amplicon could be obtained for 24/25 CNCs , except for CNC19 . Sequence analysis of 24 CNCs was performed in 32 molecularly unresolved patients . In a second step , targeted sequencing of CNCs mapping within the reduced SRO defined by the 7 . 4 kb deletion , was performed in the remaining 21 patients . Sequence analysis of the first set of patients was performed with RedTaq ( Jumpstart kit , Sigma ) under standard touchdown PCR conditions . For the second set of patients new amplicons were designed for closely mapping CNCs instead of single CNC analysis . Thus , CNC5 , 15 , 6 , 16 , 4 and 14 were pooled as follows: CNC5-15 ( amplicon size: 573 bp ) , CNC6-16 ( amplicon size: 962 bp ) and CNC4-14 ( amplicon size: 1140 bp ) . In this case , PCR amplification was carried out with the iProof High-Fidelity DNA polymerase ( BioRad ) as indicated by the manufacturer . Each amplicon was directly sequenced in forward and reverse orientation using an ABI 3130 analyser ( Applied Biosystems ) . To align and identify nucleotide variants the Sequencher software ( Gene Codes Corporation ) was used . Multispecies alignments extracted from the UCSC Genome browser were used to evaluate the conserved nature of nucleotides presenting variants . Computational transcription factor binding site predictions were performed with the MATCH interface of the TRANSFAC database [63] , [64] . In vitro luciferase assays were performed in FOXL2 expressing KGN cells , and non-expressing 293T cells ( human kidney cells , ATCC ) . Wild-type ( WT ) and variant CNCs were directly PCR amplified from normal and affected genomic DNA respectively , with the same sets of primers and PCR conditions used for CNC sequencing , except for CNC1 . For CNC1 new primers were designed as described above based on a recent conservation pattern survey . The new CNC1 amplicon adds approximately 260 bp to the original one and covers the full conserved alignment that can be observed in UCSC and that overlaps with an extremely conserved sequence with highly regulatory potential [65] . Two types of luciferase constructs were produced: ( 1 ) pTAL-Luc CNC constructs , for which each PCR product was cloned into the TOPO-TA PCR II vector after amplification ( Invitrogen ) ; colonies with insert in reverse orientation ( i . e . 3′-5′ ) were specifically selected and sequenced . Subsequent subcloning into the pTAL-Luc vector ( Clontech ) expressing the firefly luciferase was achieved by SacI-XhoI digestion of both the TOPO-CNC constructs and pTAL-Luc vector ( Clontech ) . The amplicon encompassing CNC1 contained internal SacI and XhoI , and was subcloned using SpeI-BglII restriction sites . The fragment was subsequently cloned into a modified pTAL-Luc vector containing part of the multiple cloning site of TOPO-TA II . ( 2 ) pTAL-SV40 CNC constructs , for which the pTAL-Luc backbone was digested with BglII and HindIII in order to remove the minimal TATA-like promoter and replace it by a SV40 promoter . Subsequently , all pTAL-Luc CNCs were digested with SacI-XhoI ( SpeI-BglII for CNC1 ) and subcloned into a pTAL-SV40 ( Promega ) digested with similar enzymes . In both approaches , reverse-orientated CNC constructs were obtained . We specifically decided to investigate the regulatory potential of CNCs in their native orientation with respect to FOXL2 . Ensembl Genome Browser , http://www . ensembl . org/index . html GEECEE , http://mobyle . pasteur . fr/cgi-bin/MobylePortal/portal . py ? form=geecee GenBank ( MapViewer ) , http://www . ncbi . nlm . nih . gov/mapview/static/MVstart . html Online Mendelian Inheritance in Man ( OMIM ) , http://www . ncbi . nlm . nih . gov/omim/ Palindrome , http://bioweb . pasteur . fr/seqanal/interfaces/palindrome . html Repeatmasker , http://www . repeatmasker . org UCSC Genome browser , http://genome . ucsc . edu/ | Long-range genetic control is an inherent feature of genes harbouring a highly complex spatiotemporal expression pattern , requiring a combined action of multiple cis-regulatory elements such as promoters , enhancers , and silencers . Consequently , disruption of the long-range genetic control of a target gene by genomic rearrangements of regulatory elements may lead to aberrant gene transcription and disease . To date , the contribution of mutated regulatory elements to human disease has not been studied frequently . Here , we explored the contribution of genetic changes in potentially cis-regulatory elements of the FOXL2 gene in blepharophimosis syndrome ( BPES ) , a developmental monogenic condition of the eyelids and ovaries . We identified a de novo very subtle deletion of 7 . 4 kb causing BPES . Moreover , we studied the functional capacities and chromosome conformation of the deleted region in FOXL2 expressing cellular systems . Interestingly , the chromosome conformation analysis demonstrated the close proximity of the 7 . 4 kb deleted fragment and two other conserved regions with the FOXL2 core promoter , and the necessity of their integrity for correct FOXL2 expression . Finally , our study revealed the smallest distant deletion causing monogenic disease and emphasized the importance of mutation screening of cis-regulatory elements in human genetic disease . | [
"Abstract",
"Introduction",
"Results/Discussion",
"Materials",
"and",
"Methods"
] | [
"molecular",
"biology/chromatin",
"structure",
"computational",
"biology/comparative",
"sequence",
"analysis",
"genetics",
"and",
"genomics/genetics",
"of",
"disease",
"developmental",
"biology/molecular",
"development",
"ophthalmology/inherited",
"eye",
"disorders",
"genetics",
... | 2009 | Disease-Causing 7.4 kb Cis-Regulatory Deletion Disrupting Conserved Non-Coding Sequences and Their Interaction with the FOXL2 Promotor: Implications for Mutation Screening |
A correct assessment of the quaternary structure of proteins is a fundamental prerequisite to understanding their function , physico-chemical properties and mode of interaction with other proteins . Currently about 90% of structures in the Protein Data Bank are crystal structures , in which the correct quaternary structure is embedded in the crystal lattice among a number of crystal contacts . Computational methods are required to 1 ) classify all protein-protein contacts in crystal lattices as biologically relevant or crystal contacts and 2 ) provide an assessment of how the biologically relevant interfaces combine into a biological assembly . In our previous work we addressed the first problem with our EPPIC ( Evolutionary Protein Protein Interface Classifier ) method . Here , we present our solution to the second problem with a new method that combines the interface classification results with symmetry and topology considerations . The new algorithm enumerates all possible valid assemblies within the crystal using a graph representation of the lattice and predicts the most probable biological unit based on the pairwise interface scoring . Our method achieves 85% precision ( ranging from 76% to 90% for different oligomeric types ) on a new dataset of 1 , 481 biological assemblies with consensus of PDB annotations . Although almost the same precision is achieved by PISA , currently the most popular quaternary structure assignment method , we show that , due to the fundamentally different approach to the problem , the two methods are complementary and could be combined to improve biological assembly assignments . The software for the automatic assessment of protein assemblies ( EPPIC version 3 ) has been made available through a web server at http://www . eppic-web . org .
This article is dedicated to the memory of Guido Capitani , a dear friend and mentor to all of us .
The PDB defines the biological assembly as “the macromolecular assembly that has either been shown to be or is believed to be the functional form of the molecule” [13] . Many proteins do have a single clear functional unit which accounts for the majority of the folded species in cells . However , determining the biological assembly in crystals can be less clear-cut . Modifications in the protein construct to facilitate crystallization , such as removal of disordered loops or domains , can alter or remove interfaces , giving a different assembly than would be present in vivo . In such cases , rather than representing the functional form of the molecule , the best we can hope for is representing the complex that would remain were the crystal to be dissolved in a physiological-like buffer . Weak interactions represent a further challenge . Many protein-protein interactions can be described as transient or weak , as measured by a high dissociation constant ( Kd ) . The crystal environment may or may not capture those transient assemblies . EPPIC is targeted at predicting stable biological assemblies; in cases where the protein is likely to exist in equilibrium under physiological conditions , we consider both states to be correct biological assemblies and typically predict the smaller ( more stable ) assembly . However , this is not a major issue in practice , as all cases considered in the benchmark had a clear consensus as to the correct biological assembly . Let us first introduce a few definitions that will be used throughout the manuscript: In their seminal paper , Monod , Wyman and Changeux [14] exposed the basics of protein association into oligomers by presenting a very clear argumentation on the possible ways in which homomers can associate . They argue that only two types of associations are possible between two protein chains of the same entity: In isologous associations the interacting interface patches are mutually satisfied and capped . There is no further association possible through the interfaces . However in heterologous association the interacting interface patches are exposed to the solvent and will continue associating to other protomers indefinitely . The only way that this indefinite association can stop is by the protomers cycling around and associating back to the first protomer , forming a cyclic Cn symmetry . Thus in both cases , in order to have stable oligomeric complexes in solution , symmetry must occur . Specifically , point group symmetry is necessary: cyclic ( C ) , dihedral ( D ) , tetrahedral ( T ) , octahedral ( O ) , or icosahedral ( I ) . Cyclic is the only point group that is composed by only heterologous interfaces , while the others are combinations of both isologous and heterologous interfaces . The same argument can be extended to heteromers with two or more copies of each monomer . The heteromer is reduced to the homomer case by simply fusing the heteromeric entities into one and then treating the super entity as a homomer . Symmetry is thus a necessary condition for stable protein oligomers and we found our subsequent analysis and the assembly rules on that assumption . The necessity and prevalence of symmetry has been since widely studied in the literature . The review by Goodsell and Olson [15] is a comprehensive overview of the topic . There are mechanisms that can lead to non-symmetric assemblies , for instance pseudo-symmetry or self-occlusion producing steric hindrance on an heterologous interaction [16] . However those exceptions are rare and the vast majority of known protein oligomers are symmetric . We discuss some of the exceptions in the section Exceptions to the rules below . The crystal lattice can be represented by a periodic graph with protein chains as nodes and interfaces between them as edges . Graphs that represent lattices are widely used in crystallography ( especially for small molecules ) and are also known as crystal nets . The excellent book by Sunada [17] contains an in-depth account of the mathematics of crystal nets . Here , we apply them to whole macromolecules rather than individual atoms and bonds , as is more typical in small molecule crystallography . We label nodes and edges to identify the molecular entities and the distinct mode of interactions between them , see Fig 1b . A node is identified by a chain identifier and a symmetry operator identifier ( e . g . A_1 ) , while an edge is identified by a numerical interface identifier . Additionally , all nodes corresponding to the same molecular entity are given an entity label and all edges corresponding to the same interface type are given an interface type identifier label . Although the graph is depicted in one unit cell only , it does represent all possible connections in the crystal including those across neighboring unit cells . The crystal translations associated to the interfaces are also required to fully describe the graph and are essential in finding closed cycles with 0 net translation . We represent these as an integer vector for each edge giving the difference in Miller indices for the two chains participating in the interface , with respect to a given choice of unit cell operators . Diagrams similar to our 2D graph representation of the lattice graph have been used previously in the context of quaternary structure studies , see for instance [18] and [19] . The EPPIC website includes a visualization of the lattice graph using a custom graph layout generated by a stereographic projection of the subunits along an axis of symmetry , giving graphs with geometrically consistent node positions .
Given the definitions introduced in the section above , we now establish the rules for a superassembly to be valid , from which the algorithm to find all assemblies result: The first two rules ensure consistency in the decomposition of the superassembly into assemblies . The third rule is motivated by the assumption that co-crystallization of multiple biological assemblies involving the same entities does not occur . Co-crystallization implies that the complex exists at equilibrium in the crystallization conditions , making the correct biological assembly ambiguous . By disallowing co-crystallization we effectively favor the dissociated form as the correct assembly for proteins with weak or transient interactions . Finally , the fourth rule is motivated by the hypothesis that infinite assemblies are never biological ( discussed later ) . From the rules it follows that a ) valid assemblies are point group symmetric , and b ) heteromeric assemblies must have even stoichiometry . We then implement an algorithm that follows the above rules , described in detail below . Interface classification in EPPIC is described in our previous paper [12] . However , there have been some improvements to the interface scoring and classification . When calculating the sequence entropy at each position , we now use a 6-letter reduced alphabet to represent the 20 amino acids [20] . The alphabet was proposed by Mirny et al . [21]: {ACILMV} , {DE} , {FHWY} , {GP} , {KR} , {NQST} In addition , the core surface scores are now pure Z-scores where m residues are sampled 10 , 000 times from the whole protein surface . An average sequence entropy is calculated for each of those samples and then the mean and standard deviation of the whole distribution is used for the Z-score of the m residues composing the interface core . Finally , we have introduced a probabilistic scoring for interface classification , based on a logistic regression classifier that uses 2 of our 3 previous indicators: geometry ( gm ) and core-surface ( cs ) scores [22] . The model was trained using the Many dataset [4] with R generalized linear model ( glm ) functions . The equation that describes the probability of an interface being biologically relevant ( p ) is: A ROC curve with the performance of the new method can be found in Supplementary S1 Fig and is directly comparable to the curves in [4] . We denote interface types by numerical identifiers 1 , … , n , sorted from largest to smallest area . An assembly is created when engaging a subset of those interfaces , e . g . {1 , 3} is the assembly where only interfaces 1 and 3 are engaged , or {} is the empty assembly where no interfaces are engaged . Given the set of all interface types S = {1 , … , n} , enumerating all possible assemblies is a matter of traversing the tree of its power set P ( S ) . A total of 2n assemblies are possible in principle , making the full enumeration prohibitive when n becomes large . For every set , the assembly is tested against our rules to see if it represents a valid assembly . An important observation makes the problem more tractable: if a given set is invalid , all of its children ( i . e . any other set that contains the same engaged interfaces plus any other ) will also be invalid . This dramatically prunes the tree , making it possible to quickly do the exhaustive enumeration for almost all cases . As a further optimization , heteromers with many protein entities are reduced to equivalent homomeric lattice graphs by combining entities , leading to considerably simpler graphs . Interfaces that join different entities are selected in a greedy manner . The edge corresponding to the interface is then contracted , merging the two entities into a single node . This process is iterated until a single meta-entity remains . Graph contraction preserves the structure of the graph with respect to the validity properties and relative score , while allowing considerably faster superassembly enumeration . The test of validity for a given superassembly boils down to two tests: graph isomorphism and finding closed cycles in the graph . To find the cycles we use the Paton algorithm [23] as implemented in the JGraphT library . The EPPIC software package implements all of the described algorithms in its new version 3 . The software is written in Java , using BioJava [24] as the underlying software library to handle the biological data . In order to predict the most likely biologically relevant assembly we use a combination of the probabilistic scores calculated for the pairwise interfaces . By the uniform composition rule all interfaces of a type must be engaged together , so interfaces of the same type are considered together as a binary event that can either occur or not in biological conditions . An assembly is just a subset of the interface types in the crystal occurring in biological conditions , with the remaining interface subsets not occurring . Each assembly is characterized by a boolean vector s1 , s2 , … , sn with si indicating type i is engaged in the assembly . The probability pi of interface type i being biologically relevant is calculated as the average of its interfaces ( Eq 1 ) . To estimate the probability of an assembly occurring in biological conditions is to estimate the joint probability of events coming from all interfaces in the crystal: P ( a s s e m b l y ) = P ( s 1 , s 2 , … , s n ) ( 2 ) To perform the estimation , we assume that the pairwise interfaces can be treated as independent . As EPPIC interface scores depend critically on the estimation of residue burial , this assumption is valid for the score in Eq 1 so long as the total buried surface area of the assembly is well approximated by the sum of the pairwise buried surface areas . This assumption may be violated for proteins where three or more subunits interact in a confined region , such that the calculation of buried surface area would change significantly depending on whether the third chain is included in the calculation or not . In this case the assembly cannot be decomposed into pairwise interfaces . However , this is rare and only affects small interfaces , so it is not considered by EPPIC . Using the probabilities for each interface pi , we can assign a probability of occurring in biological conditions to each assembly of the crystal: P ( s 1 , s 2 , … , s n ) = ∏ i = 1 n p i I ( s i ) + ( 1 - p i ) ( 1 - I ( s i ) ) ( 3 ) where I ( s i ) is the indicator function ( I ( s i ) = 1 if si else 0 ) . Note that interface types are not weighted according to the number of interfaces . While one does expect some cooperative effects due to avidity , including this in the assembly probability calculation ( e . g . by taking the product over individual interfaces rather than interface types ) would bias the scores towards high-order cyclic assemblies . For this reason it was decided to model the probability of the complete assembly as the product of the engaged interface types . Some combinations of engaged interfaces will correspond to invalid assemblies according to the rules above . These assemblies have a probability of occurrence of 0 , so summing P ( S ) over all valid assemblies in the crystal may be less than one . Thus , a final normalization step can be applied to redistribute the probability mass of interface events leading to invalid assemblies into the valid assemblies . Special care has to be taken with induced interfaces , which can be omitted from an assembly without changing the quaternary structure . Superassemblies which differ only by an induced interface can be easily detected by comparing the stoichiometry of their constituent assemblies . This allows all superassemblies which differ only by induced interfaces to be combined together . The superassembly with the highest number of engaged interfaces is reported along with the total probability of all equivalent superassemblies . The reported probability for an assembly is the confidence that the EPPIC call is correct . It is important not to confuse these probabilities with strength of the assembly or transitivity properties . We compiled a new dataset of biological assemblies using the annotations of deposited structures in the PDB . We started with 96 , 594 crystal structures with higher than 3 Å resolution and lower than 0 . 3 R-free value from the PDB . Structures were then grouped into 60 , 034 unique sequences and 36 , 843 70% sequence identity clusters for each of their chains . These were further filtered to clusters with at least three structures and where all structures had the same biological assembly annotation . Randomly selecting a representative from each of the remaining clusters yielded 1 , 481 proteins . This new dataset of biological assemblies from PDB1 annotations represents a diverse sample of the PDB: 53% of oligomers , from which 11% are heteromers , covering macromolecular sizes up to 24 partner subunits . Together with the command line interface ( downloadable at http://eppic-web . org/ewui/#downloads ) , we provide a web server with a graphical user interface to the EPPIC 3 software . There has been numerous improvements compared to what we described earlier . A new view provides the full enumeration of all valid assemblies found in the crystal structure with links to its constituent interfaces . The assemblies are visualized by thumbnail images of the assembled proteins and by 2-dimensional diagrams of their corresponding graphs . New lattice graph visualizations are provided . First in 2D with the help of the vis . js library [25] . An optimal 2D graph layout is achieved by performing a stereographic projection of the 3D molecule . A 3D lattice graph representation is also provided with NGL [26] by overlaying custom made spheres and cylinder objects on top of a semi-transparent cartoon representation of the unit cell . In EPPIC 2 , the 3D visualization was based in the Jmol molecular viewer . The server now uses NGL [26] as the molecular visualization software . NGL is written in JavaScript and runs natively in the browser with very good performance thanks to WebGL technologies . Its advanced features allow for showing sequence entropy surface color representation within the browser .
We validated our assembly assignment method against the dataset of 1 , 481 PDB entries with consensus quaternary structure annotations ( PDB1 dataset ) . Fig 2 shows the confusion matrix of the assembly size for EPPIC predictions , with an overall precision of 85% . While the precision is constant across the different macromolecular sizes , the recall is lower for larger assemblies . The consequence is the reduction of non-biological large macromolecular assembly predictions ( top-left of the matrix in Fig 2 ) , at the expense of predicting some partial assemblies ( bottom-right of the matrix in Fig 2 ) . As a further validation , we provide a comparison to the popular PISA method , the de-facto standard in the field . Despite very similar overall precision in the assemblies dataset , EPPIC and PISA predictions show many differences , as it can be appreciated in Fig 3 . The most important difference is that PISA makes the opposite trade-off in the prediction of large macromolecular size assemblies , achieving better accuracy for larger assemblies at the expense of predicting some non-biological large assemblies . Table 1 gives the overview of over and under predictions , whilst Table 2 contains more detailed statistics divided into 3 categories: monomers , dimers and higher oligomers . The agreement of the two methods greatly increases the confidence of a prediction . As observed in Fig 4 , when EPPIC and PISA agree , in 78% of the cases , the error rate is only 5% . On the other hand , when the methods disagree , in the remaining 22% of the structures , the error rate of each method is around 50% . Therefore , each method corrects roughly the same amount of assignments of the other . Furthermore , at least one of the two methods is correct in 95% of the cases . These results suggest that a meta-method combining EPPIC and PISA could be successful , with a potential precision of up to 95% . Indeed a recent publication [27] reports a meta-predictor method ( QSbio ) combining the predictions from QSalign [27] , PISA and EPPIC version 2 , achieving higher precisions than either method alone . Additionally to the benchmark with our dataset we have also measured the performance with the PiQSi dataset [6] , composed of 1315 biological assemblies curated with a combination of manual community annotation and automatic methods . The precision values for the PiQSi benchmark are 73% for EPPIC and 79% for PISA . It should be noted that the PiQSi dataset is less representative of the PDB compared to our dataset , for instance having fewer monomers and more very large oligomers than average in the PDB . In most cases , the quaternary structure interpretation of a crystal is unambiguous to a trained crystallographer . The unit cell shows clear blocks of symmetrically packed molecular entities . However , in more difficult cases the interpretation of the crystal is far from obvious and requires very careful observation . A good example is the crystal structure of the fimbrial adhesin FimH protein ( PDB 2VCO [28] ) . The crystal contains two FimH molecules in the asymmetric unit interacting via a heterologous interface . All other interfaces in the crystal are also heterologous , except for the very weak isologous interface 6 ( as identified by EPPIC , see http://eppic-web . org/ewui/#interfaces/2vco ) . No combination of the interfaces produces a closed cycle ( assembly rule 4 is not satisfied ) . Thus the only valid assembly in the crystal is monomeric ( see http://eppic-web . org/ewui/#id/2vco ) . However , the PDB annotation for this case engages interfaces 1 and 3 to form a tetramer . The global symmetry of the tetramer , as calculated by the RCSB PDB website [29] , is C2 , indicating that the tetramer is not point group symmetric ( the only possible point groups for an A4 stoichiometry are C4 or D2 ) . The assembly might seem reasonable since in the crystal it shows as an independent block repeated throughout ( see Fig 5a ) . Fig 5b helps explain this with a simple 2D schematic representation of a crystal packing with heterologous interfaces . The PISA software predicts in this case a different tetrameric assembly than the one annotated in PDB , formed by engaging interfaces 1 and 6 . Again this assembly does not contain point group symmetry . This example also shows how a simple search for stoichiometry-symmetry imbalance ( i . e . An stoichiometry should have Cn or Dn/2 point group symmetry ) would uncover similar cases of potentially erroneous annotations in the PDB . Another similar example is lipoteichoic acid synthase LtaP from Listeria monocytogenes ( PDB 4UOP ) , which corresponds quite closely to the schematic representation of Fig 5b: 2 molecules in the asymmetric unit interact through a heterologous interface , with the heterologous interface capped in the crystal by other molecules . The PDB annotates the asymmetric dimer in the AU as the biological assembly based on a PISA prediction . However , the protein is known to be a monomer in solution based on size exclusion chromatography [30] . Since the dimeric assembly is not symmetric , EPPIC considers it invalid following the assembly rules . A second example of a subtle lattice that is difficult to analyze manually would be that of the crystal structure of the putrescine receptor PotF from E . Coli ( PDB 1A99 [31]; see Fig 6a ) . There are 4 PotF molecules in the asymmetric unit . Two different isologous interfaces relate the 4 molecules in the AU , interfaces 5 ( D+C ) and 6 ( B+A ) . The PDB annotates a dimeric assembly through one of the interfaces in the asymmetric unit ( interface id 6 ) . In principle , the assembly is valid since it has C2 point group symmetry . However , a more careful analysis of the crystal shows that not all monomers in the lattice participate in this kind of interaction: the C and D chains do not interact in the same way throughout the crystal . Considering this assembly as a dimer would break the full coverage rule ( rule 1 ) , while considering it a co-crystal dimer + monomer breaks the isomorphism rule ( rule 3 ) . This shows why isomorphism is important: a stable assembly in solution can not occur only in some parts of the lattice and not in others . The schematic 2D view of Fig 6b helps visualize the problem . By following the assembly rule , EPPIC finds here only a monomeric assembly ( see http://eppic-web . org/ewui/#id/1a99 ) . In this case PISA predicts a disjoint assembly formed by a A2 B2 tetramer and separate monomers of chains C and D . Non-symmetric assemblies are very rare but still a possibility . In fact as of June 2017 , 96% of PDB structures are annotated with symmetric biological assemblies . A comprehensive study of asymmetric assemblies in heteromers [16] found a similar fraction of asymmetric cases for heteromers ( 9 . 8% of all heteromers have uneven stoichiometry ) . In their in-depth study , a thorough review of all cases unearthed a number of quaternary structure assignment errors , further lowering the asymmetric fraction . Different mechanisms can lead to breakage of symmetry . One major cause of exceptions is the existence of pseudosymmetry in heteromers with uneven stoichiometry ( e . g . PDB 4FI3 [32] ) , whereby one entity can bind several copies of its partner at distinct but structurally similar binding sites . Other exceptions include steric hindrance ( e . g . PDB 3Q66 [33] ) and extreme conformational flexibility ( e . g . PDB 1YGY [34] ) . An additional source of exceptions is filamentous proteins and amyloids , which violate rule 4 by definition . However , since these properties make them resistant to crystallization , such cases are rare . A prominent example of a pseudosymmetric case is that of the B12 vitamin transporter [32] . This large membrane protein complex is composed of 5 subunits , with 3 distinct molecular components ( Fig 7a ) . Two BtuD chains form a symmetric C2 dimer in the cytoplasmic domain , while the transmembrane domain is composed of two BtuC chains arranged along the same C2 axis . Capping the complex on the periplasmic side is a single BtuF chain that binds to the BtuC dimer in a symmetric way . The 1:2 symmetric binding is made possible by the internal pseudosymmetry of the BtuF chain ( see Fig 7b ) . We have presented an approach to enumerate and predict quaternary assemblies from protein crystal structures . This new method should prove very useful to the crystallographer , considerably easing the assembly interpretation of protein crystals . The automated exhaustive enumeration of assemblies represents a great improvement in the quaternary interpretation of structures , which to a large extent still requires human subjective interpretation . Our ideas are centered in the necessity of symmetry based on the simple arguments established by Monod , Wyman and Changeux [14] . Symmetry is essential for stable soluble proteins . Our method can thus help in avoiding mistaken asymmetric interpretation of assemblies . It can also serve as a validation tool for atomic models that lack symmetric or isomorphic assemblies , providing hints on possibly uninterpreted regions of electron density that need to be added to the model in order to complete it . Additionally existing methods to predict quaternary assemblies [8] are not always strict in the symmetry constraint , providing sometimes misleading interpretations of the crystal . Importantly , our assembly scoring uses evolution as the ultimate arbiters to the biological relevance of the assemblies , making this method complementary to existing methods based on thermodynamic estimations . Also , the newly introduced confidence values provide a clear guide to interpreting the predictions . At the same time , confidence estimations provide a means to more reliably estimate biological assembly annotation errors in the Protein Data Bank , as well as aiding the crystallographer in deciding when additional oligomeric experimental evidence for a particular assembly might be needed . Confidence values also allow for fully automated analyses of oligomeric complexes at the PDB wide level . Some new avenues of research are possible based on this new resource . For instance , the assembly graphs allow for more detailed study of different crystal lattices and their relationships across the PDB . We also recognize that our strict enforcement of point group symmetry is not always ideal , since , as shown in the Results section , exceptions to symmetry do occur . In future work we plan to address the problem by relaxing some of the conditions in cases where interface scoring indicate an invalid assembly could be biological . Recent publications indicate that the evolutionary approach to protein assembly prediction and classification can be significantly improved in the future . Two research lines are promising: co-evolution of inter-subunit residues in protein-protein interactions [36 , 37] and the evolutionary constraints of highly symmetric assemblies to avoid supramolecular assembly formation [38] . We believe that these additional sources of information can improve the performance of the classifier and confidence estimates , as we continue to advance the method . | X-ray diffraction experiments are the main experimental technique to reveal the detailed atomic 3-dimensional structure of proteins . In these experiments , proteins are packed into crystals , an environment that is far away from their native solution environment . Determining which parts of the structure reflect the protein’s state in the cell rather than being artifacts of the crystal environment can be a difficult task . How the different protein subunits assemble together in solution is known as the quaternary structure . Finding the correct quaternary structure is important both to understand protein oligomerization and for the understanding of protein-protein interactions at large . Here we present a new method to automatically determine the quaternary structure of proteins given their crystal structure . We provide a theoretical basis for properties that correct protein assemblies should possess , and provide a systematic evaluation of all possible assemblies according to these properties . The method provides a guidance to the experimental structural biologist as well as to structural bioinformaticians analyzing protein structures in bulk . Assemblies are provided for all proteins in the Protein Data Bank through a public website and database that is updated weekly as new structures are released . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"and",
"discussion"
] | [
"dimers",
"(chemical",
"physics)",
"crystal",
"structure",
"condensed",
"matter",
"physics",
"crystal",
"lattices",
"crystals",
"materials",
"science",
"protein",
"structure",
"crystallography",
"materials",
"by",
"structure",
"solid",
"state",
"physics",
"protein",
"st... | 2018 | Automated evaluation of quaternary structures from protein crystals |
K-Ras , one of the most common small GTPases of the cell , still presents many riddles , despite the intense efforts to unveil its mysteries . Such is the case of its interaction with Calmodulin , a small acidic protein known for its role as a calcium ion sensor . Although the interaction between these two proteins and its biological implications have been widely studied , a model of their interaction has not been performed . In the present work we analyse this intriguing interaction by computational means . To do so , both conventional molecular dynamics and scaled molecular dynamics have been used . Our simulations suggest a model in which Calmodulin would interact with both the hypervariable region and the globular domain of K-Ras , using a lobe to interact with each of them . According to the presented model , the interface of helixes α4 and α5 of the globular domain of K-Ras would be relevant for the interaction with a lobe of Calmodulin . These results were also obtained when bringing the proteins together in a step wise manner with the umbrella sampling methodology . The computational results have been validated using SPR to determine the relevance of certain residues . Our results demonstrate that , when mutating residues of the α4-α5 interface described to be relevant for the interaction with Calmodulin , the interaction of the globular domain of K-Ras with Calmodulin diminishes . However , it is to be considered that our simulations indicate that the bulk of the interaction would fall on the hypervariable region of K-Ras , as many more interactions are identified in said region . All in all our simulations present a suitable model in which K-Ras could interact with Calmodulin at membrane level using both its globular domain and its hypervariable region to stablish an interaction that leads to an altered signalling .
Ras proteins are well-known small GTPases involved in the regulation of key signal transduction pathways . Cycling from the inactive ( GDP-bound ) to the active ( GTP-bound ) state , they faithfully respond to extracellular signals due to their tight regulation by GTP-exchange factors ( GEFs ) and GTPase activating proteins ( GAPs ) . In the GTP-bound form , two regions of the protein change conformation ( switch I and II domains ) allowing its binding with different effector proteins and consequently the activation of diverse signal transduction pathways . Among those , the best characterized are the RAF1/MEK/ERK and the phosphatidylinositol-3-kinase ( PI3K ) /AKT [1] , which are known to regulate proliferation , differentiation and survival . Activating point mutations render Ras proteins that are always found in the GTP-bound state independently of the extracellular signals and are crucial steps in the development of the vast majority of cancers [2] . Ras genes were the first oncogenes identified in human cancer cells , and nowadays they are well established as the most frequently mutated oncogenes in human cancer [3] . Three different genes code for a total of four different Ras isoforms named H-Ras , N-Ras , K-RasA and K-RasB ( herein after referred to as K-Ras ) . K-Ras is the most frequently mutated oncogene in solid tumors and its oncogenic mutations occur mostly in pancreatic ductal adenocarcinomas ( 95% ) , colon ( 40% ) and adenocarcinomas of the lung ( 35% ) [3–5] . Although they all have a highly conserved globular domain ( from residue 1 to 165 ) containing the guanosine nucleotide and effector binding sites ( Switch I and Switch II ) , the last C-terminal residues of Ras proteins , named the hypervariable region ( HVR ) , which contains the membrane targeting signals , are not conserved among the different isoforms . H- and N-Ras achieve high-affinity hydrophobic membrane binding mainly through lipid modifications . By contrast , K-Ras has , adjacent to the farnesylated cysteine Cys185 , a stretch of lysine residues—known as the polybasic domain—that promotes an electrostatic interaction with the negatively charged phospholipids [6 , 7] , which confines K-Ras almost entirely to non-raft microdomains within the plasma membrane [8] . The different membrane anchors interact with lipids and proteins of the plasma membrane and , together with the hypervariable region ( HVR ) , drive the Ras isoforms into spatially and structurally distinct nanodomains , of which each then contains a cluster of molecules ( nanocluster ) [9–11] . Importantly , the nanodomains that are occupied by the three isoforms of Ras do not show any overlap . Furthermore , not only are the different Ras isoforms laterally segregated , but inactive GDP-loaded Ras occupies nanodomains that are spatially distinct from those occupied by the active GTP-loaded form . Formation of these nanoclusters is essential for activation of mitogen-activated protein kinases ( MAPKs ) , because they constitute exclusive sites in the plasma membrane for Raf-1 recruitment and ERK activation [12–14] . Because oncogenic mutations of K-RAS give rise to an always GTP-bound protein that constitutively binds to effectors , positive or negative physiologic regulation of oncogenic K-RAS was not initially expected . In recent years , interaction of K-Ras with proteins , which are not effectors but which may function as allosteric regulators or scaffolds , have been described and some proved to be crucial to fully display K-RAS oncogenic phenotype . Galectin-3 [15] , calmodulin ( CaM ) [16] , phosphodiesterase δ [17 , 18] , nucleophosmin , nucleolin [19] and heterogeneous nuclear ribonucleoprotein A2/B1 ( hnRNPA2/B1 ) [20] have been shown to interact with K-Ras and modulate its downstream signaling . The mechanism by which these proteins modulate K-Ras signaling is diverse: phosphodiesterase δ by binding to the farnesyl group facilitates the diffusion of K-Ras from endomembrane to the cytoplasm , ultimately favoring its correct relocalization to the plasma membrane and consequently enhances its signaling [18]; Galectin-3 regulates K-Ras nanoclustering at the plasma membrane and also enhances its signaling [15]; and , hnRNPA2/B1 favors the interaction of K-Ras with PI3K [20] . In contrast , while K-Ras interaction with CaM has been known for many years , there is still some controversy regarding the consequences of this interaction . Our group demonstrated that CaM interaction with K-Ras inhibits K-Ras signaling to Raf/MEK/ERK [16] and inhibits K-Ras phosphorylation at Ser181 in the HVR [21] . Interestingly , CaM also binds to PI3K enhancing its activity [22] , and the existence of a complex containing K-Ras , CaM and PI3K has been proposed [23] . CaM is a small ( 148 amino acids ) and well conserved Ca2+-binding protein [24] . The crystal structure of CaM in the Ca2+-bound form shows a dumbbell-shaped molecule with two globular domains arranged in a trans configuration . These domains are connected by a long extended central α-helix , the middle portion of which is highly mobile and acts as a flexible tether . Each domain consists of two helix-loop-helix motifs ( EF hands ) , with each binding one molecule of Ca2+ . Ca2+ binding changes the orientation of the two EF hands of each domain , inducing the appearance of hydrophobic patches that interact with proteins known as CaM-binding proteins ( CaMBPs ) [25] . Binding of CaM to CaMBPs modulates the function of these proteins and , in consequence , affects many aspects of cell regulation . The carboxyl-terminal lobe binds Ca2+ with high affinity ( Kd 10−7 M ) , whereas the amino-terminal lobe binds it with lower affinity ( Kd 10−6 M ) . The fact that the Kd values fall within the range of intracellular Ca2+ concentration exhibited for most cells ( 10−7–10−6 M ) makes it a good sensor for changes in Ca2+ intracellular levels [26–28] . The CaM binding domain of some of the CaMBPs with high affinity for CaM ( nM range ) consists of a 20-amino acid sequence that has an amphipathic α-helix conformation [29] . CaM binding domains with lower affinity for CaM ( μM range ) have also been described [30] . Some proteins like MARCKS and CAP-23/NAP-22 use the myristoyl group to interact with CaM [31 , 32] . As well as K-Ras , diverse Ras superfamily GTPases like Kir/Gem [33] , Ric [34] , Rin [35] , Rab3A [36] , and RalA [37] have been shown to bind to CaM . Biochemical data indicate that at least two different regions in the K-Ras molecule are important for K-Ras/CaM interaction: the hypervariable region and the α-helix between amino acids 151 and 166 [38] . Within the hypervariable region , both the hydrophobic farnesyl group and the positive-charged amino acids were essential for the interaction between K-Ras and CaM . Consistently , K-Ras S181D mutant , which mimics phosphorylation of Ser-181 of K-Ras , also completely abolished binding to CaM . Although the presence of the farnesyl group increases the affinity of purified K-Ras to CaM , full length non-farnesylated K-Ras still has micromolar affinity for CaM [39] . Accordingly to the above mentioned , the NMR data of this complex show that the N-terminal lobe of CaM interacts with the globular domain of K-Ras and the C-terminal lobe of CaM interacts with the HVR [40] . But controversial data exist regarding how CaM interaction with K-Ras could modulate K-Ras activity . While some data indicate that CaM could extract K-Ras from membranes in vitro , most probably by interacting with the farnesyl group [41 , 42] , it is not clear if in vivo this hydrophobic group would always be available for CaM to interact with . In fact , our group has demonstrated that K-Ras and CaM colocalize mainly in the plasma membrane , suggesting that in vivo interaction does not directly lead to K-Ras internalization [38] . CaM could be modulating interaction of K-Ras within the plasma membrane , with effectors , scaffolds or with different lipids , ultimately regulating K-Ras signaling from the plasma membrane . Thus , modelling of K-Ras/CaM interaction is important to decipher the cellular role of this interaction . To mimic the situation of K-Ras bound to the membrane , thus with farnesyl group hindered between the phosphoslipids , we aimed to model the interaction between a full length non-farnesyated K-Ras and CaM . CaM and K-Ras have been widely studied computationally [43 , 44]; in fact , CaM is one of the most studied proteins with molecular dynamics ( MD ) due to its high degree of flexibility . These systems have also been joined to a certain degree [45] , but up to date no simulations of the whole proteins have been performed . Thus , we decided to carry on conventional MD ( cMD ) on a system with both proteins in order to determine which the details of the interaction are . Furthermore , in order to increase the exploration of the conformational space of the K-Ras/CaM system , scaled MD ( sMD ) a recently developed methodology that proved to be effective to sample wider conformational areas faster than cMD [46] , was used .
In order to computationally study the interaction between K-Ras and CaM , a system with both proteins had to be prepared . Since NMR experimental data regarding the interaction between these two proteins has already been published [40] , we decided to mimic the experimental settings: oncogenic K-Ras ( G12D mutation ) full-length without post-translational modifications paired with holo-CaM . Prior to a simulation between the proteins , a system composed of GTP-bound K-Ras with a fully extended HVR was prepared . This system was used to determine whether the HVR could be found in an extended conformation in several frames or if it would be mostly bent to interact with the globular domain . Fifty nanoseconds of cMD were performed and the provided trajectories were analyzed by measuring the distance between residues 161 ( from the α-helix 5 ) and 178 ( from the HVR ) . The HVR presented an extended conformation most of the time , showing great motility ( Fig 1A ) . Interestingly , other groups have seen similar behavior when simulating K-Ras in its active state , reporting that the HVR does not significantly interact with the globular domain [47] . Since the binding of these two proteins does not seem to be mediated by the common binding mechanism of CaM ( where it wraps its lobes around a single structure , such as an α-helix ) , we decided to use the structure of CaM with PDB code 2MGU . This structure presents its lobes rather extended , which could fit with a model in which the N-lobe of CaM interacts with the globular domain of K-Ras and the C-lobe interacts with the HVR . The peptide present in the structure was replaced by the HVR of K-Ras with Modeller , and subsequently rotated to fit the model ( see Methods for more details ) . Last , the globular domain of K-Ras was attached to obtain the system with both proteins ( Fig 1B ) . Once the system was prepared , a total of 6 cMD and 4 sMD simulations were carried out , each of them with a total length of 50 ns . The trajectories were visually analyzed in order to determine which simulations had stablished a proper interaction between the two proteins , and in which K-Ras/CaM had fallen apart . Interestingly , in 9 out of 10 simulations the proteins interacted throughout most of the simulation length , even with the additional energy boost of the sMD simulations ( Fig 2 ) . Furthermore , the N-Terminal domain of CaM remained close to the α-helix 5 of K-Ras in most of the simulations . The end of the HVR maintained a close contact with the C-Terminal lobe of CaM , while the polybasic domain of K-Ras interacts with the linker region of CaM . The energy of the system was determined by performing a MMPB/GBSA analysis . The dynamics were considered stable if the last 5 ns did not present significant deviations . If any of the simulations were not stable enough , they were extended until stability was reach . The energy profile was similar for PB and GB . The interaction presented between -60 and -100 kcal/mol for GB and between -40 and -120 kcal/mol for PB both for cMD and sMD ( Fig 3 ) . The last ns of each simulation were used to calculate the contribution of each residue to the binding energy through a “per residue” analysis . The residues of CaM were studied in order to find matches with the experimental data available . Two thresholds were imposed to consider a residue as actively participating in the interaction between K-Ras and CaM: the first was a requisite of at least -0 . 7 kcal/mol of average contribution to the binding , whereas the second was its presence in at least 3 of the simulations . Up to twelve residues matched with the experimental data available , many of which are negatively charged residues ( 78 to 84 ) that can interact with the polybasic domain of K-Ras ( Fig 4A ) . Intriguingly , certain residues of CaM whose surroundings are modified when interacting with K-Ras ( experimentally ) do not present a significant implication in the interaction between both proteins in the simulations ( Fig 4A ) . The presence of changes in nearby residues when binding to other proteins can explain why there are NMR shifts assigned to those residues while no energy contribution is seen in our simulations ( Fig 4B ) . With all things considered , the model can be considered robust enough to analyze the residues of K-Ras that participate in the interaction , some of which have not been described yet . After analyzing the residues of CaM , we focused on the residues of K-Ras relevant for the interaction . A threshold of -1 kcal/mol of average was imposed to the residues that participated in the interaction . Also , their participation had to be present in at least 5 simulations . In concordance with the experimental data , most of the residues responsible for the interaction were found within the HVR . However , 5 residues were identified in the globular domain . Furthermore , most of them presented energy values below -3 . 5 kcal/mol , being arginine 135 the most significant residue in terms of average energy ( Fig 5A ) . When visualized , the simulations revealed that the selected residues of the globular domain were , in fact , closely interacting with CaM . The arginines from the α –helix 5 formed hydrogen bonds with the EF hand of the N-Terminal domain , while arginine 135 and proline 140 interact with one of the four α helixes present in the N-Terminal lobe of CaM ( Fig 5B–5E ) . To further confirm the results obtained with the performed simulations , a different methodology was used: the umbrella sampling . This kind of simulation allows a more progressive scenario for the proteins to adapt , as more time is given to position them nearer . To perform this simulation , a restriction was added to maintain the mass center of the α-helix 5 of K-Ras and the mass center of the N-Terminal lobe of CaM at a prefixed distance . Then , the restricted groups were slowly approached , at a rate of 0 . 5 Å per step , remaining for 0 . 5 ns at each distance before closing the gap between them . The initial distance was set at 20 Å , while the last step was set at 5 Å . Once the simulations were performed , structures from the US with the mass centers maintained at 5 , 6 , 7 and 8 Å of distance were obtained and 10 ns of cMD were calculated . All these simulations presented high interaction between K-Ras and CaM , with the N-Terminal lobe of CaM wrapping the α-helix 5 of K-Ras and the HVR embedded in the C-Terminal lobe of CaM ( Fig 6A ) . A MMGBSA analysis was also performed to determine the binding energy of the proteins and analyze the stability of such interaction ( Fig 6B ) . Values around -150 kcal/mol were obtained for all four simulations , exceeding the values seen in simple cMD or sMD ( whose values were around -100 Kcal/mol ) . A “per residue” analysis was also performed so as to determine if the residues described as relevant with the previous methodology were still actively participating in the binding . Since these residues had more time to accommodate and orientate in a favorable angle for the interaction , only those residues actively interacting in the four simulations with an average energy below -1kcal/mol were selected . Even though according to the US simulations some residues selected with the initial model were not relevant for the interaction , most of them matched . Furthermore , only one of the studied residues of the globular domain did not surpass the thresholds , which backs up the idea that the globular domain is playing a part in the interaction ( Fig 6C ) . Taking into account all the data provided by the simulations performed , it seems the globular domain interacts with CaM , specifically through residues R135 , P140 , R161 , R164 and , to a minor extent , K165 . With the purpose of verifying the obtained results with experimental data , three mutants of the globular domain ( 1–166 aa ) were obtained through point mutation . The mutants were designed according to the results of the simulations: R135E , R161E and R164E . The corresponding GST-K-Ras mutants were expressed in bacteria , affinity purified and then its binding to CaM determined by Surface Plasmon Resonance ( SPR ) . Biotinilated CaM was attached to a chip with streptavidin and the GST-tagged globular domain of K-Ras ( either wild-type or mutant ) was injected as an analyte . A control flow cell with no CaM , was also injected with the globular domain as a blank . To discard that the binding was due to the GST tag , recombinant GST was injected in all flow cells and no binding was observed . The injection of the globular domain of K-Ras led to an increase in the Resonance Units ( RU ) of the flow cells with CaM , which stemmed from the binding of the injected protein to CaM . The mutants also showed binding with CaM , but to a lower extent . An affinity study was performed to determine the KD , and the results reflected that the wild-type globular domain presented a lower KD than any mutant . All the mutations led to an increase in the KD , that is , in a reduction of the affinity with CaM ( Fig 7 ) . Thus , our experimental data support the results obtained through computational simulations , where these 3 residues were identified as key players in the interaction of the globular domain of K-Ras with CaM .
Even though it has been almost twenty years since the discovery of the interaction between K-Ras and CaM [16] , its subtleties have remained elusive . In the present work , thanks to the use of computational aided techniques , we have shed some light upon this interesting matter . Since these proteins have never been simulated computationally , we looked for a system with validated experimental results to be able to contrast the data generated . Thus , we used full-length K-Ras without its farnesylation , as this system had already been experimentally studied with NMR [40] . Even though we are aware of the relevance of the post-translational modifications of K-Ras for the interaction with CaM , it’s unlikely that CaM can initially interact with the farnesyl group , as it will be attached to the plasma membrane . Thus , our simulations could mimic a situation in which the farnesyl group is not available for the interaction , as it would be hidden within the membrane . We used CaM with an extended conformation , with each lobe interacting with different domains of K-Ras , rather than wrapping around an specific region , similar to the interaction with other proteins in which the lobes of CaM interact with different regions [25] . It could be observed that the interaction between the proteins was stable , as it was maintained throughout most of the simulations . Furthermore , several residues of CaM matched the experimental data , despite the motility present in the system and the energy boosts provided in the sMD simulations . Our group had previously described the participation of the globular domain and specially α-helix 5 of K-Ras in the interaction with CaM [38] . In the present work , we have taken another step forward to model this interaction and have confirmed and identified new residues of the globular domain that are implicated in said interaction . Arginine 161 , arginine 164 ( α-helix 5 ) and , to a minor extent , arginine 135 ( α-helix 4 ) seem to be responsible of the interaction with the N-terminal lobe of CaM , since single point mutations in any of those residues lead to an increase in the experimental values of KD between the globular domain of K-Ras and CaM . Interestingly , previous publications have highlighted the relevance of residue 135 in Ras signaling , as mutations in this residue led to enhanced binding with C-Raf RBD [48] . This phenomena may be explained by the diminished interaction with CaM , as the binding of this protein to K-Ras is known to diminish MAPK pathway signaling [16] . While we previously showed that the K-Ras switch II mutant , R68D/R73D , had a compromised interaction with CaM [38] , our present model does not predict direct interactions of these two arginines with CaM . The most plausible explanation is that the substitution of the two positive residues by negative ones induces a conformational change in K-Ras , and indirectly , an increase in the negative charge density in the surface of CaM interaction that prevents the binding with this acidic protein . As for the HVR , the simulations we have performed here proven to fit the available data . The highly negatively charged linker region of CaM is attracted to the polylysine domain of K-Ras , where they interact through electrostatic couplings . This fact has already been described experimentally by other groups [39] . As for the last C-Terminal residues of K-Ras , they are embedded by the C-Terminal lobe of CaM . However , it must be considered that this interaction may vary after the post-translational modifications , as the–AAX residues are removed and the farnesyl group is attached . Interestingly , in most of the simulations ( six out of nine ) cysteine 185 is not embedded within the C-terminal lobe of CaM , which would fit with a model in which the farnesyl group of this residue would be attached to the plasma membrane . Our simulations can help to understand why the phosphorylation of K-Ras leads to the abrogation of this interaction with CaM [49] . As shown in the performed MD , the polybasic region of K-Ras plays an important role in the interaction with CaM , creating electrostatic interactions with the acidic linker region . Thus , the addition of a phosphate group , highly negatively charged , is bound to have a negative impact on the K-Ras/CaM interaction . Our model may also provide one of the reasons why CaM does only interact with K-Ras when bound to GTP ( its active state ) [16] . Since the α-helixes of K-Ras are oriented towards the membrane when bound to GDP [50] , the N-Terminal lobe of CaM would not be able to reach its interaction zone with the globular domain of K-Ras , as it would be covered by the PM . When active , K-Ras would expose its α-helix 4 and 5 , giving the N-Terminal lobe of CaM a chance to interact with it . However , lack of interaction of full-length GDP-loaded K-Ras has also been described in the absence of lipid membranes . In this case the proposed autoinhibitory effect of the HVR could prevent CaM binding [44]: the globular domain would be inaccessible for CaM due to its binding with the HVR when K-Ras is in its GDP bound state , but would become reachable when GTP is loaded and the HVR is released . In fact , our simulations support the idea that , when bound to GTP , the HVR of K-Ras is not stably interacting with the globular domain . The study we performed of the dynamism of the HVR revealed that , even though there are some interactions between these groups , they are neither stable nor prolonged through much more than a few nanoseconds , thus giving to CaM the opportunity to interact with the HVR . However , it must be considered that experimental data show that CaM fails to interact with the purified inactive globular domain of K-Ras [16] , so , despite our model being able to provide some explanations , a few details remain elusive . Beyond the ins and outs of the interaction , the biological significance of such binding is becoming more interesting day after day . Although other interaction models between CaM and K-Ras may be feasible , especially with cytosolic K-Ras , our simulations would support a model in which K-Ras and CaM would interact at membrane level without indispensably inducing an extraction of K-Ras from the membrane ( Fig 8 ) , a fact that has been previously described by our group [38] , and lately supported by recent publications that demonstrate that CaM can bind to K-Ras even when attached to nanodiscs emulating diverse types of PM [45] . In fact , CaM interaction with K-Ras may be modulating K-Ras clustering and signaling from the PM . For instance , CaM is thought to form a ternary complex with K-Ras and PI3K , which would enhance K-Ras signaling through AKT signaling while diminish it through Raf [23] . Our simulations provide interesting data suggesting that , while keeping one of its lobes interacting with K-Ras ( probably the C-Terminal due to the interaction of the linker region with the polybasic domain ) , CaM could use its other lobe to interact with PI3K . Also regarding K-Ras signaling , several authors have described the relevance of K-Ras dimerization in the activation of downstream effectors . While dimerization through α-helix 1 and β sheets 1/2 would inhibit the binding of effectors such as Raf or PI3K , due to the overlapping interaction surfaces , dimerization of K-Ras using the α-helixes 3/4/5 and β sheet 2 has been proposed to promote Raf dimerization and hence its activation [51 , 52] . As shown in the present work , CaM would also attach to the region of α-helixes 4 and 5 of the globular domain of K-Ras . Interestingly , this region has recently been described as relevant for proper K-Ras dimerization , as arginine 161 forms a salt bridge when forming the dimer [53] . As stated above , according to our model the region used by CaM to interact with K-Ras may overlap with the one used to form K-Ras dimers . Moreover , not only do these interactions share the surfaces by which they interact but also certain residues used in the interaction such as arginine 161 . Thus , one can conclude that CaM’s interaction with K-Ras would most probably interfere with K-Ras dimerization , and consequently this would be another mechanism ( besides inhibiting phosphorylation [21] ) by which CaM is negatively regulating K-Ras-Raf-ERK signaling . All in all , we can affirm that our simulations ( and later experimental validation ) propose a reliable model in which residues R135 , R161 and R164 play a significant role in the interaction of the globular domain of K-Ras with CaM , while the polybasic domain of the HVR interacts with the acidic linker region of CaM .
The joining of the proteins was performed in several steps . In order to have a model to work with , we first examined the original structure or the NMR structure of CaM with the HIV-1 matrix peptide ( PDB code 2MGU ) . The peptide presented a certain degree of homology with the HVR of K-Ras , and we took advantage of that fact by replacing the existing peptide with a fragment of K-Ras ( residues 165 to 188 ) . To this end , the peptide was replaced using the program modeller ( https://salilab . org/modeller/ ) . The best structure generated by modeller was selected to perform the following simulations . The HIV-1 matrix peptide was replaced by the K-Ras peptide . However , the homology between the sequences did not match the real orientation of the interaction between K-Ras and CaM . Thus , once the HIV-1 matrix peptide was replaced by the K-Ras fragment , another program was developed to rotate it . This software creates a vector between two given atoms ( one from the CaM and another from the K-Ras fragment ) and increases the module ( the distance between those atoms ) . Afterwards , it performs a rotation on its axis ( rotating the whole K-Ras fragment ) and decreases the module ( diminishes the distance between the selected atoms ) . Even though several combinations were tried , the distances were finally set to 10 Å and 5Å and the rotation angle was fixed at 180° . The system was then minimized in a multi-step manner , applying the same restraints as in the simulations with K-Ras . The minimized complex was heated up to 300 k in a step wise manner , at a rate of 30 K every 20 ps . The protein backbone atoms were restrained with a force constant of 0 . 5 kcal/mol·Å . Additionally , 200 ps of simulation at constant pressure ( NPT ensemble ) were performed without any restraint in order to allow density equilibration . Then , a short MD simulation of 2 ns length within the NVT assembly was done to allow small structural readjustments . The final structure after this process was used as a reference to add the full-length oncogenic K-Ras ( mutation G12D ) ( PDB code 4DSN ) . The lacking residues ( a majoritarian part of the globular domain , residues 1–164 ) were added by merging the two systems . This step was done by superimposing the residues 165 to 168 of K-Ras and removing the leftover atoms of the K-Ras peptide . The final complex was placed in a cubic periodic box filled with TIP3P water molecules , imposing a minimal distance of 15 Å between the protein and the box walls . Water molecules closer than 2 . 2 Å were removed and neutralizing counter-ions ( sodium ions ) were added at positions of lowest electrostatic potential . Minimization was carried in a multistep procedure: 1 ) Full complex restraint , both K-Ras and CaM were restrained with a 10 kcal/mol Å constant , including GTP and calcium ions; 2 ) CaM and globular domain of K-Ras restricted while releasing the lateral chains of the HVR; 3 ) Release of the lateral chains of CaM; 4 ) Progressive release of the HVR and CaM by diminishing the restriction constants from 10 to 0 . 5 kcal/mol Å; 5 ) Progressive release of the globular domain of K-Ras ( in the same way as with HVR and CaM ) ; 6 ) Minimum restraint on all backbones ( 0 . 1 kcal/mol Å ) ; 7 ) No restraints minimization . These minimizations were performed through 5000 steps of the conjugate gradient algorithm keeping fixed the selected parts of the system fixed with the indicated restriction constants , except for the last minimization , which was carried out for 10000 steps . First , the cMD simulations were carried out . To do so , the minimized systems were heated up to 300 K in increments of 30 degrees per step of 20 ps . Afterwards , 200 ps of simulation at NPT ensemble were performed . Also , a short MD simulation of 2 ns length within the NVT assembly was carried out . The MD simulations of the systems were performed in a multi-step procedure ( each step of 10 ns ) . The temperature was regulated by using the Langevin thermostat with a collision frequency ɣ of 2 . 0 ps-1 . All bonds involving hydrogen atoms were constrained to their equilibrium value using the SHAKE algorithm , allowing the use of a 2 fs integration time step in all of the simulations . Non-bonded interactions were truncated at a cut-off of 10 Å , and long range electrostatic interactions were treated with the particle-mesh Ewald method . A total of 6 molecular dynamics simulations using different sets of initial velocities aimed at providing a better sampling [54] were performed of at least 50 ns each one . As for the sMD simulations , the initial coordinates were taken from the first 5 ns of the cMD simulation . A λ factor of 0 . 8 was applied , and a total of 4 simulations of at least 50 ns were produced . In order to determine the binding energy between the proteins , the AmberTools module of AMBER was used with Molecular Mechanics Poisson Boltzmann ( Generalized Born ) Solvation Area . To perform the calculations , structures for the conformational ensemble were extracted from the MD trajectories . The water molecules and the counter-ions were removed to obtain the topology of the systems without solvent . The calculations were performed with the following parameters . GB: GB method 5 , salt concentration 0M and surface tension value 0 . 005 . PB: Cavity offset -0 . 920 , cavity surften 0 . 00542 , external dielectric constant 80 . 0 , internal dielectric constant 1 and ionic strength 0M . CaM and K-Ras were moved closer with a restriction between their suspected regions of interaction . A mass center was created in the globular domain of K-Ras , selecting the alpha carbons of the residues 161 to 169 , and another was created in one of the lobes of CaM , selecting the alpha carbons of the residues 29 , 32 , 48 , 52 , 63 and 67 . These two mass centres were restrained to keep a certain distance , between 20 and 5 Å . A restriction constant of 5 kcal/mol Å was applied at higher or shorter distances so that the mass center would remain close to the specified separation; if the simulations tend to approach or separate the mass centers , the restriction constant corrects the deviation . The distance was analyzed every 5 frames to make sure it was maintained as stablished . The backbone of residues 161 to 169 of K-Ras were also restricted so as to maintain their mass center stable . Five nanoseconds of each distance were obtained , decreasing the distance by 0 . 5 Å at each step ( obtaining a total of 30 steps ) . Afterwards , the last coordinates of the simulations between 5 and 12 Å were obtained and a cMD of 10 ns was performed for each structure , and the interaction stability was analyzed with MMPB/GBSA analysis . Byotinilated CaM was used as the ligand and GST-K-Ras as the analyte . A Sensor Chip SA was used for these experiments . This chip has a matrix of carboxymethylated dextran pre-immobilized with streptavidin , which binds to Byotinilated CaM while K-Ras was injected as an analyte . The matrix was first activated and prepared as follows: all four channels of the chip were treated with 100 mM HCl , 50 mM NaOH , 0 . 1% SDS and then water at a flow rate of 100 μL/min: and the signal was normalized by injecting glycerol at 70% and then two cycles of running buffer ( 50 mM Tris-HCl , pH 7 . 6 , 150 mM NaCl , 0 . 05% Tween 20 , 2 mM MgCl2 and 1 mM CaCl2 ) . Then the ligand was loaded: the first channel was loaded with biocytin alone ( a biotin analogue ) , in order to have a negative control and being able to detect any unespecificities; the rest of channels were loaded with biotinylated CaM; the loading was performed at 25°C with a concentration of 20 μg/mL of biotinylated CaM solved in running buffer; the ligand was added at a constant flow rate of 5 μL/min for 20 seconds ( until 1000 resonance units ( RU ) were reached ) ; afterwards , 0 . 3μg/mL of biocytin in running buffer were injected to block the streptavidin molecules that had not bound to biotinylated CaM . Next , the analyte , the globular domain of K-Ras ( whether the WT form or the mutated forms ) , was injected at a flow rate of 5 μL/min for 1 minute . The protein was allowed to dissociate for a minute . Samples were loaded at concentrations between 0 . 0625 μM and 1μM . Last , the chips were regenerated by injecting running buffer without MgCl2/CaCl2 and supplementing with 6 mM EDTA , to obtain the apo form of CaM and dissociating any remains of K-Ras . The regeneration step was performed for a minute at 30 μL/min . The data obtained were analysed with the Scrubber software . Both the kinetics and the affinity were analysed: 1 ) after loading the data file , the baseline was fixed as 0 for each flow cell ( channel ) ; 2 ) the injection points were aligned for all the samples; 3 ) the region of the sensogram regarding the injection of the sample was selected and cropped; 4 ) the samples were blanked subtracting the flow cell 1 , which has not biotinylated CaM but biocytin; 5 ) the spikes generated by the injection or buffer exchange ( when injecting the regeneration solution ) were removed; 6 ) the bound section was selected as the more stable section within the injection process , usually 10 seconds before the ending of the injection; 7 ) for the affinity analysis , the amount of bound protein at each concentration of the sample was analysed to determine the KD . At least 4 concentrations were injected in each run . To obtain the purified globular domain of K-Ras , the cDNA of residues 1–166 of K-Ras4B ( either WT or with mutations in 1 or 2 bases ) were cloned into a pGEX-KG bacterial expression vector with a GST tag . BL21 ( DE3 ) cells were transformed with the plasmids . The expression of the protein was induced by adding 0 . 5mM IPTG to a 0 . 5 L culture of the transformed cells and incubating for 4 hours . Cells were harvested at 6000 r . p . m . for 10 minutes at 4°C . The supernatant was discarded and the cell pellets were stored at -80°C . After checking a proper protein expression , cell pellets were lysed with a lysis buffer ( 50 mM TrisHCl pH 7 . 5 , 100mM NaCl , 1mM EDTA , 5mM MgCl2 , 10% Glycerol , 1% Triton X-100 , 0 . 5% Β-Mercaptoethanol ) . Cells were lysed for 20 minutes on ice and then sonicated for three minutes , also on ice . The lysates were centrifuged for 10 minutes at 14000 r . p . m . at 4°C . The supernatant was collected and added to 800 μL of glutathione-sepharose beads . The mixture was incubated for 1 hour at 4°C . The supernatant was removed by centrifugation ( 3 minutes 3000 r . p . m . ) and the beads were washed with 20mL of lysis buffer four times and once with exchange buffer ( Tris-HCl pH 7 . 4 10 mM , NaCl 150 mM , MgCl2 2 mM , glycerol 10% , DTT 1 mM ) ) . Afterwards , the beads were incubated with 1 . 6 mL of exchange buffer complemented with 1mM GTP , 10mM EDTA and 1mM DTT . The mix was incubated for 30 minutes at room temperature . Magnesium chloride was added until a final concentration of 15 mM was reached , and then the beads were washed with 15 mL of exchange buffer with 2mM of MgCl2 . The resin was separated by centrifugation and 1 mL of lysis buffer supplemented with 20mM of glutathione was added . The supernatant was then collected by centrifugation . The products of the purification were confirmed by western blot . | K-Ras is one of the most mutated oncogenes in human cancer . Although several studies validate K-Ras protein as good candidate for direct therapeutic targeting , pharmacologic targeting has not been successful . During the last years increasing evidences demonstrate that oncogenic K-Ras activity can be modulated in vivo by dimerization , nanoclustering at the plasma membrane or interaction with non-effector proteins , consequently opening new therapeutic strategies . We have previously demonstrated that Calmodulin , an ubiquitous Ca2+-binding protein , is one of this K-Ras interacting proteins and that it negatively modulates K-Ras signaling . Although experimental data were available showing the relevant regions for this interaction , a model of K-Ras and Calmodulin interaction was missing . In the present work by using different computational modeling techniques we obtained a model for this interaction that agrees with the experimental data . We believe the present model will help to better understand K-Ras regulation , and to design new inhibitors . For instance , base on our model , we can predict that the interaction can take place at the plasma membrane , and that since the surface of K-Ras that interact with Calmodulin is the same that it uses for dimerization , that Calmodulin could be inhibiting K-Ras dimerization . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"protein",
"interactions",
"bioelectrochemical",
"analysis",
"molecular",
"dynamics",
"biochemical",
"analysis",
"simulation",
"and",
"modeling",
"membrane",
"proteins",
"bioassays",
"and",
"physiological",
"analysis",
"cellular",
"structures",
"and",
"organelles",
"researc... | 2018 | Modeling and subtleties of K-Ras and Calmodulin interaction |
In placental mammals , inactivation of one of the X chromosomes in female cells ensures sex chromosome dosage compensation . The 17 kb non-coding Xist RNA is crucial to this process and accumulates on the future inactive X chromosome . The most conserved Xist RNA region , the A region , contains eight or nine repeats separated by U-rich spacers . It is implicated in the recruitment of late inactivated X genes to the silencing compartment and likely in the recruitment of complex PRC2 . Little is known about the structure of the A region and more generally about Xist RNA structure . Knowledge of its structure is restricted to an NMR study of a single A repeat element . Our study is the first experimental analysis of the structure of the entire A region in solution . By the use of chemical and enzymatic probes and FRET experiments , using oligonucleotides carrying fluorescent dyes , we resolved problems linked to sequence redundancies and established a 2-D structure for the A region that contains two long stem-loop structures each including four repeats . Interactions formed between repeats and between repeats and spacers stabilize these structures . Conservation of the spacer terminal sequences allows formation of such structures in all sequenced Xist RNAs . By combination of RNP affinity chromatography , immunoprecipitation assays , mass spectrometry , and Western blot analysis , we demonstrate that the A region can associate with components of the PRC2 complex in mouse ES cell nuclear extracts . Whilst a single four-repeat motif is able to associate with components of this complex , recruitment of Suz12 is clearly more efficient when the entire A region is present . Our data with their emphasis on the importance of inter-repeat pairing change fundamentally our conception of the 2-D structure of the A region of Xist RNA and support its possible implication in recruitment of the PRC2 complex .
In mammals , the transcriptional silencing of one of the two X chromosomes in female cells ( X chromosome inactivation , XCI ) ensures sex chromosome dosage compensation [1] . Once acquired early in development , the inactivated state is faithfully inherited through successive cell divisions . XCI initiation is associated with increased Xist RNA transcription . Whilst first retained near its transcription site , Xist RNA then spreads along the entire X chromosome from which it has been transcribed [2]–[5] whilst , a series of epigenetic marks , which include the repressive histone modifications H3K27me3 , H3K9me3 , are recruited to the presumptive inactive X chromosome . Xist RNA is a long non-coding RNA ( 17 kb in length in the mouse ) , which is capped , spliced , and polyadenylated . Little is known about its structure and mechanism of action . The Xist gene has a complex origin . It includes degenerated pieces of an ancient protein gene Lnx3 as well as genomic repeat elements derived at least in part from transposon integration events [6] , [7] . The most conserved Xist RNA regions correspond to repeat elements ( denoted A to E in mouse [8] ) , which are organized as tandem arrays . The A region ( positions 292 to 713 in mouse , accession no . gi|37704378|ref|NR_001463 . 2| [2] , and 350 to 770 in human , accession no . gi|340393|gb|M97168 . 1| [5] ) is the most highly conserved of the repeat regions and is critical for initiation of XCI . The observation that female mouse embryos carrying a mutated XistΔA gene inherited from males are selectively lost during embryogenesis underlines the importance of this element [9] . Recent data have shown that an early event in silencing is the formation of a Xist RNA compartment and that the A region whilst not necessary for formation of this compartment is needed for relocation of X linked genes into this territory [10] . Over-expression of a XistΔA RNA in transgenic mouse ES cells indicates that the A region whilst not necessary for Xist coating is implicated in the recruitment of the PRC2 complex [11]–[16] . The PRC2 complex contains the Suz12 , Eed , Ezh2 , and Rbap46–48 proteins [17] , [18] . Eed and Suz12 have been proposed to bind nucleic acids [19] , [20] , whereas Rbap46–48 may interact with nucleosome protein components [17] . Lysine 27 tri-methylation of histone H3 is catalysed by Ezh2 [12] , [14] and both Eed and Suz12 are required for this activity [20] . Recently a short 1 , 600 nucleotide-long RNA which contains the A region at its 5′ extremity was suggested to be expressed early in XCI initiation and to bind the PRC2 complex [21] . Since the function of Xist RNA is expected to depend on its 2-D structure , studies aimed at establishing the 2-D structure of the Xist A region have considerable interest . Based on nucleotide sequence of the A region and computer prediction , Wutz and colleagues have proposed that each repeat forms two short stem-loop structures [11] . Recent NMR analysis confirmed that one of these stem-loop structures can be formed in vitro by an RNA molecule bearing a single copy of the mouse repeat A sequences [22] . In Xist RNA , the repeat sequences are , however , separated by long spacer regions ( 21 to 48 nt long for mouse ) . Since current models fail to take account of this sequence complexity , an experimental analysis of the entire A region was thought likely to provide valuable information on the structure of the A region . As conventional probing experiments are , however , hindered by the presence of the repeated sequences and long U tracks , we applied a combined approach exploiting both chemical and enzymatic probing of RNA structure in solution and FRET experiments using fluorescent oligonucleotide probes complementary to different parts of the A region . Using this dual approach , we could show that repeats in the A region interact with each other to form long irregular stem-loop structures . Such inter-repeat interactions appear to be required for the binding of the various components of the PRC2 complex . We identified the minimal number of repeats necessary for such binding . The implications of our results within the wider context of X-inactivation and of the XCI mechanism ( s ) underlying silencing are discussed .
We analysed in parallel both the entire mouse and human A regions , as the sequence divergence of the inter-repeat linking sequence between mouse and human was expected to provide insight into how the spacer regions might influence the A repeat structure . The specific primers for the two A regions are listed in Table S1 . To test if the A region interacts with neighbouring Xist RNA sequences , an RNA that contained only the mouse A region ( positions 277 to 760 in mouse Xist RNA ) and a larger RNA including the sequence extending from positions 1 to 1137 were studied in parallel by limited enzymatic digestion . Very similar digestion profiles ( Figure 1A and B ) were obtained for the two RNAs when digestions were performed on the T7 RNA transcripts after folding under the conditions outlined in Materials and Methods . We conclude that the A region probably folds on itself without major interaction with other upstream and downstream Xist sequences . Hence , our subsequent analyses of the 2-D structure of the Xist A region were carried out in the absence of flanking sequences ( positions 227 to 760 and 330 to 796 for the mouse and human A regions , respectively ) . Each enzymatic digestion and chemical modification assay was carried out in duplicate using different transcript preparations and each extension analysis repeated 2 or 3 times for each primer . Representative examples of the primer extension analyses are provided in Figure 1 and Figure S1A–S1E . The structure proposed by Wutz and colleagues , in which each of the repeats fold into a double stem-loop structure , could not explain the numerous V1 RNase cleavages that we observed with both the mouse and human A regions ( Figure 2 ) [11] . We explored the possibility that each repeat folds into a unique longer stem-loop structure . Such folding was similarly unable to explain V1 RNase cleavages ( Figure S2 ) . We conclude that the 2-D structure may involve interactions between repeats and spacers and inter-repeat interactions . There is , however , a multitude of potential ways for duplex formation between repeats ( Figures 3–5 , models 1–3 ) . Our design of the putative structure was orientated by the detection of six successive strong V1 RNase cleavages in the central poly A sequence ( positions 550 to 555 ) , suggesting the involvement of this segment in a helical structure . The strong modification by DMS of a sequence immediately downstream ( positions 555 to 561 ) was an indication for a single-stranded state . One possible explanation for these data was the formation of a central stem-loop structure called SLS2M with a U track on one strand and an A track on the other strand ( Figure 5 ) . Formation of this central stem-loop structure was subsequently imposed as a constraint when exploring the possible folding of the mouse A region . This excluded structures in which two successive repeats would interact with each other ( Model 1 , Figure 3 ) , since in this case , the entire poly A region would interact with the poly U track located upstream of repeat 3 , which was not in agreement with the probing data ( Figure 3 ) . Another possible structure involved formation of an interaction between the 5′ and 3′ halves of the A region . This would generate a very long irregular stem-loop structure with an A rich terminal loop ( Model 2 , Figure 4 ) . An alternative structure involved the folding on themselves of the 5′ and 3′ parts of the A region , with SLS2M in between ( Model 3 , Figure 5 ) . Several other alternative pairings of the repeats were also explored—none fitted the chemical and enzymatic data perfectly . The notion of independent folding of the 5′ part of the A region ( positions 318 to 521 in mouse RNA ) was supported by M-fold analysis of this segment , which identified a highly stable long irregular stem-loop structure , SLS1M ( ΔG = −41 . 96 kcal/mol at 0°C in 3 M NaCl ) , in which repeat 1 interacts with repeat 4 and repeat 2 interacts with repeat 3 . It is the most thermodynamically stable structure proposed for this 5′ segment and was predicted irrespective of whether the experimental data were introduced as a constraint in the M-fold search . In SLS1M , each repeat interacts both with another repeat and with a spacer segment , increasing the stability of the overall structure . Similarly , M-fold analysis of the 3′ part of the mouse A region suggested that repeat 5 interacts with repeat 8 and a spacer region and repeat 6 with repeat 7 and a spacer region . The resulting SLS3M structure was predicted as the most favourable structure by M-fold when the experimental data were incorporated as a constraint . The overall predicted three stem-loop structure ( Model 3 ) has a low calculated free energy ( −77 . 76 kcal/mol ) and has a better fit to the experimental data than Models 1 and 2 ( Figures 3 and 4 ) , suggesting that , in solution , Model 3 is the most likely structure among the numerous possibilities . If a structure has biological relevance , it is generally conserved throughout evolution . Therefore , we tested whether the most favourable structures identified for the mouse A region were relevant to the human A region in solution . The sequence of the human A region differs from that of mouse by the presence of an additional repeat 5 and the absence of a long central polyA region . Experimental data ( Figures 6 and 7 and Figure S3A–E ) suggested that the central repeat 5 forms a central stem-loop structure ( SLS2H ) . Based on this , structures similar to mouse Models 2 and 3 could be proposed for the human A region which involve either a long irregular stem-loop structure including all the repeats ( Model 2 ) or a three stem-loop structure ( Model 3 ) with repeats 1 to 4 forming a first stem-loop structure ( SLS1H ) , repeat 5 folded alone in a second stem-loop ( SLS2H ) and repeats 6 to 9 involved in a third stem-loop structure ( SLS3H ) . As for the mouse A region: ( i ) A 2-D structure in which each repeat interacts with its immediate downstream repeat ( repeats 1 , 3 , 6 , and 8 with repeats 2 , 4 , 7 , and 9 , respectively , Model 1 ) was not supported by the probing data ( segment 688 to 696 ) ( Figure S4 ) ; ( ii ) M-fold analysis of the 5′ portion of the human A region ( positions 370 to 530 ) either with or without the experimental data as a constraint identified SLS1H as the most stable structure ( ΔG = −42 . 70 kcal/mol ) ( Figure 7 ) ; ( iii ) SLS3H was retained as the most stable structure for the 3′ part of the A region , when the experimental data were added as a constraint to an M-fold search; and ( iv ) the 3 stem-loop structure corresponding to Model 3 ( ΔG = −86 . 6 kcal/mol ) had the best fit with probing data compared to the other 2-D models . Further support for Model 3 was provided by our observation of identical patterns of enzymatic cleavage for the entire human A region and for the isolated SLS1H portion ( Figure 6 ) . The maintenance of interactions between both spacers and repeats during mammalian evolution of the A region implies that the nucleotide sequences involved in these interactions were either conserved or subjected to compensatory base changes . This was confirmed by the alignment of the mouse , human , orangutan , baboon , lemur , dog , rabbit , cow , and elephant A region sequences ( Figure S5 ) . Nucleotide sequence conservation extends out beyond the repeats themselves for the majority of the repeats , allowing formation of the SLS1 and SLS3 structures in all sequenced Xist RNAs ( Figure S6 ) . Three oligonucleotide pairs ( P1–P5 , P2–P4 , and P6–P7 ) were retained in order to test Model 3 by FRET experiments ( Figure 8 ) . This Model predicts that the P1–P5 and P2–P4 pairs of oligonucleotides interact with the single-stranded segments which border the helix formed by repeats 1 and 4 , whilst the P6–P7 pair of oligonucleotides should interact with the single-stranded segments bordering the helix formed by repeats 5 and 8 . A marked FRET effect would therefore be expected for these three oligonucleotide pairs if the A region was folded as in Model 3 . The distance between the fluorophores of these three pairs of oligonucleotides would , on the other hand , be expected to be much larger if region A was folded as in structures 1 or 2 . Whilst tertiary structural interactions might decrease the distances , a lower level of FRET would still be expected to be observed for the three pairs of oligonucleotides if the A region was folded according to structures 1 or 2 ( Figure 8 ) . The P1 and P6 oligonucleotides bind to two single-stranded segments which flank the helix formed by repeats 1 and 8 in structure 2 . A strong FRET effect for P1 and P6 would therefore be expected if the A region was folded according to structure 2 . Upon binding to the A region , oligonucleotide P7 partially disrupts the base-pair interactions formed by the central poly A stretch . However , as similar levels of destabilization are expected for the three possible structures , binding of this oligonucleotide was not expected to favour one structure more than the two other ones . The same is true for oligonucleotide P5 that binds to the partner U stretch of the poly A sequence . To monitor the level of FRET obtained for oligonucleotides bordering a helix , we used the short R1–2 transcript containing repeats 1 and 2 and their bordering sequences , which adopt a single unique 2-D structure and the P1–P3′ oligonucleotide pair ( Figure S7 ) . Other controls exploited the P3–P6 and P3–P5 pairs , which were not expected to be in close proximity in any of the three proposed structures ( Figure 8 ) . The oligonucleotide pairs used are shown in Figure 8 , along with examples of typical fluorescence intensity spectra recorded in FRET experiments for the P2–P4 and P3–P6 pairs ( Figure 8D ) . High FRET signals in the range of 50% were obtained for the P1/P5 , P2/P4 , and P7/P6 oligonucleotide pairs , whilst lower FRET signals were observed for the P1–P6 ( 35% ) and especially the P3–P6 and P3–P5 oligonucleotide pairs ( 25% and 22% , respectively ) . This is compatible with a large part of the molecules being folded in solution into structure 3 . Folding of a large number of molecules into structure 2 would have led to a strong FRET signal for the P1–P6 pair and lower signals for the five other pairs , which was not observed . The strong FRET effects obtained for the P1/P5 , P2/P4 , and P7/P6 oligonucleotide pairs argues strongly against folding according to structure 1 . Based on our FRET data , we conclude that folding predominantly occurs according to Model 3 . Previous studies have shown that the PRC2 complex interacts with the Xi [12] , [14] , [15] , [23] and the A region has been proposed to recruit the PRC2 complex through the Ezh2 subunit , which would act as an RNA-binding subunit [21] . We wished to explore further the binding of the PRC2 complex to the A region in the light of our structural data . In particular , we were interested in determining how many A region repeats were required to bind the individual Eed , Ezh2 , RbAp46 , RbAp48 , and Suz12 components of the PRC2 complex . We initiated a proteomic approach based on affinity chromatography purification of complexes formed upon incubation of in vitro transcribed Xist A region RNAs with nuclear extracts , followed by protein identification by mass spectrometry and Western blot analysis . Mouse ES cells are a widely exploited model for the study of XCI initiation , and we reasoned that as Xist RNA acts as an initiator of XCI , proteins which have to interact with this RNA to ensure early Xist functions should already be present in the nuclear extract of ES female mouse cells prior to differentiation . We used a control RNA containing only the three MS2 protein binding sites and tested four RNAs containing different segments of the mouse Xist A region flanked by the three MS2 binding sites at their 3′ end ( Figure 9A ) . These RNAs denoted as 1R/MS2 , 2R/MS2 , 4R/MS2 , Aregion/MS2 , and HIV/MS2 contained , respectively , repeat 1 without any neighbouring sequence , repeats 3 and 4 and their bordering spacers ( positions 401 to 552 in mouse Xist RNA ) , the SLS1M stem-loop structure , the entire A region , and a fragment of HIV-1 RNA ( positions 5338 to 5514 in the BRU RNA ) used as a negative control . In order to get an idea of the proteins capable of associating with the entire A region , the proteins bound to purified complexes formed on the Aregion/MS2 RNA were analysed by mass spectrometry . Among numerous proteins detected were protein PTB and components of the PRC2 complex ( Ezh2 , RbAp46 , RbAp48 , and Suz12 ) ( Figure S8 ) . We then evaluated by Western blot experiments the relative amounts of each of the PRC2 components in RNPs formed by the various RNAs tested . Whilst Eed , Ezh2 , and PTB were detected in complexes formed on RNAs containing two or more repeats , binding of RbAp46 and RbAp48 was detected only when using RNAs with at least four repeats and Suz12 when the entire A region was used ( Figure 9C ) . The control HIV-1 RNA bound none of these proteins ( Figure 9B and 9C ) . To further explore these data , we performed a series of experiments in which fragments of the A region were transcribed in vitro as radio-labelled RNA without MS2 fusion and these RNAs were incubated with mouse ES nuclear extracts . Three distinct RNAs ( the complete A region , 4R , and 2R RNAs; Figure 9D and 9E ) were used for these experiments . In confirmation of the possible interaction of Eed with an RNA containing only two repeats , trace amounts of 2R RNA were retained on the beads when an anti-Eed antibody was used . Only complexes containing the 4R RNA or the entire A region were retained when anti-Suz12 , anti-Ezh2 , anti-RbAp46 , and anti-RbAp48 antibodies were used . These observations confirmed the importance of the corresponding regions for association of these proteins . Clearly , however , higher amounts of the entire A region compared to 4R RNA were bound when the anti-Suz12 and Ezh2 antibodies were used . We conclude that whilst some segments of the A region allow the binding of particular PRC2 components , the entire A region is required for efficient association of the entire complex .
Until now , both computer [11] and experimental analyses [22] of the possible 2-D structure of the A region of Xist RNA have privileged the individual repeat as the unit of folding . However , the presence of long intervening spacer sequences between the repeats suggests that these spacer sequences may participate in 2-D structure formation , and points to the potential inadequacy of previous models . Our detailed chemical and enzymatic probing of the A region structure in solution involving the design of specific primers for reverse transcriptase extension analysis enabled us to identify for the first time the double-stranded and single-stranded segments making up the A region structure in solution . The data obtained clearly demonstrate that the repeats do not fold on themselves but rather fold one with the other . Chemical and enzymatic probing of an RNA structure in solution often allows the building of a unique 2-D structure model in agreement with the experimental data . Studies on the A region were , however , complicated by the high degree of sequence redundancy . Use of a recently proposed biophysical approach , based on FRET assays [24] , helped overcome these difficulties by providing information on the relative distances between the sequences flanking the various repeats . To our knowledge , up to now , this approach has only been used to confirm the 2-D structure model of telomerase RNA [24] . This method , which involves the utilization of oligonucleotides carrying donor and acceptor fluorescent dyes complementary to single-stranded segments in the studied RNA , proved particularly well suited to the study of the A region , since our probing data identified several long single-stranded segments which were able to bind the oligonucleotide probes . Among the possible 2-D structures for the A region , only one , structure 3 , showed perfect agreement with the FRET data . Structure 3 , which contains two long irregular stem-loop structures , each involving four repeats ( four-repeat structure ) , also shows the best agreement with the chemical and enzymatic data . The two long stem-loop structures are separated by a short stem-loop structure , corresponding to a divergent region between mouse and human Xist RNAs . One repeat in this segment is common to all the sequenced Xist RNAs ( Figure S5 ) , except mouse RNA . In the latter , it is replaced by a poly A sequence forming a short stem loop with a poly U sequence . Interestingly , nucleotide sequence conservation in the A region extends to the spacer extremities , which contribute to the possibility of forming the four-repeat structures ( Figure S5 ) . Although the presence of large internal and terminal loops decreases the stability of stem-loop structures , the predicted free energies of the two four-repeat structures in both mouse and human RNAs have strongly negative values ( between −33 and −45 kcal/mol ) , explaining why they are proposed by the M-fold software . In addition , these four-repeat structures may be stabilized in cellulo by RNA-protein interactions . Interestingly , protein PTB , which contains 4 RNA recognition motifs ( RRMs ) , which are each able to interact with UCUU ( C ) , UUCUCU , or CUCUCU sequences , showed high affinity for the A region in nuclear extract binding experiments ( Figure 9 ) [25] . As UCUU motifs are present on each side of the large internal loops in the four-repeat structures and in the terminal loop , one might imagine that interactions of the RRMs of a single PTB molecule with these various segments may stabilize the four-repeat structure as suggested by previously proposed models for protein PTB-RNA interaction [26] . In spite that model 3 has the best fit with all the data compared to other models , this does not exclude the possibility of some local dynamic in small areas of the SLS1 and SLS3 structures . More precisely , the instability of a few base-pair interactions can explain the presence of both V1 RNase cleavages and chemical modifications in a limited number of very small segments of the A repeat region . Our adaptation of the affinity purification chromatography , originally developed for purifying spliceosome complexes [27] , to complexes formed upon incubation of different fragments of the A region with nuclear extracts prepared from undifferentiated mouse ES cells , coupled with mass spectrometry and Western blot analyses , was powerful . Together with immunoselection assays performed on assembled RNP complexes , it revealed the capability of four components of the PRC2 complex to associate with an RNA corresponding to one of the four-repeat structures formed by the A region . However , our observation that the entire A region is needed for efficient association of the Suz12 protein suggests a putative additional functional role for the entire A region in either binding Suz12 or in stabilising the binding of Suz12 to the four-repeat structure . This is too early to give a convincing molecular explanation of this observation . Further experiments are needed to understand why Suz12 displays different association properties compared to other members of the PRC2 complex . Whilst UV cross-linking of the RNP complexes formed with ES nuclear extract using the entire A region has confirmed the direct binding of PTB to this RNA region , direct binding of components of the PRC2 complex was not detected ( unpublished data ) . Neither Ezh2 nor Eed , which were previously proposed to be recruited by Xist in an A region dependent manner [21] , were cross-linked in significant amounts , suggesting that their association with the A region is mediated via association with other nuclear proteins . Therefore , the peculiar SLS1 and SLS3 structures in the A region may be needed to recruit nuclear proteins which have an affinity for components of the PRC2 complex or to reinforce the RNA affinity for these components . Mass spectrometry analysis of RNP complexes formed with the entire A region showed that , in addition to components of the PRC2 complex , a large number of other nuclear proteins can associate with this RNA region . In further studies , it will be important to identify which of these proteins are required for PRC2 association and which ones bind directly to the A region structure . Our finding that Suz12 requires the entire A region , or more simply more than four repeats for efficient association with the RNA , is in good agreement with the observation of Wutz and colleagues ( 2002 ) that the presence of at least 5 . 5 repeats is needed to initiate inactivation [11] . Additional support for the functional significance of the four-repeat model comes from a reworking of data obtained by Wutz and colleagues , who tested the effect of a series of mutations within the A region on XCI initiation [11] . Our structural studies show that all the variants ( XR , XSR , XCR ) classed by Wutz et al . as active are able to form the four-repeat structure , whereas the two inactive variants ( XS1 and XNX ) cannot ( Figure S9 ) . Although several data argue in favour of a major role of the four-repeat structure in A repeat activity , we cannot exclude a possible role of alternative structures , for instance in modulating A repeat activity . Although it is clear that the A region is essential for the X inactivation process , the precise role and mechanisms involved in the action of the A region remain unclear . Its deletion was shown to block silencing but not the coating of the X chromosome by Xist [11] , an observation in agreement with a possible role of the A region in PRC2 recruitment . PRC2 is needed for apposition of some , but not all , of the epigenetic marks which are specific features of silenced chromatin in general and the inactive X in particular ( methylation of histone H3 at position 27 ) [20] . The association of PRC2 with long ncRNAs before transfer of the PRC2 complex to chromatin may be a general mechanism for chromatin silencing processes that depend on long ncRNAs . Both the HOTAIR and Kcnq1ot1 long ncRNAs , which are involved in gene silencing , were recently found to bind the PRC2 complex [19] , [28] . Recruitment of PRC2 is a relatively early event in X inactivation [14] in agreement with a possible early association of this complex with Xist RNA prior to extensive Xist coating of chromatin . One could imagine that PRC2 is associated with the chromatin upon Xist coating through its interaction with proteins bound to Xist RNA . Alternatively coating of the Xist RNP may facilitate PRC2 transfer to chromatin by interaction of some of the RNP components with proteins of the chromatin structure . Lee and colleagues recently reported the existence of the 1600 nucleotide long RepA RNA carrying the A region at its 5′ extremity , which may be expressed prior to expression of the entire Xist RNA and has been reported to recruit the PRC2 complex in a very early step of XCI [21] . Independent confirmation of these findings will be of major importance to the field . Screening of the numerous other proteins that we have found to be capable of association with the entire A region by mass spectrometry for their eventual specific involvement in the recruitment of genes to the X inactivation domain [10] or other early events characterising the onset of X initiation and silencing will be of potential major importance to our understanding of X inactivation .
DNA fragments coding for the entire A regions of mouse and human Xist RNAs and their subfragments were PCR amplified using mouse or HeLa cell genomic DNA , and cloned into plasmid pUC18 under the control of a T7 promoter . RNAs were generated by run-off transcription with T7 RNA polymerase as previously described [29] . DNA templates were digested with RNAse-free DNAse I and RNA transcripts were purified on denaturing 3% to 8% polyacrylamide gels . RNA 2-D structures in solution were probed as follows [29]: 200 ng of transcripts dissolved at a 80 nM concentration in buffer D ( 20 mM Hepes-KOH , pH 7 . 9 , 100 mM KCl , 0 . 2 mM EDTA pH 8 . 0 , 0 . 5 mM DTT , 0 . 5 mM PMSF , 20% ( vol/vol ) glycerol ) were renatured by 10 min heating at 65°C , followed by slow cooling at room temperature with the addition of 1 µl of 62 . 5 mM MgCl2 to a final concentration of 3 . 25 mM MgCl2 . After a 10 min preincubation at room temperature , RNase T1 ( 0 . 02 or 0 . 0375 U/µl ) or T2 ( 0 . 025 or 0 . 0375 U/µl ) was added under conditions such that it cleaved single-stranded segments . V1 RNase ( 2 . 5×10−3 or 5×10−3 U/µl ) was used to cleave double-stranded and stacked residues . DMS ( 1 µl of a 1/4 or 1/8 ( V/V ) DMS/EtOH solution ) was employed to modify single-stranded A and C residues and CMCT ( 4 or 5 µl of a 180 mg/ml solution ) to modify single-stranded U and to a lower extent G residues . Reactions were stopped as described in [29] . Cleavage and modification positions were identified by primer extension [29] . Stable secondary structures having the best fit with experimental data were identified with the Mfold software , version 8 . 1 [30] . Probing data were introduced as a constraint in the search . Fluorescence spectra were recorded at 4°C , with an excitation wavelength of 515 nm and scanning from 500 to 750 nm ( excitation and emission bandwidth of 3 nm ) . The procedure used was derived from [24] . The RNA and Cy3-oligonucleotide were mixed at a 1∶1 molar ratio in 160 µl of 150 mM NaCl , 3 . 25 mM MgCl2 , and 15 mM Na citrate ( pH 7 . 0 ) to a final concentration ( 0 . 38 µM ) superior to the Kd , incubated at 85°C for 5 min , and slowly cooled at room temperature for 15 min . After 4 h of incubation at 4°C , the yield of oligonucleotide association was determined by electrophoresis in a non-denaturing gel . Fluorescence in the gel was measured with a Typhoon ( 9410 ) Healthcare scanner . When a satisfying yield of association was detected , the emission of the Cy3-labeled complex was measured on a flux spectrofluorometer ( SAFAS ) . Ten spectra were averaged . Then , the Cy5-labeled oligonucleotide was added at a 1∶1 molar ratio , and incubation carried out at 4°C for 4 h . Ten spectra were recorded and the Fluorescence Resonance Energy Transfer ( FRET ) for the Cy3–Cy5 pair was calculated taking into account the bound/unbound ratio of Cy3-oligonucleotide . Each FRET experiment was repeated three times using different batches of transcripts . The entire mouse A region and several fragments were cloned 3′ to a T7 promoter and 5′ to the MS2 tag present in plasmid pAdML3 [27] , [31] . Nuclear extracts from undifferentiated female ES cells ( LF2 ) were prepared according to [32] , and dialyzed against buffer D . One hundred pmol of MS2-tagged RNAs were denatured , renatured , as described above , and incubated with a 5-fold molar excess of purified MS2-MBP fusion protein [31] at 4°C for 15 min . The RNA-MS2-MBP complexes formed were incubated with amylose beads ( 40 µl , GE Healthcare ) equilibrated in buffer D for 2 h at 4°C . After three washes with 500 µl of buffer D , 1 mg of nuclear extract in 150 µl of buffer D containing 5 µM of yeast tRNAs was added . After 15 min of incubation at 4°C with constant agitation , three successive washes were performed in Buffer D and RNP complexes eluted by incubation with 80 µl of Buffer D containing 10 mM maltose ( 30 min at 4°C ) . Half of the eluted RNP complex formed with the entire A region was fractionated by 10% SDS-PAGE for mass spectrometry analyses . For all the purified RNP complexes , 10% of the eluted material was used for Western blot analysis performed according to [33] . Each lane of the SDS-PAGE was cut into 2 mm sections , and proteins submitted to in-gel trypsin digestion . Analysis of extracted peptides was performed using nano-LC-MS-MS on a CapLC capillary LC system coupled to a QTOF2 mass spectrometer ( Waters ) according to standard protocols ( Figure S8 ) . The MS/MS data were analyzed using the MASCOT 2 . 2 . 0 . algorithm ( Matrix Science ) for search against an in-house generated protein database composed of protein sequences of Rattus and Mus downloaded from UniprotKB http://beta . uniprot . org/ ( August 07 , 2008 ) and protein sequences of known contaminant proteins such as porcine trypsin and human keratins concatenated with reversed copies of all sequences . Spectra were searched with a mass tolerance of 0 . 3 Da for MS and MS/MS data , allowing a maximum of 1 missed cleavage with trypsin and with carbamidomethylation of cysteines , oxidation of methionines , and N-acetyl protein specified as variable modifications . Protein identifications were validated when one peptide had a Mascot ion score above 35 . Evaluations were performed using the peptide validation software Scaffold ( proteome Software ) . RNA transcripts were dephosphorylated , 5′-end labelled with [γ-32P]ATP ( 3 , 000 Ci/mmol ) , purified and quantified according to [34] . About 70 pmol of the RNA were denatured , renatured as described above , and incubated with 30 µg of nuclear extract for 30 min at room temperature with constant agitation . About 40 µl of Protein G-sepharose beads suspension blocked with BSA ( 2 µg ) and coated for 2 h at 4°C with 10 µl of each antibodies were incubated with the RNP complexes for 2 h at 4°C in 300 µl of immunoseletion buffer ( 150 mM NaCl , 10 mM Tris-HCl , pH 8 . 0 , NP40 0 . 1% ) . Beads were washed three times for 10 min at 4°C with 750 µl of immunoselection buffer containing 0 . 5% NP40 . RNAs were phenol extracted , ethanol precipitated , fractionated on 7% polyacrylamide gel , and analysed by autoradiography . | In placental mammal females , Xist RNA is crucial for inactivation of one of the two X chromosomes in order to maintain proper X chromosome dosage . It is known that the conserved A region of Xist RNA , which contains eight or nine repeated elements , plays an essential role in this process , however , little is known about its structure and mechanism of action . By using chemical and enzymatic probes , as well as FRET experiments , we performed the first experimental analysis of the solution structure of the entire Xist A region . Both mouse and human A regions were found to form two long stem-loop structures each containing four repeats . In contrast to previous predictions , interactions take place both between repeats and between repeats and spacers . Affinity-purification of RNA-protein complexes formed by incubation of RNA in mouse ES cell nuclear extract , followed by mass spectrometry and antibody-based analyses of their protein contents , showed that the isolated 4-repeat structures from the A region can recruit components of the PRC2 complex that is needed for X chromosome inactivation . However , association of one component of this complex , Suz12 , was more efficient when the entire A region was used . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"genetics",
"and",
"genomics/epigenetics",
"molecular",
"biology/histone",
"modification",
"biochemistry/rna",
"structure"
] | 2010 | 2-D Structure of the A Region of Xist RNA and Its Implication for PRC2 Association |
Developmental genes can harbour multiple transcriptional enhancers that act simultaneously or in succession to achieve robust and precise spatiotemporal expression . However , the mechanisms underlying cooperation between cis-acting elements are poorly documented , notably in vertebrates . The mouse gene Krox20 encodes a transcription factor required for the specification of two segments ( rhombomeres ) of the developing hindbrain . In rhombomere 3 , Krox20 is subject to direct positive feedback governed by an autoregulatory enhancer , element A . In contrast , a second enhancer , element C , distant by 70 kb , is active from the initiation of transcription independent of the presence of the KROX20 protein . Here , using both enhancer knock-outs and investigations of chromatin organisation , we show that element C possesses a dual activity: besides its classical enhancer function , it is also permanently required in cis to potentiate the autoregulatory activity of element A , by increasing its chromatin accessibility . This work uncovers a novel , asymmetrical , long-range mode of cooperation between cis-acting elements that might be essential to avoid promiscuous activation of positive autoregulatory elements .
DNA cis-acting elements play key roles in the regulation and evolution of gene expression by controlling spatiotemporal transcription patterns . A major class of cis-regulatory elements are transcriptional enhancers , which can recruit combinations of transcription factors ( TFs ) to modulate transcription initiation from ( a ) cognate gene promoter ( s ) , in general independently of their relative distance and orientation [1–3] . So far , most enhancers have been functionally characterized by assay of their transcriptional activity using transgenic constructs carrying the enhancer and a reporter gene driven by a minimal promoter [4] . Another strategy consists in the random insertion of a transposon that senses enhancer activity within the surrounding genomic region . It is particularly useful to detect multiple cis-regulatory elements with similar activities and long-distance modulation of gene expression [5 , 6] . Transgenesis using BACs allows the introduction of large DNA fragments containing enhancers in their native context . This approach is helpful in the analysis of multiple enhancers controlling the same gene [7] , but can be challenging for the study of mammalian enhancer that are located far away from the promoter that they control . These different approaches provide useful information on spatial and temporal activity of the putative enhancer element , but they usually do not establish whether and how the enhancer actually participates in the control of the expression of its suspected cognate gene in its full normal genomic context . Answer to this latter question requires in vivo analyses involving deletion or mutation of the endogenous enhancer . This issue is particularly important in situations where multiple , overlapping enhancers operate within the same cis-regulatory landscape . In such cases , various types of regulatory crosstalk can occur between the enhancers , resulting in additive , synergistic , competitive or repressive effects [3] . In vertebrates , very few studies have addressed such situations . Enhancer activity is intimately linked to chromatin organization . Hence , association of pioneer TFs to an enhancer can lead to chromatin decompaction and facilitate the binding of additional TFs and/or recruitment of various epigenetic machineries [8] . In return , chromatin configuration can affect gene expression by modulating long-range interactions between enhancers and promoters [9] , that are usually constrained within regions called topologically associated domains ( TADs ) [10 , 11] . TADs , which are approximately Mb-sized in mammals , form constitutive “regulatory neighbourhoods” that provide specificity to enhancer-promoter interactions by reducing aberrant contacts between cis-regulatory elements located in distinct TADs [6 , 10] . To provide insights into the mechanisms involved in the regulation of a vertebrate gene by multiple enhancers during development , we investigated the case of the mouse Krox20/Egr2 gene [12] for which several hindbrain-specific enhancers have been identified [13] . The hindbrain is an attractive model to investigate the genetic control of morphogenesis in vertebrates , as it is subject to a transient segmentation process leading to the formation of 7–8 segments called rhombomeres ( r ) [14 , 15] . The formation and specification of segments r3 and r5 are governed by the transcription factor KROX20/EGR2 [15–17] . So far three evolutionarily conserved sequences exhibiting enhancer activity in the hindbrain have been identified within the Krox20 locus and are termed A , B and C [13] . Element A , located 217 kb upstream of the promoter in the mouse , is active in both r3 and r5 . This element carries several KROX20 binding sites and requires direct binding of the protein for its activity , suggesting that it acts as an autoregulatory element [13] . Indeed , upon deletion of element A , Krox20 expression is normally initiated in r3 and r5 , but is not amplified nor maintained at later stages [18] . Additional studies have indicated that element A underlies a positive feedback loop that acts as a binary switch for specification of odd- versus even–numbered rhombomere identity [18] . Element B , located 164 kb upstream of the promoter , drives the expression of reporter constructs specifically in r5 [13 , 19 , 20] . Finally , element C , located 144 kb upstream of the promoter , is active in the r3-r5 region [13 , 19–21] . Several observations suggest that elements B and C , in contrast to the autoregulatory element A , are involved in the initial steps of Krox20 expression in r3 or r5 ( initiator elements ) : i ) they are transcriptionally active at the early stages of Krox20 hindbrain expression [13]; ii ) they are activated by transcription factors known to act upstream of Krox20 [19–21]; iii ) they do not require the presence of the KROX20 protein for their activity [13] . In the present study , we have investigated the contribution of element C to Krox20 expression , as it was the only characterized initiator element with an activity in r3 . Using a conditional knock-out mutation of element C , we show that , unexpectedly , this element is not necessary for Krox20 initial expression in r3 . In contrast , it appears absolutely required for the maintenance of Krox20 expression in this rhombomere . This activity involves a cooperation in cis with element A , element C potentiating its activity and increasing its accessibility . These observations reveal that a cis-acting element can cooperate with other enhancers within the same locus according to different modalities and suggest a scheme for protecting autoregulatory elements from inappropriate activation .
To assess the contribution of element C to the regulation of Krox20 expression , we generated a mouse line carrying a deletion of this element . The details of the knock-out strategy are presented in Fig 1A . Two alleles were generated: Krox20Cflox , in which element C is present , but flanked by two loxP sites , and Krox20ΔC , in which element C is deleted . The impact of element C deletion on Krox20 transcription was analysed by mRNA in situ hybridization on 4 to 14 somite stage ( s ) embryos , comparing homozygous ( Krox20ΔC/ΔC ) with heterozygous mutants ( Krox20+/ΔC ) , the knock-out of one allele of Krox20 having no phenotype [15 , 16 , 22] . Unexpectedly , up until 6s the expression of Krox20 in r3 and r5 is similar in Krox20ΔC/ΔC embryos and control littermates ( Fig 1B ) . However , at 8s , Krox20 expression is severely reduced in r3 from Krox20ΔC/ΔC embryos as compared to controls and , at later stages , it is completely lost ( Fig 1B ) . During the considered period , although Krox20 expression does not appear to be dramatically affected in r5 , in contrast to r3 , the width of the corresponding domain of expression appears to be slightly reduced after 10s ( Fig 1B ) . To investigate longer-term consequences of element C deletion on cell specification , we analysed the expression of a KROX20 target gene , EphA4 [22] , which is known to persist beyond the period of Krox20 expression in r3 and r5 [18] . In control embryos ( Krox20+/ΔC ) , at 10s and 25s , EphA4 is expressed at high levels in r3 and r5 and at a lower level in r2 ( S1 Fig ) . At both stages , the r3 domain , as demarcated by EphA4 expression , is reduced in Krox20ΔC/ΔC embryos as compared to controls , whereas the r5 domain is similar in both genotypes ( S1 Fig ) . This is consistent with the premature loss of Krox20 expression in r3 , known to reduce the extension of this rhombomere [18 , 22] . The limited variation of Krox20 expression in r5 in Krox20ΔC/ΔC embryos after 10s does not appear to perturb the size of this rhombomere at later stages , consistent with the fact that Krox20 expression is not required for the maintenance of EphA4 expression in r3 and r5 [18] . In conclusion , these data indicate that i ) element C is dispensable for the initiation of Krox20 expression in r3 or r5 , suggesting the existence of other elements in charge of these functions; ii ) in r3 , in contrast , element C is absolutely required for expression beyond 6s , leading to a reduction in size of this rhombomere at later stages . Notably , the phenotype observed in r3 in Krox20ΔC/ΔC embryos is very similar to what was previously described in Krox20ΔA/ΔA embryos ( Figs 1B and 2A , and S1 Fig ) [18]; iii ) in r5 , the contribution of element C to Krox20 expression is rather limited , without significant effect on the size of this rhombomere . The similarity of the phenotypes observed in r3 upon deletion of elements A or C led us to investigate the possibility of an involvement of element C in Krox20 autoregulation , together with element A . For this purpose , we first analysed the expression of Krox20 in composite heterozygous embryos , Krox20ΔA/ΔC , carrying deletions of element A on one allele and of element C on the other ( Fig 2A ) . Although this combination does not affect Krox20 expression in r3 at early stages , at 8s Krox20 mRNA level is severely reduced and , at 12s , it is completely lost , mimicking the phenotype observed in Krox20ΔA/ΔA or Krox20ΔC/ΔC embryos at both stages ( Fig 2A ) . In r5 , Krox20 expression is only slightly affected in Krox20ΔA/ΔC embryos , similarly to Krox20ΔC/ΔC embryos ( Fig 2A ) . This defect in the maintenance of Krox20 expression in r3 , combined with apparently normal expression at early stages , contrasts with the fact that a single wild type Krox20 allele is sufficient to activate and maintain the autoregulatory loop ( Fig 2A , Krox20+/Cre ) [12] . This suggests that in Krox20ΔA/ΔC embryos the level of expression of Krox20 is not a limiting factor for the activation of the only wild type allele of element A . Therefore , the most likely explanation for the defect in Krox20 maintenance is that , in r3 , the deletion of element C impairs the activity of element A located on the same chromosome and that the two elements synergistically cooperate , in cis , for the establishment and/or maintenance of the autoregulatory loop . A more conventional , partial redundancy between elements A and C appears much less likely . This cooperation does not preclude an early involvement of element C , for instance to poise element A for the subsequent autoregulation phase . To investigate whether element C has a function only at the early phase of Krox20 activation , or whether it is required during the autoregulation phase as well , we generated a genetic condition in which element C is initially active , but is deleted at a later stage . This was achieved by combining the Krox20Cflox allele ( Fig 1A ) with a knock-in allele , Krox20Cre , in which the Krox20 coding sequence has been replaced by the coding sequence of the Cre recombinase [23] . In such embryos , Krox20 and Cre are expected to be synthetized at early somitic stages . Subsequently , the recombinase leads to deletion of element C in r3 and r5 . In Krox20Cflox/Cre embryos , Krox20 expression is progressively reduced in r3 from 6s to 10s , as compared to Krox20+/Cre controls ( Fig 2B ) , although less abruptly than in Krox20ΔC/ΔC mutants ( Figs 1B and 2A ) . At 12s , Krox20 expression is completely abolished in r3 in Krox20Cflox/Cre embryos ( Fig 2B ) . These data indicate that the presence of element C only during the early phase of Krox20 expression is not sufficient to establish and/or maintain the autoregulatory loop . The higher level of Krox20 in r3 in Krox20Cflox/Cre as compared to Krox20ΔC/ΔC embryos is likely to originate from transient activation of the loop , followed by termination of its activity , due to Cre excision of element C . In conclusion , these results indicate that elements A and C synergistically cooperate in cis for establishing and/or maintaining this loop in r3 . More precisely , they show that element C is permanently required for activity of the Krox20 feedback loop . The existence of a cooperation in cis between elements A and C led us to investigate the existence of possible physical 3D interactions between the different Krox20 cis-elements , which are separated by large distances on the mouse chromosome . A previous Hi-C analysis [11] in embryonic stem cells identified a TAD that includes the gene and its cis-regulatory elements ( Fig 3A ) . The left boundary of the TAD spreads out over a relatively large and undefined transition zone ( S2A and S2B Fig ) . To better characterize the Krox20 regulatory neighborhood , we used circular chromosome conformation capture ( 4C-seq ) on multiple viewpoints in the locus [24] . In samples prepared from total embryos at embryonic day ( E ) 9 . 5 , when Krox20 is no more transcribed [25] , the Krox20 gene and its distant regulatory element A ( separated by over 200 kb ) show highly similar distributions of 4C-seq signal ( Fig 3A and S2B and S2C Fig ) preferentially located in the Krox20 TAD . In contrast , the distribution of the Nrbf2 gene , which is located in the TAD transition zone and is separated from element A by a 35 kb genomic interval ( including a cluster of CTCF binding sites ) spreads its interactions about equally over the two neighboring TADs ( S2C Fig ) . Repositioning of the TAD boundary to the cluster of CTCF binding sites results in strongly increased separation of signal between the Nrbf2 gene on one hand and the Krox20 gene and its regulatory elements on the other hand , indicating they are located in different regulatory neighborhood ( Fig 3A and S2C Fig ) . To determine if 3D chromatin interactions in the Krox20 regulatory neighborhood were dynamic at these early stages of embryogenesis , and possibly linked to the autoregulatory loop , we compared our E9 . 5 samples to micro-dissected embryonic heads at E8 . 5 , when the autoregulatory loop is active in a subset of cells [18] . For all viewpoints , very similar patterns were obtained between E8 . 5 heads and E9 . 5 ( Fig 3B ) . At both stages , the Krox20 promoter forms long-range interactions within the Krox20 TAD that cover elements A and B ( Fig 3B ) . In addition , bi-directional interactions are formed between elements A and B and , to a lesser extent , between elements A and C ( Fig 3B ) . In conclusion , these data reveal that the Krox20 regulatory neighbourhood adopts a higher-order configuration that incorporates long-range interactions between the various cis-regulatory elements and is mostly invariant at different positions in the early embryo . To investigate the correlation between the activity of the Krox20 cis-regulatory elements and their chromatin modifications and conformation , we first performed ChIP-seq experiments [26] to analyse two histone modifications: H3K4me1 ( broad peaks covering active enhancers ) and H3K27ac ( punctuated peaks covering active enhancers and promoters ) [27] . In E8 . 5 wild type embryo heads , a number of H3K4me1 peaks were observed , including those that expectedly correspond to the A , B and C elements and to a previously known neural crest element ( NCE; Fig 4A ) [28] . The signals observed for the H3K27ac mark were low across the Krox20 locus except for the promoter ( S3A Fig ) . We can observe the same pattern of H3K27ac at the EphA4 locus with low signal at the enhancer driving its expression in r3 and r5 [29] and higher signal at the promoter ( S3B Fig ) . In contrast , a gene widely expressed at E8 . 5 in the whole embryo head , like Sox2 [30] , displays a high H3K27ac enrichment ( S3C Fig ) . The low signals observed for the Krox20 and EphA4 genes are most likely due to the limited number of Krox20-expressing cells in the sample . To overcome this limitation , we performed micro-dissections and assessed chromatin structure by ATAC-seq [31] , a technique that requires much lower cell numbers ( a few thousand ) . Enhancer activity is usually associated with increased local chromatin accessibility . E8 . 5 ( 8s-10s stage ) or E9 . 5 embryos were dissected to isolate samples derived from three regions: an anterior region , obtained by cutting within r2 and r4 ( r3 sample ) ; an intermediate region , for which cutting was performed within r4 and r6 ( r5 sample ) , and a posterior region , for which cutting was performed within r6 and r7-8 ( posterior sample ) . We observed peaks of accessibility at the level of the promoter at both stages and in all of the regional samples ( Fig 4A ) . At both stages , element A was specifically accessible in the r3 and r5 samples , but not in the posterior sample ( Fig 4A and 4B ) , in accordance with its activity restricted to r3 and r5 [13] . At E8 . 5 , element B was compacted in the r3 sample , but highly accessible in the r5 and posterior samples ( Fig 4A and 4B ) . This accessibility largely decreased at E9 . 5 ( Fig 4A ) . The limited accessibility of element B in r3 is in agreement with its lack of activity in this rhombomere [13] . Finally , element C was particularly accessible in the r3 and r5 samples at E8 . 5 ( Fig 4A and 4B ) , consistent with its activity that spans the r3-r5 region [13] . This accessibility was only maintained in the r5 sample at E9 . 5 ( Fig 4A ) . The pattern of chromatin accessibility observed in our ATAC-seq experiments revealed additional potential enhancers involved in the regulation of the Krox20 gene in the hindbrain . Indeed , we have identified an element located 107 kb downstream to Krox20 with high accessibility at E8 . 5 ( S3A Fig ) . We have tested the transcriptional activity of this new element ( NE ) by transgenesis in the zebrafish by cloning it upstream of a GFP reporter gene . In a transgenic line , this element drives specific expression in r3 at the time of the initiation of Krox20 expression in this rhombomere ( S4 Fig ) . These data raise the possibility that element NE might be the missing element involved in the initiation of Krox20 expression in r3 , although its activity still needs to be verified in the mouse . In conclusion , this analysis reveals that the patterns of accessibility of the different known elements largely correlate with their enhancer activities and helped us to identify a novel candidate element for the regulation of Krox20 expression in r3 . A final step was to investigate the effects of enhancer deletions on the accessibility of the other elements . Deletion of element A did not significantly affect the accessibility of elements B or C in any samples ( Fig 4B and 4C and S3A Fig ) . In contrast , deletion of element C significantly reduced the accessibilities of element A in r3 and of element B in r5 ( Fig 4B and 4C and S3A Fig ) . These data establish that element C has the capacity to specifically modulate the accessibility of elements A and B and therefore probably their activities . They may provide a mechanism for the involvement of the late activity of element C in the control of Krox20 autoregulation governed by element A . Furthermore , this analysis establishes the existence of an asymmetry in the relationship between elements A and C: whereas element C affects A accessibility and presumably potentiate its activity , the reverse is not true .
Previous analyses had suggested a rather straightforward mode of regulation of the Krox20 gene . Element C was responsible for the initiation of its expression in r3 , whereas element B , possibly together with element C , was in charge of the initiation in r5 . Subsequently , element A governed the maintenance of the expression through a positive feedback loop towards a definitive engagement into an odd-numbered rhombomere fate [13 , 18 , 19] . Knock-out analysis of element C now leads to major revisions of this scenario . Despite the early r3 enhancer activity of element C , as demonstrated by transgenic experiments in mouse and fish [13 , 32] , the deletion of mouse element C does not affect early Krox20 expression in r3 ( Fig 1B ) . This suggests that another cis-acting element contributes to this expression . This is not element B , which is only active in r5 , as revealed in transgenic experiments [13] , nor element A , which is absolutely dependent on the presence of the KROX20 protein [13 , 18] . Therefore , another enhancer , active in r3 and not dependent on KROX20 , must be involved . Indeed , the identification of the NE element , fully accessible at early time in the hindbrain and specifically active in r3 , makes it an attractive candidate for being involved in the initiation of Krox20 expression in this rhombomere ( S3A Fig and S4 Fig ) . The absence of phenotype in Krox20ΔC/ΔC embryos during the early phase of Krox20 expression does not preclude a role for element C during this phase . In support of this idea , enhancer activity of element C in r3 is dependent on the binding of Meis and Hox/Pbx factors [21] , as well as of the Sp5 factor mediating FGF signalling [19 , 20] , factors that are precisely known to act upstream of Krox20 in r3 [33–39] . It is therefore possible that elements C and NE cooperate in a redundant manner ( S4 Fig ) and further analyses will be required to determine whether this is indeed the case . Several examples of redundancy have been reported for enhancers governing the expression of developmental genes [3 , 40–42] . Redundant enhancers , or shadow enhancers , often share the same regulatory logic , since their activities have to be , at least in part , concomitant [43] . The analysis of the characteristics of the NE enhancer should reveal whether it depends on the same TFs as element C . In a few cases of redundant cis-acting elements that have been investigated in detail so far , it has been shown that redundancy provides robustness to the system and that , in specific genetic or environmental conditions , phenotypes can be revealed in absence of one of the elements [44] . Our study also revealed an unexpected function of element C: it is required for autoregulation , which was thought to be only dependent on element A . Using genetic approaches , we showed that an interaction must occur in cis between the two elements and that it is permanently required during the autoregulatory phase . ATAC-seq experiments indicated that element C is likely to act by modulating the accessibility of element A . Therefore , simultaneous to its classical enhancer function , element C performs another type of activity , which we propose to name enhancer potentiator . Potentiator characteristics ( asymmetrical interaction , permanent requirement , long-range ) clearly distinguishes this function from classical enhancer cooperative activities ( additive , synergistic ) and possibly from other hierarchical logic modes of interactions [3 , 45] . At this point , it is not known whether this additional enhancer potentiator function of element C , which is functionally distinct from its classical enhancer activity , is dependent on enhancer activity . Several recent studies have shown that enhancers can be transcribed and that the products of this transcription can act locally in cis to promote the expression of the target gene [46] . It is possible that such a mechanism could be involved in the potentiator activity of element C . It is interesting to note that Nlz factors , which are likely to repress Krox20 expression by reducing the accessibility of the KROX20 protein to element A [18] , are also involved in repressing element C [32] . It is therefore possible that Nlz factors only indirectly affect accessibility of KROX20 on element A , by modulating the potentiator activity of element C . In r5 , Krox20 is almost normally expressed in the absence of element C , suggesting that cooperation between elements A and C is not essential in this rhombomere to support element A activity . It is possible that element B , which is likely to constitute the major initiator element in r5 and physically interacts with element A ( Fig 3B ) performs a dual function similar to element C and potentiates the activity of element A in this rhombomere , in addition to its classical enhancer activity . Analyses by Hi-C in embryonic stem cells [11] and by 4C-seq in various embryonic samples ( this report ) revealed the existence of a regulatory neighbourhood that contains interactions between the Krox20 promoter and element A , irrespective of the considered stages or regions of the embryo ( Fig 3 ) . This chromatin configuration might therefore constitute a permissive environment for the activation of the autoregulatory loop . Such a situation , in which an autoregulatory element might be only dependent on the presence of its cognate TF for its activity , would be rather dangerous for an organism , as any transcription of the TF gene , even illegitimate , could end up activating the feedback loop and lead to high-level and long-term expression of the gene . This danger would be increased by the stochastic nature of the activation of the autoregulatory loop [19] . Furthermore , developmental genes may possess several positive autoregulatory enhancers that have to function at specific stages or in different tissues . This is exemplified by the Krox20 gene , that has at least three of such elements and that are differentially active in r3/r5 , the neural crest and developing bones [13 , 28] . Therefore , mechanisms must exist as well to prevent the inappropriate activation of these elements in the other embryonic tissues . Indeed , we have shown that ectopic expression of exogenous Krox20 in the entire zebrafish embryo only leads to activation of the autoregulatory loop in the r2-r6 region of the hindbrain [18] . The introduction of an enhancer potentiator within a positive feedback loop may constitute an efficient prevention ( safety lock ) against inappropriate activation of autoregulatory elements . According to our model ( Fig 5A ) , in the ground state , the autoregulatory element ( element A in the case of Krox20 ) is locked in a configuration that is not accessible to its cognate TF and therefore unable to activate transcription , despite its possible interaction with the promoter . This lock will be released when the potentiator function is provided by a second cis-acting element ( element C ) . It is interesting to note that in transgenic constructs , element A is able to activate a promoter in the absence of element C in cis . This difference in behaviour might be explained by two types of reasons: in transgenic constructs , element A is very close to the promoter , in contrast to the endogenous enhancer , located far upstream to the promoter; the chromatin context of a transgene is likely to be different , possibly more permissive than that of a highly regulated endogenous locus . In the endogenous locus , if the unlocker element is also responsible for the early accumulation of the cognate TF , through its classical enhancer activity , the autoregulatory element will be placed under the control of the upstream factors regulating the initial expression of the TF . In this way , the asymmetrical cooperation between the two cis-acting elements becomes essential for establishing the appropriate specificity of the positive feedback loop . As indicated in the model , such a feedback loop can be broken by mutation of either the TF or of any of the two cis-acting elements ( Fig 5B ) .
All animal experiments were performed in accordance with the guidelines of the council of European Union directive n°2010/63/UE and were approved by the "Comité d'éthique pour l'expérimentation animale Charles Darwin" ( Project Number: CE5/2012/120 ) . All mouse lines were maintained in a mixed C57BL6/DBA2 background . We used the following alleles: Krox20Cre [23] and Krox20ΔA [18]; the mouse Krox20Cflox line was generated at the Institut Clinique de la Souris ( Illkirch , France ) by homologous recombination in ES cells; the Krox20ΔC allele was obtained as described in Fig 1 , using the maternally expressed PGK-Cre transgene as deletor [47] . In situ hybridizations were performed on whole embryos as previously described [48] , with the following digoxigenin-labelled riboprobe: Krox20 [49] and EphA4 [50] . ChIP-seq experiments were performed as previously described [51] . Briefly , 10 embryos at E9 . 5 or 20 embryos at E8 . 5 were dissected in cold PBS . Cell suspensions were obtained by passing them through a 21G needle fitted onto a 5ml syringe . The cells were cross-linked with 1% formaldehyde for 10 min and washed twice in PBS , 1 mM PMSF , 1 X PIC ( Protease Inhibitor Cocktail ) . Sonication was performed on a Covaris S220 using the following programme: duty factor = 10/5 , peak incident power = 140 , cycles per burst = 200 during 600/480 seconds . 5–10 μg of chromatin was used for each IP using 3 μg of the following antibodies: anti H3K4me1 ( C15410037 , Diagenode ) and anti H3K27ac ( ab4729 , Abcam ) in RIPA buffer . The libraries were prepared with the MicroPlex Library Preparation kit ( Diagenode , E8 . 5 embryos ) and with the NEXTflex ChIP-Seq Kit ( Bioo Scientific , E9 . 5 embryos ) . ChIP Seq experiments involved biological duplicates . Sequencing was performed on multiplexed samples using 50 bp single-end reads on an Illumina HiSeq system ( E9 . 5 embryos ) or using 42 bp paired-end reads on an Illumina NextSeq ( E8 . 5 embryos ) according to the manufacturer’s specifications . Chip-seq data were analysed as follows , using Eoulsan [52] with extended support for ChIP-seq workflows ( https://github . com/GenomicParisCentre/eoulsan/tree/branch-chip-seq ) . First , reads were filtered out when they would not pass Illumina filters ( module filterreads with option illuminaid ) . Files corresponding to technical replicates were merged ( module technicalreplicatemerger , with option format = fastq ) , followed by trimming of the reads using Trim Galore ! ( http://www . bioinformatics . babraham . ac . uk/projects/trim_galore/; version 0 . 4 . 1 in module trimadapt , with cutadapt v1 . 8 . 1 and options: length = 41 , quality = 20 , error = 0 . 1 , stringency = 8 , and is . paired = yes for E8 . 5 data and is . paired = no for E9 . 5 data ) . Mapping was performed using STAR [53] ( version 2 . 4 . 0k in module mapreads with mapper . arguments = “—outFilterMultimapNmax 1—outFilterMismatchNmax 999—outFilterMismatchNoverLmax 0 . 06—alignIntronMax 1—alignEndsType EndToEnd—alignMatesGapMax 2000—outSAMunmapped Within” ) . Further filters were applied to the mapped reads before conversion into BAM ( module filtersam with removeunmapped = true; module sortsam; module rmdupgalax with is_sort = true; module sam2bam with compression . level = 5 ) . BIGWIG files were created from the resulting BAM files using deepTools’ bamCoverage [54] ( version 1 . 6 . 0 , with options:—binSize 1—normalizeTo1x 200000000—fragmentLength 200—outFileFormat bigwig ) . 4C-seq libraries were constructed as previously described [55] with small adjustments . 35 E9 . 5 embryos ( for each biological duplicate ) or 250 E8 . 5 embryos were dissected in cold PBS , followed by dissociation in collagenase type I ( Gibco ) . DpnII ( New England Biolabs , Ipswich , MA ) was used as the primary restriction enzyme and NlaIII ( New England Biolabs ) was used as the secondary restriction enzyme . For each viewpoint , up to 600 ng of each 4C-seq library were amplified using 16 individual PCR reactions with inverse primers including Illumina adapter sequences ( S1 Table ) . Illumina sequencing was performed on multiplexed samples , containing PCR amplified material of up to 7 viewpoints , using 100 bp single-end reads on the Illumina HiSeq system according to the manufacturer's specifications at the iGE3 Genomics Platform of the University of Geneva ( Switzerland ) . Reads were sorted , aligned , translated to restriction fragments and smoothed ( 11 fragments running mean ) using the 4C-seq pipeline of the BBCF HTSstation [56] according to ENSEMBL Mouse assembly NCBIM37 ( mm9 ) . For the calculation of the 4C-seq signal distribution , reads were normalized to the entire chromosome 10 , based on an approach adapted from [http://www . ncbi . nlm . nih . gov/pubmed/25959774] . For visualizations , smoothed 4C-seq reads were normalized to the 5 TADs surrounding the Krox20 locus ( chr10:62 , 880 , 000–70 , 720 , 000 ) . Position of mouse TADs in ES cells and associated 40 kb normalized Hi-C matrices [11] were obtained from http://promoter . bx . psu . edu/hi-c/download . html . Directionality indexes were calculated as described using tools described previously [7] . Interaction matrices are visualized using standard cut-offs . ATAC experiments were performed according to Buenrostros and colleagues [31] , using a homemade transposome [57] . 7–8 embryos at 10-12s were dissected in cold PBS for each genotype and cells were mechanically dissociated . Biological duplicates were performed for ATAC experiments . Cells were lysed before transposition using 1 μl of transposome and purified using a Qiagen MinElute Kit with 10 μl of Elution Buffer . Transposed DNA was amplified by PCR as previously described [57] and quantified by qPCR using 5 μl of PCR products . The number of additional cycles was determined by plotting linear Rn versus cycle and corresponded to one third of the maximum fluorescence intensity . The remaining PCR products ( 45 μl ) were treated with the additional number of cycles . The final product was purified with Qiagen PCR Cleanup Kit and eluted in 20 μl Elution Buffer . Sequencing was performed on multiplexed samples using 42 bp paired-end reads on an Illumina NextSeq according to the manufacturer’s specifications . For computational analysis , paired-end reads were mapped onto the mouse genome assembly mm9 , using STAR ( outFilterMultimapNmax 1; outFilterMismatchNmax 999; outFilterMismatchNoverLmax 0 . 06; alignIntronMax 1; alignEndsType EndToEnd; alignMatesGapMax 2000 ) . Duplicates reads were removed using Picard ( http://picard . sourceforge . net ) ( MarkDuplicates , REMOVE_DUPLICATES = true ) . To consider only fragments coming from transcription factors protected DNA ( and not from nucleosomes ) , only fragment with size lower than 100 bp were kept . Bigwig tracks were obtained using DeepTools BamCoverage ( 1 . 5 . 9 . 1 ) . Peak calling was performed using MACS2 ( 2 . 1 . 0 . 20140616 ) , using a q-value< = 0 . 01 threshold ( other parameters as default ) . For quantification , we first defined a set of non-redundant enriched regions for all samples by taking the union of all peak summits from both replicates of all samples , grouped together all summits distant from less than 50 bp , and for each group kept only the summit with the lowest q-value ( calculated by MACS2 ) . We then quantified the signal at all summits in each sample by counting the number of fragments ( using the R bioconductor package csaw , v . 1 . 0 . 7 ) . Normalisation and statistical analysis were performed using the bioconductor DESeq2 package ( 1 . 6 . 3 ) . Library size factors were calculated on fragment counts in genomic bins of 10 kb . Comparison between wild type , Krox20ΔC/ΔC and Krox20ΔA/ΔA embryos was performed using negative binomial Wald Test ( DESeq2 ) . The data have been deposited in the Gene Expression Omnibus ( GEO ) under accession number GSE94716 and is available at the following address: https://www . ncbi . nlm . nih . gov/geo/query/acc . cgi ? acc=GSE94716 | The formation of multicellular organisms from the egg to the adult stage is largely under genetic control . The activation of specific genes is governed by regulatory DNA sequences present nearby on the chromosome . Most of these sequences promote activation and are called enhancers . In this paper , we study two enhancers governing the expression of a gene involved in the formation of the posterior brain in vertebrates . One of these enhancers is involved in a positive feedback loop: it is itself activated by the protein product of the gene that it regulates . The other enhancer was thought to be only involved in the initial accumulation of the protein , necessary for the subsequent activation of the feedback loop . Here we show that the second enhancer directly cooperates with the autoregulatory enhancer to increase its accessibility and its activity . Our work uncovers a novel , long-range mode of cooperation between enhancers that restricts the domain of action of autoregulatory enhancers within embryos and might be essential to avoid their inappropriate activation . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"embryology",
"developmental",
"biology",
"genetic",
"loci",
"enhancer",
"elements",
"cell",
"biology",
"chromosome",
"biology",
"gene",
"expression",
"genetics",
"gene",
"regulation",
"biology",
"and",
"life",
"sciences",
"epigenetics",
"embryos",
"genomics",
"mammalia... | 2017 | Krox20 hindbrain regulation incorporates multiple modes of cooperation between cis-acting elements |
Arboviral infections are a public health concern and an escalating problem worldwide . Estimating the burden of these diseases represents a major challenge that is complicated by the large number of unapparent infections , especially those of dengue fever . Serological surveys are thus required to identify the distribution of these diseases and measure their impact . Therefore , we undertook a scoping review of the literature to describe and summarize epidemiological practices , findings and insights related to seroprevalence studies of dengue , chikungunya and Zika virus , which have rapidly expanded across the globe in recent years . Relevant studies were retrieved through a literature search of MEDLINE , WHOLIS , Lilacs , SciELO and Scopus ( 2000 to 2018 ) . In total , 1389 publications were identified . Studies addressing the seroprevalence of dengue , chikungunya and/or Zika written in English or French and meeting the inclusion and exclusion criteria were included . In total , 147 studies were included , from which 185 data points were retrieved , as some studies used several different samples . Most of the studies were exclusively conducted on dengue ( 66 . 5% ) , but 16% were exclusively conducted on chikungunya , and 7 were exclusively conducted on Zika; the remainder were conducted on multiple arboviruses . A wide range of designs were applied , but most studies were conducted in the general population ( 39% ) and in households ( 41% ) . Although several assays were used , enzyme-linked immunosorbent assays ( ELISAs ) were the predominant test used ( 77% ) . The temporal distribution of chikungunya studies followed the virus during its rapid expansion since 2004 . The results revealed heterogeneity of arboviruses seroprevalence between continents and within a given country for dengue , chikungunya and Zika viruses , ranging from 0 to 100% , 76% and 73% respectively . Serological surveys provide the most direct measurement for defining the immunity landscape for infectious diseases , but the methodology remains difficult to implement . Overall , dengue , chikungunya and Zika serosurveys followed the expansion of these arboviruses , but there remain gaps in their geographic distribution . This review addresses the challenges for researchers regarding study design biases . Moreover , the development of reliable , rapid and affordable diagnosis tools represents a significant issue concerning the ability of seroprevalence surveys to differentiate infections when multiple viruses co-circulate .
Arboviral infections have become a significant public health problem with the emergence and re-emergence of arboviral diseases worldwide in recent decades . Arboviruses are considered emerging or re-emerging pathogens based on their geographic spread and increasing impact on susceptible populations . For instance , dengue virus ( DENV ) infection , once rare , is now estimated to be the most common arboviral infection globally , with transmission occurring in at least 128 countries and with nearly 4 billion people at risk [1 , 2] . Over the period 2000–2010 , an unprecedented increase in the number of cases was reported in the Americas , circulating all four serotypes ( DENV1-DENV2-DENV3-DENV4 ) and reaching the highest record of cases ever reported over a decade [3] . DENV is now hyperendemic in many parts of the tropics and subtropics . The recent emergence of chikungunya virus ( CHIKV ) in the Caribbean in 2013 and its rapid spread to 45 countries and territories in North , Central , and South America highlight its high potential for epidemics [4] . In the aftermath of this emergence , Zika virus ( ZIKV ) aroused global attention due to its rapid spread since its first detection in May 2015 in Brazil to 22 other countries and other territories in the Americas [5] . Given the increasing number of cases; geographic spread; and health , social and economic impact of arboviral outbreaks , estimating their true burden represents a crucial issue but remains a difficult task . In their acute stages , arboviral infections cause a broad spectrum of disease , ranging from asymptomatic infection to severe disease , which can lead to misclassification in case reporting , especially when several arboviruses co-circulate [6] . Furthermore , surveillance systems , which generally rely on clinicians , hospitals and laboratory reports , are appropriate for helping detect outbreaks promptly but are not designed to estimate the real disease burden and tend to underestimate the total number of cases . In fact , because of the nature of arboviral infections with 75% , between 3 and 25% and 80% of asymptomatic cases for DENV , CHIKV and ZIKV respectively [1 , 7 , 8] and because healthcare seeking can vary greatly based on access to care , surveillance data alone can be unreliable [9] . Accordingly , some studies have estimated the burden of DENV outbreaks using a range of empirical or extrapolative methods and disease-modeling approaches [1 , 10 , 11] . However , the most reliable data for empirical assessments are drawn from seroprevalence studies , which are often lacking . In fact , these seroprevalence surveys are expensive and difficult to perform; such surveys require important logistical resources , including a large workforce ( e . g . , supervisors , technicians , physicians , nurses or phlebotomists , epidemiologists , statisticians , and field investigators ) and biological support ( e . g . , sufficient freezer space for sample storage and reagents and kits for testing ) . Moreover , establishing good and reliable tests for arboviruses is an important task for public health institutions , especially when symptoms are difficult to distinguish from other common febrile illnesses and when cross-reactivity is observed [12] . The problem of cross-reactivity , as a result of the co-circulation of multiple arboviruses belonging to the same family in the same area , requires additional tests and thereby increases overall cost , time and labor [13] . However , data on arboviruses prevalence rates are essential for understanding their geographical distribution as well as their contribution to global morbidity and mortality . Such information is critical for determining the optimal allocation of the limited resources available for disease control and evaluating the impact of prevention policies and strategies such as vaccination . The rationale for conducting serological studies is straightforward; these studies provide surveillance that complements traditional symptom-based and laboratory-based surveillance . Serological studies provide an alternative approach for monitoring immunity levels in a population and do not require that people be tested during the short period when they are symptomatic [14] . In our research , seroprevalence can be defined as the frequency of individuals in a given population presenting evidence of a prior infection based on serological tests or a combination of serological and virological tests . Seroprevalence studies can be conducted using multiple designs and among various populations involving a general population or specific or relevant population subgroups . The general population concept is widely used in seroprevalence studies , but few studies provide a clear definition , and ambiguities related to the definition exist in the context of almost every country . Here , we present a definition that will be used throughout the review to classify serologic surveys according to the study population . A “general population” refers to the people ( without any ethnic , socio-economic or health status restrictions ) who inhabit a given area , usually in terms of political or geographical boundaries . The area may be quite small in size and population ( e . g . , a village of one hundred people ) or quite large ( e . g . , a nation of one million people ) . A general population survey involves the collection of data to characterize all , or nearly all , people living in the area . Because of financial and logistical constraints , the data are typically collected from a representative sample of people residing in that area through a combination of personal interviews , administered on site using a standardized questionnaire , and blood samples drawn by skilled personnel ( doctor , nurse or phlebotomist ) . Although surveys of the general population may gather data about inhabitants of all ages , lower and/or upper age limits are typically placed on eligible respondents , especially when blood samples are needed . In contrast to general population surveys , specific population surveys focus on subgroups , ( e . g . , pregnant women , school children , blood donors , and patients ) . These subgroups are defined by membership in or contact with some social institution or by the presence of exposure . Furthermore , regardless of the type of population , because a census is resource-intensive , random sampling is highly recommended as a cost-effective method for obtaining seroprevalence estimates that are representative of the target population . Convenience sampling , such as selecting administrative units or schools that are easy to sample , is expected to result in bias . The reason is that administrative units selected because of convenience may not be generalizable to the larger population [9] . Seroprevalence studies can also use different designs , including cross-sectional , prospective , and retrospective designs , and can refer to cohort or case-control studies . In the context of emerging and re-emerging arboviral diseases worldwide , we undertook a scoping review of the literature to describe and summarize the epidemiological practices , findings and insights related to seroprevalence studies reported worldwide over the recent period of 2000 to 2017 , which was marked by an unprecedented increase in the number of arboviruses cases registered across the globe .
Screening was first conducted through an online MEDLINE ( United States National Library of Medicine ) search for English- or French-language literature published between January 2000 and March 2018 . Between November 2016 and March 2018 , we searched several electronic databases with reference to the expanded Medical Subject Headings ( MeSH ) thesaurus , using the following search terms: [“arbovirus” or “arbovirus infection” or “dengue” or “chikungunya” or “zika”] AND [“seroepidemiologic studies” or “seroprevalence” or “seroepidemiology” or “serosurvey”] . The databases included the following: MEDLINE , World Health Organization Library database ( WHOLIS ) , Latin American and Caribbean Health Sciences Database ( Lilacs ) , Scientific Electronic Library Online ( SciELO ) and Scopus . A free search was also conducted through the Google search engine . Additional studies were identified through manual searches of the reference lists of identified papers . No attempt was made to identify unpublished studies . After deleting duplicates , the literature review group systematically screened the title , abstract and full text of each study for the inclusion and exclusion criteria . Articles were excluded if ( i ) the studies were published before January 1 , 2000 , or after March 15 , 2018; ( ii ) the studies were published in languages other than English or French; ( iii ) the study sample included febrile patients , hospitalized patients , suspected or confirmed cases , or HIV or malaria patients because they are likely to provide biased estimates of seroprevalence , as well as if the study sample included immigrants , military personnel , travelers , or relief workers; and ( iv ) they were prospective/retrospective cohort studies that did not provide a baseline seroprevalence , because these study designs are likely to be associated with a specific first objective that only rarely focuses on determining seroprevalence rates . We included cross-sectional and cohort studies analyzing samples from the general population , pregnant women , blood donors , age-specific subgroups , healthy volunteers and school children as possible sources of information about arboviruses seroprevalence . Data from the selected sources were collated and summarized using a table consisting of a series of Excel spreadsheets . Eligible articles were abstracted for publication metadata , settings , design , population sampling approach , sample size , laboratory assays , age categories , seroprevalence rates , ethical approval and reported biases . When a study used several separate samples ( e . g . , from different countries or different study populations or age group ) , it was separated , and each sample was considered a unique data point . Duplicate citations were removed . When articles were not available or did not provide sufficient information , we contacted the authors for additional information .
We identified 265 unique studies reporting the seroprevalence of dengue , chikungunya or Zika that were eligible for full-text review ( Fig 1 ) . Among these studies , 18% ( n = 48 ) were prospective or retrospective cohort or case-control studies , among which 16 studies provided a seroprevalence at baseline and enrolled participants according to our inclusion criteria . With respect to the study populations , 39 . 6% ( n = 105 ) of these studies targeted febrile patients , hospitalized cases , suspected or confirmed cases , malaria or HIV patients , travelers , immigrants , relief workers or military personnel . Incomplete information was available for three studies . In total , 118 studies did not fulfill the inclusion criteria . Ultimately , the review was based on 185 data points from 147 unique studies ( Fig 1 ) . A description of the included studies is available in S1 Appendix . The majority of the studies were exclusively conducted on dengue [15–112] ( n = 123 ) , with 16 . % exclusively conducted on chikungunya ( n = 29 ) [113–136] and 12% conducted on both dengue and chikungunya ( n = 23 ) [137–154]; furthermore , seven studies were conducted on Zika [8 , 155–160] , one study was conducted on both dengue and Zika [161] and two studies were conducted on both viruses [162] . Overall , as shown in the maps in Fig 3A , the studies were primarily conducted in inter-tropical areas , with some disparities within this region . We identified data from eight world regions , including eight studies from North America [15 , 22 , 55 , 60 , 60 , 71] , three from Europe [65 , 126 , 134] , 12 from Oceania [8 , 29 , 32 , 33 , 114 , 161 , 163 , 165] , 38 from Central America and the Caribbean [20 , 24 , 24 , 36–38 , 48 , 48 , 53 , 58 , 63 , 91 , 94 , 95 , 97 , 98 , 100 , 103 , 105 , 106 , 111 , 117 , 117 , 118 , 120 , 121 , 121 , 141 , 158] , 21 from Latin America [17 , 19 , 19 , 19 , 23 , 25 , 28 , 54 , 56 , 72 , 73 , 78–80 , 101 , 102 , 108 , 128 , 160 , 162] , 44 from Africa [18 , 21 , 26 , 26 , 31 , 31 , 31 , 47 , 52 , 57 , 68 , 74 , 76 , 81 , 86 , 88 , 107 , 116 , 120 , 120 , 124 , 127 , 130–132 , 132 , 133 , 137 , 140 , 143 , 144 , 144–149 , 149 , 149 , 149 , 151 , 153 , 156 , 158] , and 59 from Asia [16 , 27 , 30 , 34 , 35 , 39 , 41–45 , 45 , 46 , 49–51 , 59 , 61 , 62 , 62 , 62 , 62 , 64 , 66 , 67 , 69 , 70 , 75 , 77 , 82–85 , 87 , 89 , 89 , 90 , 92 , 93 , 96 , 99 , 104 , 109 , 110 , 112 , 113 , 115 , 117 , 123 , 125 , 129 , 135 , 136 , 138 , 139 , 141 , 150 , 152 , 152] . Dengue studies were primarily conducted in the Americas ( 39% ) and in Asia ( 33% ) ( Fig 3B ) . The countries most heavily involved in the implementation of the surveys over the past two decades were Brazil ( 12 studies ) [17 , 19 , 19 , 19 , 23 , 28 , 78–80 , 101 , 108 , 128] , Singapore ( ten studies ) [16 , 35 , 62 , 62 , 62 , 62 , 82 , 92 , 93 , 96] , Thailand ( eight studies ) [40 , 46 , 66 , 67 , 87 , 90 , 139 , 150] and India ( seven studies ) [34 , 45 , 45 , 69 , 75 , 138 , 152] . Chikungunya studies were primarily conducted in Africa ( 46% ) and Asia ( 24% ) . The most represented countries were Kenya , with six studies , and India [123 , 125 , 138 , 152] , Madagascar [128 , 128 , 128 , 128] and French Polynesia [114 , 163] , with four studies ( Fig 3C ) . Finally , Zika studies were conducted in Oceania , the Caribbean , Africa and Latin America , with three studies in French Polynesia [161 , 165 , 165] , one in Micronesia [8] , one in the French Indies ( Martinique ) [159] , one in Zambia [156] , one in Cameroon [158] , one in French Guiana [160] and one in Bolivia [162] ( Fig 3D ) . The inclusion criteria restricted the analysis to studies published between January 2000 and March 2018; however , 14 studies were conducted before 2000 ( Fig 4 ) . DENV seroprevalence studies were conducted between 1989 and 2017 . Their distribution over the last decade indicated two peaks , one in 2004 and one in 2009–2010 . The number of studies observed in 2004 might be enhanced by the re-emergence of CHIKV in Africa and the large DENV epidemic in Reunion Island . In 2010 , the first phase III clinical trial for the now available tetravalent vaccine was initiated . This event may have encouraged seroprevalence studies to provide data for future vaccine programs . There were difficulties in interpreting the distribution of DENV studies with respect to the study year and location given the expansion of the virus in the Pacific , Southeast Asia , Africa , the Americas and the Middle East before the 1990s [166] . Moreover , at the time of this study , many countries were hyper-endemic with the co-circulation of four serotypes and with repeated epidemics every three to five years . The first exclusive CHIKV seroprevalence study was conducted in 2004 in Kenya , with the re-emergence of the virus causing a large outbreak in 2004 [130] . This study was followed in 2005 by two studies , one in the Grande Comoros Island [131] , where an outbreak occurred , and in Mayotte before the 2006 epidemic [132] . In 2006 , four studies were conducted on Reunion Island and Mayotte during and after the 2006 epidemic [120 , 120 , 132 , 133] and in Benin , where no cases have been reported [116] . In 2007 , three studies were conducted: one in Malaysia [129] ( after the 2006 outbreak ) when the virus subsequently spread to Asia , one in Gabon before the 2007 outbreak [153] and one in Italy [126] when CHIKV was imported to Europe , causing an outbreak . In 2008 , two studies were conducted in India and Malaysia , where two outbreaks occurred [115 , 125] . In 2009 , two studies were conducted in India and Kenya [123 , 124] , and in 2010 , a study was conducted in Congo after a 2010 CHIKV outbreak [127] . In 2013 , CHIKV emerged in the Americas , and in 2014 and 2015 , five studies were conducted in the Caribbean [118 , 118 , 119 , 121] ( Saint-Martin , Guadeloupe , Martinique and Puerto Rico ) and Central America [122] ( Nicaragua ) during an outbreak in Saint Martin and post-outbreak in the other locations . One study was conducted in Vietnam in 2015 , where little was known about CHIKV transmission and where dengue is endemic [136] . The last study was conducted in 2016 in Brazil in a post-outbreak context [128] . There were seven ZIKV studies . The first study was conducted in Yap Island during the 2007 outbreak [8] , and the second study was conducted in Zambia in 2013 [156] , where no information on ZIKV was available . Two studies were conducted in French Polynesia in two distinct populations during and after the 2014 outbreak and one in 2015 [155] . Another study was conducted in Cameroon in 2015 [158] , and one study was conducted in 2016 in Martinique ( West Indies ) [159] . Finally , the last study was conducted in French Guiana during the ZIKV outbreak in 2016 [160] . All seroprevalence data are presented in S1 Appendix .
Arboviral infections are common causes of disabling fever syndromes worldwide . In many countries , the concomitant co-circulation of dengue , chikungunya and Zika viruses represents a major recent public health and biomedical challenge . Prior to the introduction and subsequent spread of CHIKV and ZIKV in the Americas , dengue was the predominant arboviral infection worldwide . In this context of emerging and re-emerging arboviral diseases worldwide , estimating the burden of these diseases represents a major challenge to more efficient planning for disease control and reducing the risk of future re-emergence of arboviruses . Several affected countries face challenges in estimating the burden of arboviruses . Nonetheless , estimating the true burden of arboviral infections remains a difficult task given the large number of unapparent infections , especially those of dengue fever [1] . Serological surveys are thus required to identify the distribution of these diseases and measure their epidemic impact . A recent estimate indicated that the number of cases affected by any of these three arboviruses dramatically increased after 2013 , reaching over 3 . 5 million by the end of 2015 in the Americas [169] . This review emphasizes several aspects of arboviruses epidemiology and describes current challenges and implications for dengue , chikungunya and Zika seroprevalence studies worldwide . Overall , our results highlight the highly heterogeneous nature of study designs and serological tests used in arboviral seroprevalence studies . Seroprevalence surveys have the benefit of not being affected by surveillance system sensitivity or symptomatic case reporting rates but still have several limitations inherent to the adopted methodology . Selection biases , defined in our review as a distortion in the seroprevalence rate , may occur due to the use of a non-probabilistic sampling frame or poor field worker practices , such as replacing a selected household with one that is easier to reach [170] . Furthermore , the use of serum samples collected for various purposes frequently hinders the representativeness of the population sample and , consequently , that of the provided estimations . Even if the use of convenience samples is a good strategy for increasing the volume of serological data produced , the potential biases such sampling introduces must be considering during the analysis process to produce valid results . These limitations in the literature underscore the challenge of estimating global prevalence in the absence of nationally representative age-specific databases . Whenever surveys are conducted , all efforts to ensure high-quality collected data should be made . In particular , probabilistic samples should be used , and the sample size and number of clusters should be selected appropriately to rigorously estimate population seroprevalence rates . The review also highlights the variety of serological tests used to measure antibodies activities . ELISA tests are the most common diagnostic method ( used in more than half of the studies included in this review ) . Moreover , we noted that more than half of the studies that performed IgG ELISA tests used the indirect method , which is recommended , as it allows for the detecting of lower levels of antibodies than the direct method does and is thus more sensitive [171] . Most commercially available diagnostic IgG ELISAs that are adjusted to measure past arboviral exposure tend to have high sensitivity but suffer from low specificity due to high cross-reactivity with other arboviruses ( flaviviruses or alphaviruses ) circulating in a given geographical area or with Japanese encephalitis ( JE ) and YFV recommended immunization [172 , 173] . The resulting false positives could lead to information bias that can be overcome through control with neutralization tests . Both tests provide complementary results because one test is a biochemical assay ( ELISA ) measuring antibodies binding to the antigen and the other is a biological assay measuring antibodies’ capacity to neutralize an infecting virus . Only neutralization tests measure the biological parameters of in vitro virus neutralization and are the most virus-specific serological tests [174] . Indeed , for seroepidemiological studies , neutralization tests remain the “gold standard” for confirming and serotyping DENV infections in regions where two or more flaviviruses are co-circulating . These tests , however , are time-consuming , labor-intensive and expensive and are not as amenable to testing large numbers of sera as the ELISA is . When neutralization tests are not performed to complete results from IgG ELISA , country-specific contexts , including the presence of other circulating flaviviruses or alphaviruses and immunization programs for JE and YF , must be considered when interpreting seroprevalence results . Although we stratified seroprevalence data by assay to allow for comparisons , more than half of the ELISA tests were performed using different commercial kits and in-house assays with variable sensitivities and specificities . Moreover , differences in assay formats , usage of antigen , and detection systems make it difficult to estimate the performances value of each individual assay by proper comparison [175] . In a multicenter evaluation using a commercial assay , it was shown that the sensitivities and the specificities varied between studies depending on the serum samples of the respective collaborating centers used for the performance evaluation [176] . These variations reported from several studies [177] indicated the need to develop the most sensitive and specific diagnosis tool to provide recommendations for future serological studies . Although the diversity of study designs and serological tests used in the selected seroprevalence studies represents a major limitation for the comparison of seroprevalence rates by geographical region , our literature search highlights the highly heterogeneous seroprevalence of DENV and CHIKV worldwide as well as the significant variability among regions in the same country . The review also clearly shows that seroprevalence was the highest in island environments for both arboviruses . Some of these variations may stem from methodological differences , as well as the choice of study population , sample size and diagnostic test . This heterogeneity may also reflect differential exposure to mosquitoes . Indeed , disease transmission can substantially differ between regions characterized by different environmental and climatic determinants of vector density . For instance , in Kenya , alphavirus antibodies , specifically those against CHIKV , were detected only in children from the Kisumu District ( lowlands ) and not in children from the Nandi District ( highlands ) [144] . Geographic and climatic differences between these two regions could provide evidence for varying environmental factors related to arboviruses transmission risk . For instance , mosquito vectors are not as prolific in the colder climate of the highlands , whereas the lowlands offer warmer and wetter areas for mosquito development and could provide an appropriate environment for mosquito vectors and subsequent arboviral transmission . Moreover , the seroprevalence heterogeneity is related to the different transmission dynamics of these arboviruses , including force of infection , reproductive number and others factors such as strain [178] . The review highlights some factors associated with arboviral seroprevalence . We noted an increase in seroprevalence among older people in 44% of DENV studies and 18 . 5% of CHIKV studies mainly conducted in endemic areas , whereas only one ZIKV study obtained this result . The epidemiological context of affected countries appears to be associated with the relationship between age and seroprevalence . For instance , in an emergence context , few studies reported this association , as the target population is naïve . However , some CHIKV studies not conducted in an endemic area reported this association , which may suggest that age is associated with level of exposure . We noted that 14 studies found an association between gender and seroprevalence . These findings suggest that there may be gender-related differences; however , these discrepancies require further exploration , as the health gender gap may stem from various patterns that affect exposure to mosquitoes [179] . We also observed that , regarding DENV and CHIKV , this association varied between countries , and no study reported an association between gender and ZIKV seroprevalence . However , given the transmission issues associated with ZIKV , we can expect that in future studies , sexual transmission will correlate with higher seroprevalence among women . Although few studies revealed an association between ethnicity and seroprevalence , this finding can be related to background prevalence in the country of origin , in combination with increased early life exposure before migration or exposure during travel to their region of origin post migration . Moreover , these associations might also be partly explained by increased susceptibility related to the lower socioeconomic position of certain ethnic groups . The relationship between health and socio-economic status is well documented , and research has revealed a graded association in which people of lower socio-economic status have much worse health outcomes than those of higher socio-economic status [180 , 181] . Finally , 20 studies also reported an association between environmental factors and seroprevalence , as certain environmental conditions , such as house structure or objects collected in the yard , are more hospitable to A . aegypti [182] . Although the proportion of seropositivity depends on the diagnostic method used , it also relies on study planning; if a serosurvey is conducted long after the end of an outbreak , the signal for the antibodies may be lower than in a study conducted close to the end of an outbreak . Our results indicate that DENV seroprevalence in the Americas was higher than that in Asia , which is surprising because dengue has been endemic in Southeast Asia for decades [183] , and the Asian burden , including the Western Pacific , accounted for nearly 75% of the global burden worldwide [184] . Analysis revealed that studies conducted in the Americas were performed significantly more frequently in an outbreak or post-outbreak context ( p<0 . 01 ) via IgM ELISA , which could explain the discrepancy between America and Asia . Our review included 185 studies worldwide according to well-defined inclusion criteria . The findings indicate that the distribution of our studies follows the same pattern observed for the expansion of their vectors [185] . Moreover , all of the studies reflect areas of arboviral circulation in an epidemic pattern . However , the maps clearly demonstrated where seroprevalence survey data are lacking and identified potential places for implementing future seroprevalence studies; the maps also highlighted places in tropical regions where no data are available , especially for CHIKV and ZIKV , which are considered emergent or re-emergent viruses . Some countries located in the tropics and subtropics were not represented , although they are considered at risk of transmission by the WHO; this is especially true for Africa , where few studies were conducted and where the epidemiology of these arboviruses is under-exploited . In addition , a recent review suggested that dengue transmission is endemic to 34 countries in all regions of Africa [186] . The temporal distribution of Chikungunya studies followed the timeline of CHIKV outbreaks during its rapid expansion since 2004 , suggesting that ZIKV surveys , in the context of its recent emergence in the Americas , may be currently in process and , if not , that such surveys should be addressed rapidly . Serological surveys provide the most direct measurement for defining the immunity landscape for infectious diseases , but they remain difficult to implement . Overall , dengue , chikungunya and Zika serosurveys have followed the expansion of these arboviruses , but there remain gaps in their distribution . Serological studies can address future challenges in identifying trends in arboviruses transmission over time , and age-specific antibody prevalence rates can be used to estimate when major changes in transmission occurred . | Arthropod-borne viruses ( arboviruses ) are among the most important of the emerging infectious disease public health problems facing the world . The actual impact of arboviruses worldwide remains unknown , and estimating the true burden of these diseases represents a current challenge . Serological surveys are the most reliable tool for estimating the impact of arboviruses outbreaks in a given territory , and the results of such surveys have implications for potential mitigation measures such as vaccination . We undertook a thorough review of the literature produced from 2000 to March 15 , 2018 , addressing the seroprevalence of dengue , chikungunya and/or Zika to describe and summarize methodological approaches and map the geographical distribution of seroprevalence studies for these three viruses worldwide . A total of 185 studies addressing the seroprevalence of dengue , chikungunya and/or Zika were included in the review . Most of the studies were exclusively conducted on dengue ( 66 . 5% ) , but 16% were exclusively conducted on chikungunya , and 7 studies were exclusively conducted on Zika; the remainder were conducted on multiple arboviruses . Our study reveals that a wide range of methodological designs were applied regarding population , recruitment and/or laboratory testing . This study also highlights the high seroprevalence heterogeneity between continents and within a given country for dengue , chikungunya and Zika viruses . The results underscore existing gaps in seroprevalence studies distribution worldwide and the need to develop the most sensitive and specific diagnosis tool to provide recommendations for future serological studies . | [
"Abstract",
"Introduction",
"Materials",
"and",
"method",
"Results",
"Discussion"
] | [
"dengue",
"virus",
"medicine",
"and",
"health",
"sciences",
"enzyme-linked",
"immunoassays",
"pathology",
"and",
"laboratory",
"medicine",
"togaviruses",
"chikungunya",
"infection",
"pathogens",
"tropical",
"diseases",
"microbiology",
"alphaviruses",
"viruses",
"chikungunya... | 2018 | Current challenges and implications for dengue, chikungunya and Zika seroprevalence studies worldwide: A scoping review |
Failure of the human heart to maintain sufficient output of blood for the demands of the body , heart failure , is a common condition with high mortality even with modern therapeutic alternatives . To identify molecular determinants of mortality in patients with new-onset heart failure , we performed a meta-analysis of genome-wide association studies and follow-up genotyping in independent populations . We identified and replicated an association for a genetic variant on chromosome 5q22 with 36% increased risk of death in subjects with heart failure ( rs9885413 , P = 2 . 7x10-9 ) . We provide evidence from reporter gene assays , computational predictions and epigenomic marks that this polymorphism increases activity of an enhancer region active in multiple human tissues . The polymorphism was further reproducibly associated with a DNA methylation signature in whole blood ( P = 4 . 5x10-40 ) that also associated with allergic sensitization and expression in blood of the cytokine TSLP ( P = 1 . 1x10-4 ) . Knockdown of the transcription factor predicted to bind the enhancer region ( NHLH1 ) in a human cell line ( HEK293 ) expressing NHLH1 resulted in lower TSLP expression . In addition , we observed evidence of recent positive selection acting on the risk allele in populations of African descent . Our findings provide novel genetic leads to factors that influence mortality in patients with heart failure .
Heart failure ( HF ) is a common clinical condition in which the heart fails to maintain blood circulation adequate to meet the metabolic demands of the body without increased cardiac filling pressures . HF is the result of chronic ventricular remodelling initiated by myocardial injury , volume/pressure overload , or intrinsic cardiomyopathic processes . Progression of HF is a complex process involving many tissues , driven by activation of neurohormonal pathways , which induce gradual myocardial hypertrophy , ventricular dilation , and deterioration of cardiac function , often resulting in death from low cardiac output , arrhythmia , or thromboembolic complications [1] . Activation of such neurohormonal pathways in the short term increases cardiac output when necessary . However , long-term activation results in accelerated ventricular remodelling and myocyte death . Inhibitors of deleterious neurohormonal pathways , including adrenergic [2–4] and renin-angiotensin-aldosterone ( RAAS ) [5–8] pathways have been shown to improve ventricular function and survival in patients with HF and are the mainstay of current pharmacological treatment of HF [9–10] . Despite advances in therapy with neurohormonal antagonists , mortality after onset of HF remains high [9–13] and continued progress to identify additional therapeutic targets is needed . Genome-wide association ( GWA ) studies have the potential to identify in an agnostic manner genetic variants related to clinical outcomes in humans and has led to the identification of novel pathways [14] and potential treatments [15] for cardiovascular traits . Heritable factors have been shown to be predictive of mortality in certain heart failure patients [16] . We therefore implemented a genome-wide association approach to identify novel molecular determinants of mortality in patients with new-onset HF .
We expanded our previously published GWA study [17] of HF mortality with additional samples and extended follow-up in Stage 1 . Stage 1 included 2 , 828 new-onset HF patients from five community-based cohorts , thus representative of the general population of HF patients , as part of the Cohorts for Heart and Aging Research in Genomic Epidemiology ( CHARGE ) consortium [18]: the Atherosclerosis Risk in Communities ( ARIC ) Study , the Cardiovascular Health Study ( CHS ) , the Framingham Heart Study ( FHS ) , the Health , Aging and Body Composition ( Health ABC ) Study , and the Rotterdam Study ( RS ) . Cohorts are described in detail in S1 Text . HF was defined using international published criteria as outlined in S1 Table . Subjects in Stage 1 cohorts were of European ancestry , predominantly male , and approximately 20–30% had a history of myocardial infarction at the time of HF diagnosis . Additional characteristics are shown in Table 1 . During an average follow-up time of 3 . 5 years , 1 , 798 deaths occurred . The sample-size weighted average 1-year mortality rate was 28% . Among deaths , 51% were classified as cardiovascular , 19% were due to neoplasms , 10% were respiratory deaths , and the remaining were due to other miscellaneous causes . Genotyping using high-density Illumina or Affymetrix single nucleotide polymorphism ( SNP ) arrays , followed by imputation to the HapMap CEU release 22 imputation panel was performed in each cohort . Population stratification was assessed and corrected in each cohort as described in S1 Text . Association with time to death following HF diagnosis was examined in each cohort using Cox proportional hazards models with censoring at loss to follow-up . Mild inflation of test statistics was observed only in the Framingham Heart Study ( FHS ) as shown in S1 Fig ( λGC = 1 . 07 , other cohorts ≤ 1 . 03 ) , and genomic control was applied in each individual study . In the meta-analysis of all cohorts , there was no evidence of inflated test statistics overall ( λGC = 1 . 00 ) as shown in S2 Fig , so no further genomic control was needed . Results for all SNPs across the genome are plotted in S3 Fig . Single nucleotide polymorphisms ( SNPs ) passing a significance threshold specified a priori as P < 5 . 0x10-7 , as used in our previous article [17] , were carried forward to a second stage of genotyping in independent cohorts . Five SNPs on chromosome 5q22 and one SNP on chromosome 3p22 passed the pre-specified P-value threshold . Results for all six SNPs are shown in Table 2 and S3 Table . The five SNPs on chromosome 5q22 were highly correlated ( pairwise r2 > 0 . 9 ) . Two sentinel SNPs , rs9885413 and rs12638540 , on chromosomes 5q22 and 3p22 , respectively , were next genotyped in 1 , 870 European-ancestry subjects with new-onset HF from four independent cohorts in Stage 2: Malmö Diet and Cancer , Malmö Preventive Project , Physicians’ Health Study , and the PROSPER trial . Characteristics of populations in Stage 2 are shown in S2 Table . During an average sample-size weighted follow-up of 4 . 3 years in Stage 2 samples , 889 patients died . We observed evidence of association with mortality for rs9885413 on chromosome 5q22 ( P = 0 . 006 ) but not for the SNP rs12638540 ( P = 0 . 18 ) which reached nominal significance in our previous analysis [17] . Results for both SNPs are shown in Table 2 . In the combined results from Stages 1 and 2 , rs9885413 was associated with a 36% relative increase in mortality per minor allele ( P = 2 . 7x10-9 ) . There was no evidence for effect heterogeneity across cohorts in the two stages ( P for heterogeneity = 0 . 39 ) as shown in S4 Table . The SNP had a similar minor allele frequency ( MAF = 0 . 07 ) across cohorts . Information on cause-specific mortality was available from death certificates in a subset of cohorts ( S5 Table ) and was explored descriptively due well-known problems with substantial misclassification in death certificate data and low power for agnostic GWAS of individual causes . The minor allele frequency was slightly higher for several causes of death associated with heart failure , including renal , pulmonary and endocrine mortality and death from ischemic heart disease . We next examined whether rs9885413 on chromosome 5q22 that was associated with HF mortality was also associated with differences in myocardial structure and function , which could potentially mediate the association ( S6 Table ) . In 12 , 612 individuals from the EchoGen Consortium [19] , the SNP was not associated with major echocardiographic characteristics . The SNP rs9885413 was not associated with incident HF in 20 , 926 individuals from the general population in the CHARGE-HF study [20] , or with cardiac endocrine function , as determined by plasma levels of atrial and B-type natriuretic peptides ( all P > 0 . 05 ) , in a GWA study of 5 , 453 individuals from the population-based Malmö Diet and Cancer study [21] . No association was observed with electrocardiographic measures of cardiac conduction ( n = 39 , 222 ) [22] or repolarization ( n = 74 , 149 ) [23] , which confer risk of ventricular arrhythmia , or with sudden cardiac death in 4 , 496 sudden death cases and over 25 , 000 controls from the general population ( described in S1 Text ) . The lead SNP rs9885413 on chromosome 5q22 that was associated with mortality is located in an intergenic region , 100 kb downstream of the gene SLC25A46 , 114 kb upstream of TMEM232 , and 230 kb upstream of TSLP as shown in Fig 1 . The SNP is not in linkage disequilibrium with any known coding SNP in the 1000 Genomes Project database ( no coding SNP with r2 > 0 . 01 to the sentinel SNP ) . We therefore sought to evaluate gene regulatory functions of this SNP . In 129 human tissues from the ROADMAP Epigenomics project [24] , we studied whether rs9885413 or strongly correlated SNPs ( a total of 9 at r2 > 0 . 8 ) are located in regulatory regions , as determined by histone modification patterns . None of the 9 SNPs was located in an active regulatory region in cardiac tissues ( S7 Table ) , but rs9885413 was located in a predicted enhancer in several epithelial or mesenchymal tissues , including keratinocytes , gastrointestinal cell types and adipose cells ( Fig 2 and S7 Table ) . Regulatory motif annotations in HaploReg indicate that the SNP causes a change in a regulatory motif predicted to bind the transcription factor NHLH1 as shown in S8 Table . Interestingly , NHLH1-null mice have been shown to be predisposed to premature , adult-onset unexpected death in the absence of signs of cardiac structural or conduction abnormalities , in particular when mice were exposed to stress [25] . Little is known about the function of NHLH1 , but it is widely expressed in human tissues and has been shown to regulate expression of key inflammatory genes [26] . To experimentally test the effect of rs9885413 on enhancer activity , the 100 bp region flanking the SNP ( 50 bp on either side ) was cloned into a reporter vector and transfected into HEK293 cells expressing NHLH1 ( S1 Text ) . Luciferase activity measured after 24 hours was 4-fold higher with a construct corresponding to the risk allele as compared to the wild-type allele ( S4 Fig , P < 0 . 001 ) , indicating that the risk allele of rs9885413 substantially increases enhancer activity . We next explored the association of rs9885413 with DNA methylation at the locus , providing functional evidence of epigenetic association and regulation of gene expression . DNA methylation was determined by a microarray assaying in total over 480 000 CpG methylation sites in whole blood samples from 2408 participants of the FHS . Of the 84 CpG methylation sites on the microarray within +/- 500 kb of the SNP , two were significantly associated with rs9885413: cg21070081 ( beta 0 . 017 per T allele , P = 9 . 0x10-69 ) and cg02061660 ( beta -0 . 015 per T allele , P = 4 . 5x10-40 ) , thus constituting strong methylation quantitative trait loci ( mQTLs ) at the locus . Other , correlated SNPs at the locus were more strongly associated with each of these mQTLs as shown in S5 Fig: rs244431 for cg21070081 ( P = 6 . 7x10-369 ) and rs72774805 for cg02061660 ( P = 7 . 0x10-85 ) . The SNP rs72774805 ( perfect proxy SNP rs3844597 used ) but not rs244431 was associated with heart failure mortality ( P = 3 . 3x10-3 and 0 . 08 , respectively ) , indicating that the methylation site cg02061660 is more strongly related to the underlying signal for heart failure mortality . The association of rs9885413 with lower probability of methylation at cg02061660 was replicated in 731 participants from the Rotterdam study ( beta -0 . 029 per T allele , P = 1 . 7x10-11 ) . Adjustment for cell types from direct measurement instead of estimates from methylation patterns did not abolish the association ( beta -0 . 029 per T , P = 1 . 2x10-6 ) . Interestingly , differential methylation at this CpG site was also correlated with a SNP at the locus previously associated with allergic sensitization [27] ( rs10056340 , P = 4 . 7x10-29 for mQTL ) , suggesting a link to inflammatory disease . This SNP was also modestly correlated with rs9885413 ( r2 = 0 . 28 ) and associated with heart failure mortality ( P = 0 . 01 ) . The association of cg02061660 with rs9885413 ( P = 0 . 52 ) and rs10056430 ( P = 0 . 87 ) was abolished in analyses conditioning for rs72774805 , for which the association was also markedly attenuated ( P = 7 . 0x10-33 and 2 . 1x10-46 , respectively ) indicating that these correlated SNPs may reflect the same underlying signal . We further assessed the association of rs9885413 with gene expression . No gene was significantly associated with rs9885413 in the diverse tissues from the Gene-Tissue Expression ( GTEx ) project [28] after correction for multiple tests ( S1 Text , S9 Table ) , although conclusions were limited by a small sample size . We next assessed association of the SNP with gene expression in two large datasets with each of the tissues most relevant for the phenotype under study: heart tissue and whole blood . We observed no convincing evidence of association ( S1 Text , S10 Table ) with gene expression in 247 left ventricular samples from patients with advanced heart failure ( n = 116 ) undergoing transplantation and from unused donors ( n = 131 ) . Finally , we tested the association of rs9885413 with the expression of genes at the locus in whole blood from 5257 FHS participants [29] , and with DNA methylation at cg02061660 among 2262 FHS participants . All five genes at the locus ( Fig 1 ) except TMEM232 were expressed in blood . We did not observe association of the SNP rs9885413 with any transcript , but expression of one gene ( TSLP ) was significantly associated with the methylation status of cg02061660 ( P = 1 . 1x10-4 ) . The TSLP gene encodes a cytokine released from epithelial cells that induces release of T cell-attracting chemokines from monocytes , promotes T helper type 2 cell responses , enhances maturation of dendritic cells and activates mast cells . It has also been linked to angiogenesis and fibrosis . A monoclonal antibody targeting and inhibiting TSLP is currently in clinical phase III trials for asthma and allergic inflammation after a promising phase II trial [30–32] . In the myocardium , the TSLP gene has very low expression ( S10 Table ) but expression has been described in mature myocardial fibroblasts , which are abundant in the myocardium but of substantially smaller volume than cardiomyocytes and likely contribute little to the overall myocardial RNA pool [31 , 32] . To examine whether the transcription factor NHLH1 affects the expression of any of the five genes in the locus ( Fig 1 ) , we knocked down NHLH1 in HEK293 cells using siRNAs . A 50% decrease in NHLH1 mRNA levels was seen 48 hours after transfection , confirming efficient knock down ( p<0 . 05 , S6A Fig ) . TSLP was the only gene at the locus affected by NHLH1 knock down , showing a 30% decrease compared to cells transfected with negative control siRNA ( p<0 . 05 , S6A Fig ) . Moreover , we observed a dose-response relation between level of NHLH1 knockdown and expression of TSLP in HEK293 cells ( r2 = 0 . 74 , p<0 . 0001 , S6B Fig ) . Finally , distribution of the risk allele of rs9885413 in human populations was assessed using data from HapMap phase II . The derived ( non-ancestral ) T allele ( risk allele for mortality ) was highly differentiated among human populations ( S7 Fig ) having risen to an allele frequency of 0 . 59 in a Nigerian population ( HapMap YRI sample ) but only 0 . 06 in a European population ( CEU sample ) . The fixation index ( Fst ) , a measure of population differentiation in allele frequencies , for comparison of YRI and CEU was 0 . 48 and more extreme Fst was observed in only 2 . 4% of SNPs in the HapMap phase 2 dataset . Consistent results were observed for another signature of recent positive selection , based on longer runs of haplotype homozygosity in carriers of the derived allele ( standardized integrated haplotype score -0 . 766 in YRI , where negative score values indicate longer haplotypes on the background of the derived allele ) [33] . These observations are consistent with positive selection in recent human history , with a selective sweep resulting in high frequency of the derived allele in western African populations . These findings are of particular interest as HF mortality is well known to be higher in populations of African ancestry , although the current study has not tested for the association with HF mortality in such populations [34] .
We identified a SNP on chromosome 5q22 associated with increased mortality in subjects with HF . Although previous genome-wide association studies have described hundreds of loci associated with risk of disease onset , few have examined prognosis in subjects with manifest disease . This approach has the potential to generate targets for novel disease-modifying medications . Through a series of analyses in silico and in vitro we show that the SNP is located in an enhancer region , and confers increased activity of this enhancer . Interestingly , mice deficient in the transcription factor NHLH1 predicted to bind a motif in this enhancer region have been reported to be predisposed to premature , adult-onset unexpected death in the absence of signs of cardiac structural or conduction abnormalities . NHLH1 has also been shown to regulate expression of key inflammatory cytokines such as interleukin-6 and tumor necrosis factor α . The SNP was not associated with any electrocardiographic , endocrine , or echocardiographic marker of increased risk in the general population , suggesting a mechanism specific to heart failure , an extracardiac pathway of importance in cardiac pathophysiology , or interaction with therapy for heart failure which we were unable to further test given the inception cohort design of this study . We also did not observe any robust eQTL associations for the SNP in heart . The SNP was however associated with a DNA methylation signature in whole blood that was also associated with a SNP previously associated with allergy , and with expression of the cytokine TSLP in blood . Knockdown of NHLH1 also resulted in lower expression of TSLP in HEK293 cells . This non-coding SNP may thus exert an influence on TSLP expression via altered NHLH1 enhancer function and DNA methylation at the methylation site cg02061660 . Detailed characterization of causal variants and different association signals at the locus would however require finemapping and sequence data . The TSLP cytokine is released from epithelial cells and fibroblasts and is considered important in initiation of inflammatory responses to tissue damage , particularly in the type 2 T-helper ( Th2 ) pathways . Th2 pathways are central in the response to extracellular parasites but also play a key role in the pathophysiology of allergies and hypersensitivity reactions . A small subset of HF is known to be caused by Th2-mediated inflammation ( eosinophilic cardiomyopathy ) , yet Th2 cells have received limited attention in HF pathophysiology . Recent experimental work implicates an important role of T-helper cells in HF progression for both systolic and diastolic heart failure , but has mainly focused on type 1 T-helper pathways [35 , 36] . It remains unclear if the mechanism for rs9885413 is through a specific etiology characterized by high mortality such as eosinophilic cardiomyopathy or a pathway involved in outcomes with manifest disease . The lack of association with HF incidence suggests that it may not act through incidence of a specific etiology , although firm conclusions are limited by sample size . We did not observe significant associations of the SNP with gene expression in any tissue . It is possible that adequately powered samples with a specific cell subtype in a specific context is needed to detect such associations , as illustrated by a recent study which only observed certain eQTLs with single-cell but not across averaged cells [37] . Indeed , baseline expression of TSLP was low in our samples , and is induced by tissue injury , microbes , viruses and proinflammatory cytokines [38] . Evidence of recent positive selection in individuals of African descent suggests that the HF risk allele may have been beneficial in some environments in recent human history . Inflammatory pathways are enriched for signals of recent positive selection , reflecting that infectious disease has been an important cause of mortality throughout recent evolution . Genes such as HBB and APOL1 have also been reported to have been subject to recent positive selection in Africa by conferring protection against infectious diseases such as Malaria and Trypanosomiasis ( sleeping sickness ) [39] , and APOL1 alleles have also been linked to cardiovascular disease [40] . As cardiovascular disease and heart failure often presents after reproductive age , increased mortality in such patients would not be expected to exert purifying ( negative ) selective pressure . Whether SNPs at 5q22 contribute to higher mortality in subjects of African ancestry remains to be shown . Thus , although additional work is needed to further clarify the tissues and pathways perturbed by this genetic variant and the mechanisms linking it to mortality in HF patients , the current findings implicate rs9885413 as a novel marker of increased risk among patients with HF . Complementary epigenomic evidence demonstrated candidate regions and genes , which may be mediators in cardiac pathophysiology and potential therapeutic targets to improve prognosis in patients with HF .
A genome-wide association ( GWA ) study was performed in a total of 2 , 828 subjects of European ancestry with HF from seven samples collected within five large community-based prospective cohort studies including the Atherosclerosis Risk in Communities ( ARIC and ARIC2 ) Study , the Cardiovascular Health Study ( CHS ) , the Framingham Study ( FHS ) , the Health ABC ( Health ABC ) study and the Rotterdam Study ( RS and RS2 ) . Sample characteristics , data collection and clinical definitions have been described previously and are summarized in S1 Text . [41–46] First diagnosis of heart failure ( new-onset ) was ascertained using a variety of methods based on international published criteria , as detailed in S1 Table . Mortality was ascertained from telephone contacts with relatives and from medical records , death certificates and/or municipal records ( S1 Text ) . Genotyping was performed using commercially available assays for genome-wide SNP detection . Imputation of non-genotyped SNPs was performed using CEU reference panels of SNP correlations from the HapMap project phase II ( S1 Text ) , to characterize a total of 2 . 5 million SNPs . Imputation quality was assessed for each SNP from the ratio of observed over expected variance of allele dosage . All-cause mortality following initial HF diagnosis was examined for association with additive allele dosage of each genotyped or imputed SNP using Cox proportional hazards models , with censoring at the end of or loss to follow-up . Models were adjusted for age at diagnosis , sex , and recruitment site in multicenter cohorts . In the family-based FHS , Cox models were implemented with clustering on pedigree to account for relatedness . Genomic control was applied to results from each cohort . Cohort-specific GWA results were combined using fixed effects meta-analysis with inverse variance weights . SNPs were excluded from cohort-level analyses if exhibiting implausible beta coefficients ( < -5 or > 5 ) and from the meta-analysis for low heterozygosity ( sample size-weighted minor allele frequency ≤ 0 . 03 , corresponding to < 100 minor allele carriers with an endpoint ) . SNPs passing a P-value threshold defined a priori as P < 5 . 0x10-7 in the genome-wide stage 1 were carried forward to the second stage with targeted genotyping in 1 , 870 HF patients from four independent cohorts . For 2 . 5 million tests , this threshold limits the expected number of genome-wide false positives to approximately 1 , assuming statistical independence of tests . The second stage included four independent cohorts; the Malmö Diet and Cancer Study ( MDCS ) , the Malmö Preventive Project ( MPP ) , the Physicians’ Health Study ( PHS ) and the Prospective Study of Pravastatin in the Elderly at Risk ( PROSPER ) [47–50] . Heart failure ascertainment and time of death in these cohorts was similar to in stage 1 cohorts , as shown in S1 Table and S1 Text . Genotyping was performed as outlined in S1 Text . Association analyses and meta-analysis of results were performed as in the first stage . Meta-analysis of stage 1 and 2 was performed , and a combined P-value < 5 . 0x10-8 was considered significant . Heterogeneity was assessed across the combined stage 1 and 2 cohorts using Cochran’s Q test for heterogeneity , computed as the sum of the squared deviations of each study’s effect from the weighted mean over the study variance , and the I2 test , the percentage of total variation across studies that is due to heterogeneity rather than chance ( I2 = [Q—df] / Q ) [51 , 52] . The association of replicated SNPs with measures of cardiac structure and function was evaluated from summary results of the following GWA consortia: EchoGen [19] , CHARGE-HF [20] , CHARGE-QRS [22] , natriuretic peptides in 5453 subjects from the Malmö Diet and Cancer study [21] , QT-IGC [23] , and the CHARGE Sudden Cardiac Death consortium ( manuscript in preparation ) . Each of these consortia is described in S1 Text . The correlation of replicated SNPs with known coding SNPs was examined in the databases for the 1000 Genomes Project and phase III of the HapMap project , using SNAP [53] . The location of SNPs in relation to regulatory motifs was explored using histone methylation patterns generated as part of the ROADMAP Epigenomics project [24] . Enhancers were identified in each of the 129 ROADMAP tissues using the ChromHMM algorithm [54] from patterns of monomethylation ( H3K4Me1 ) of the fourth residue ( lysine ) and acetylation of the 27th residue ( H3K27Ac ) of histone H3 . The location of SNPs in relation to transcription factor binding sites was assessed in silico using HaploReg version 4 . 1 ( http://www . broadinstitute . org/mammals/haploreg/haploreg . php ) [55] and the UCSC Genome Browser ( http://genome . ucsc . edu ) . In HaploReg , position weight matrices ( PWMs; probabilistic representations of DNA sequence ) were computed with p-values based on literature sources and ENCODE ChIP-Seq experiments as previously described [55] , and only instances where a motif in the sequence passed a threshold of P < 4−7 were considered . The NHLH1-binding motif was retrieved into HaploReg from the manually curated TRANSFAC database . Complementary DNA oligonucleotides corresponding to the 100 bp genomic region flanking rs9885413 ( 50 bp on either side of the SNP ) were cloned into the luciferase reporter vector pGL3-Promoter ( Promega , Madison , WI ) using the MluI and BglII sites . Two different sets of oligos were cloned , one corresponding to the major allele of rs9885413 ( pGL3P-G ) and one to the minor allele ( pGL3P-T ) . Oligonucleotide sequences were as following: major allele sense: CGCGTCCTGCCTCACATAATCTTTTTGTTTGTCCCCCTGAAATGGATTCTCAGCTGTTGCCCAAACATTTCATCTTAGCGTTCCAGGTTTGAACTCGCCCTCACGA , minor allele sense: CGCGTCCTGCCTCACATAATCTTTTTGTTTGTCCCCCTGAAATGTATTC TCAGCTGTTGCCCAAACATTTCATCTTAGCGTTCCAGGTTTGAACTCGCCCTCACGA , and the corresponding antisense sequences . The reporter vectors were co-transfected with the pRL-null vector at a ratio of 10:1 into HEK293 cells using Lipofectamine LTX ( Life Technologies ) according to the manufacturer’s instructions . 24 hours post-transfection , luciferase activity was assayed using the Dual-Luciferase Reporter Assay System ( Promega ) and Glomax 20/20 Luminometer ( Promega ) . The signal from the reporter vector was normalized to the signal from the pRL-null vector . Samples of left ventricular cardiac tissue from patients undergoing cardiac surgery were genotyped for the SNP rs9885413 and for all five transcripts within +/- 500 kb of the SNP . Samples of cardiac tissue were acquired from patients from the MAGNet consortium ( http://www . med . upenn . edu/magnet/ ) . Gene expression levels were determined using the Affymetrix ST1 . 1 gene expression array ( Affymetrix , Santa Clara , CA , USA ) in a cohort including 247 heart samples . Genotyping was performed using the Illumina OmniExpress array . Left ventricular free-wall tissue was harvested at time of cardiac surgery from subjects with heart failure undergoing transplantation or from unused transplant donors . In all cases , the heart was perfused with cold cardioplegia prior to cardiectomy to arrest contraction and prevent ischemic damage . Tissue specimens were then obtained and frozen in liquid nitrogen . Genomic DNA from left ventricle was extracted using the Gentra Puregene Tissue Kit ( Qiagen ) according to manufacturer’s instruction . Total RNA was extracted from left ventricle using the miRNeasy Kit ( Qiagen ) including DNAse treatment on column . RNA concentration and quality was determined using the NanoVue Plus spectrophotometer ( GE Healthcare ) and the Agilent 2100 RNA Nano Chip ( Agilent ) . For all samples , genome-wide SNP genotypes were generated using the Illumina OmniExpress Array . Caucasian Ancestry was verified using multi-dimensional scaling of genotypes . For Gene expression array experiments , the Affymetrix ST1 . 1 Gene array was used . Data were normalized using the Robust Multi-array Average algorithm and batch effects were adjusted for using ComBat . Transcript expression levels were considered significantly higher than background noise if expression values from robust multiarray analysis in at least 10% of either cases or controls exceeded of the 80% quantile of expression of genes on the Y-chromosome in female hearts ( 5 . 24 ) . Associations of expression levels for expressed genes with SNP genotypes were tested using a likelihood ratio test . Specifically , we fit a linear regression model Y = β0 + β1*D + β2*g + β3* ( g x D ) where Y is the log2 transformed expression level of a given probe , g is the genotype ( coded as 0 , 1 , and 2 ) of the test SNP , and D is heart failure disease status ( D = 1 for heart failure cases and D = 0 for unused donor controls ) . Association between the probe and test SNP was assessed by testing H0: β2 = β3 = 0 using a likelihood ratio test . Significance of the test statistic was evaluated by comparing with a Chi-squared distribution with two degrees of freedom . All models were additionally adjusted for age , gender , and study site . The association of the SNP rs9885413 with DNA methylation was examined in 2408 participants from the FHS Offspring cohort . Methylation at cytosine-guanine dinucleotides ( CpG ) at the 5q22 locus ( +/-500 kb from rs9885413 ) were ascertained from a gene-centric DNA methylation array ( Infinium HumaMethylation450 BeadChip , Illumina , San Diego , CA , USA ) which allows interrogation of 485 , 512 methylation sites across the genome . The array has coverage of at least one methylation site near 99% of RefSeq genes and 96% of CpG islands . Briefly , bisulfite-treated genomic DNA ( 1 μg ) from peripheral blood samples underwent whole-genome amplification , array hybridization and scanning according to manufacturer instructions . Genotyping of rs9885413 was performed as described in S1 Text . Association of rs9885413 and the methylation probe cg02061660 with expression of the five genes at the locus ( +/-500 kb from rs9885413 ) was examined from microarray data ( Affymetrix Human Exon Array ST 1 . 0 ) in 5257 participants from the FHS Offspring cohort and Third Generation cohort . Procedures for RNA extraction , processing and analysis have been described previously ( 28 ) . Linear mixed effect ( LME ) models were fit accounting for familial correlation , cell count heterogeneity and technical covariates to account for batch effects using the pedigreemm package in R [56] . Specifically , the mQTL model utilized a two-step approach: first , the DNA methylation beta-value ( ratio of methylated probe intensity to total probe intensity ) was residualized with adjustment for age , sex , cell count proportions ( imputed using the Houseman method for granulocytes , monocytes , B-lymphocytes , CD4+ T lymphocytes , CD8+ T lymphocytes and NK cells ) [57] , measured technical covariates ( row , chip , column ) , and the family structure covariance matrix . Second , DNA methylation residuals were specified as dependent variable , SNP genotype dosage as independent variable with additional adjustment for 558 SVAs ( surrogate variable analysis ) [58] and ten principal components from eigenstrat [59] to account for unmeasured batch effects . The eQTL models similarly residualized gene expression with adjustment for age , sex , imputed cell count proportions ( imputed in Offspring Cohort participants utilizing gene expression markers of cell count proportions developed from the Third Generation participants with both gene expression and measured complete blood counts ) , and family structure covariance matrix . The residual of gene expression was specified as dependent variable and SNP dosage as independent variable adjusted for 20 PEER ( probabilistic estimation of expression residuals ) factors [60] to account for unmeasured technical and batch effects in the gene expression data . The eQTM models specified gene expression residual as dependent variable and DNA methylation residual as independent variable adjusted for 20 methylation SVAs and 20 expression SVAs to account for unmeasured technical and batch effects . Replication of the association of rs9885413 with cg02061660 including the same covariates in the model as in FHS was attempted in blood samples from 750 randomly selected participants of the Rotterdam study ( RS3 ) not included in the GWA stage , where information from the same DNA methylation array as FHS was available . DNA was extracted , bisulfite-treated using the Zymo EZ-96 DNA-methylation kit ( Zymo Research , Irvine , CA , USA ) and hybridized to arrays according to manufacturer instructions . During quality control samples showing incomplete bisulfite treatment were excluded ( n = 5 ) as were samples with a low detection rate ( <99% ) ( n = 7 ) , or gender swaps ( n = 4 ) . Probes with a detection P-value>0 . 01 in >1% samples , were filtered out . A total number of 474 , 528 probes passed the quality control and the filtered β values were normalized with DASEN implemented in the wateRmelon package in R statistical software . Genotyping was performed using the Illumina 610quad array . Cell counts were estimated using the same method as in FHS and also directly measured on a Coulter AcT Diff II Hematology Analyzer ( Beckman Coulter , Brea , CA ) for granulocytes , monocytes , lymphocytes ) . Models including both estimated and directly measured cell counts were explored . HEK293 cells were seeded at 100 , 000 cells/well in a 6-well plate the day before transfection . Cells were transfected using Lipofectamine and 50 nM of siRNA designed to target human NHLH1 or negative control siRNA ( Life Technologies , Carlsbad , CA , USA ) according to the manufacturer’s instructions . After 48 hours , cells were harvested and total RNA extracted using the miRNeasy Mini Kit ( Qiagen , Hilden , Germany ) according to the manufacturer’s instructions . cDNA was synthesized using the RevertAid H- First Strand cDNA Synthesis Kit ( Thermo Fischer Scientific , Waltham , MA , USA ) using random hexamer primers and qPCR was performed with TaqMan assays for NHLH1 , TMEM232 , SLC25A4 , WDR36 , TSLP , CAMK4 and GAPDH on a StepOne Plus Real-Time PCR System ( Life Technologies ) . Gene expression was normalized to GAPDH and expressed relative to cells transfected with negative control siRNA according to the ΔΔCt-method [61] . The frequencies of ancestral and derived alleles of rs9885413 were examined in populations from the International HapMap Project ( http://www . hapmap . org/ ) [62] and the Human Genome Diversity Project ( HGDP , http://hagsc . org/hgdp/ ) [63] . The fixation index ( Fst ) was estimated as described by Weir and Cockerham [64] , based on allele frequencies in HapMap stage II as also previously described [65] . The integrated haplotype score ( iHS ) was calculated from HapMap stage II data as described by Voight et al ( http://haplotter . uchicago . edu/ ) [33] . Allele frequency distributions in HGDP populations were visualized using the HGDP selection browser ( http://hgdp . uchicago . edu/ ) [66] . Informed consent was obtained from all participants and all contributing studies were approved by the respective ethics committee as described in S1 Text . | In this study , we applied a genome-wide mapping approach to study molecular determinants of mortality in subjects with heart failure . We identified a genetic variant on chromosome 5q22 that was associated with mortality in this group and observed that this variant conferred increased function of an enhancer region active in multiple tissues . We further observed association of the genetic variant with a DNA methylation signature in blood that in turn is associated with allergy and expression of the gene TSLP ( Thymic stromal lymphoprotein ) in blood . Knockdown of the transcription factor predicted to bind the enhancer region also resulted in lower TSLP expression . The TSLP gene encodes a cytokine that induces release of T-cell attracting chemokines from monocytes , promotes T helper type 2 cell responses , enhances maturation of dendritic cells and activates mast cells . Development of TSLP inhibiting therapeutics are underway and currently in phase III clinical trials for asthma and allergy . These findings provide novel genetic leads to factors that influence mortality in patients with heart failure and in the longer term may result in novel therapies . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"Methods"
] | [
"death",
"rates",
"genome-wide",
"association",
"studies",
"medicine",
"and",
"health",
"sciences",
"body",
"fluids",
"demography",
"genome",
"analysis",
"epigenetics",
"dna",
"population",
"biology",
"dna",
"methylation",
"chromatin",
"cardiology",
"genomics",
"chromos... | 2016 | Discovery of Genetic Variation on Chromosome 5q22 Associated with Mortality in Heart Failure |
Vinculin can interact with F-actin both in recruitment of actin filaments to the growing focal adhesions and also in capping of actin filaments to regulate actin dynamics . Using molecular dynamics , both interactions are simulated using different vinculin conformations . Vinculin is simulated either with only its vinculin tail domain ( Vt ) , with all residues in its closed conformation , with all residues in an open I conformation , and with all residues in an open II conformation . The open I conformation results from movement of domain 1 away from Vt; the open II conformation results from complete dissociation of Vt from the vinculin head domains . Simulation of vinculin binding along the actin filament showed that Vt alone can bind along the actin filaments , that vinculin in its closed conformation cannot bind along the actin filaments , and that vinculin in its open I conformation can bind along the actin filaments . The simulations confirm that movement of domain 1 away from Vt in formation of vinculin 1 is sufficient for allowing Vt to bind along the actin filament . Simulation of Vt capping actin filaments probe six possible bound structures and suggest that vinculin would cap actin filaments by interacting with both S1 and S3 of the barbed-end , using the surface of Vt normally occluded by D4 and nearby vinculin head domain residues . Simulation of D4 separation from Vt after D1 separation formed the open II conformation . Binding of open II vinculin to the barbed-end suggests this conformation allows for vinculin capping . Three binding sites on F-actin are suggested as regions that could link to vinculin . Vinculin is suggested to function as a variable switch at the focal adhesions . The conformation of vinculin and the precise F-actin binding conformation is dependent on the level of mechanical load on the focal adhesion .
The focal adhesion is a critical for cell-substrate adhesions [1] , [2] , necessary for Cell movement [3] , [4] , wound healing [5] , cancer cell metastasis [6] , , and other processes [8]–[10] . At the focal adhesion , actin filaments of the cellular cytoskeleton are linked to the extra-cellular membrane ( ECM ) of the substrate [5] . Once linked , both mechanical forces [11] originating from within the cell – such as myosin induced contraction of the actin filaments [12] during cell migration – can act on the ECM and the substrate , and , mechanical forces originating from outside the cell – such as flow induced cyclic stress in the case of endothelial cells [13] – can be transduced to the cellular machinery [14] . Formation of the focal adhesion involves linkage of focal adhesion proteins to ECM bound integrins [15]–[18] , linkage of focal adhesion proteins to each other [19] , [20] , and linkage of the ECM-focal adhesion complex to actin filaments [21]–[23] . The simplest focal adhesion complex would consist of a talin molecule bound to integrin via its head domain and bound to actin via its tail domain [24] , [25] . Talin has 11 cryptic binding sites for vinculin and activation of these binding sites along with subsequent recruitment of vinculin to the growing focal adhesion correlates with the strengthening of the focal adhesion [26] . Recruitment of vinculin would reinforce the focal adhesion as vinculin can crosslink an actin filament to the talin molecule [27] . This binding of the focal adhesion to actin filaments , by vinculin or other focal adhesion forming molecules , is a critical step in completing formation of a mechanical link between the cell and its substrate . The actin filament itself is composed of numerous individual actin subunits bound together to form a polar double-stranded filament [28] . Multiple filaments can be crosslinked by actin crosslinkers [29] , [30] . Both the actin subunits and the actin filament are polar , with a barbed-end ( + ) and a pointed-end ( − ) . In this paper , the actin subunit at the barbed-end is referred to as subunit n , with the next subunit towards the pointed-end referred to as subunit n-1 , and the subsequent subunit to n-1 is referred to as n-2 , and so on ( Figure 1 ) . Polymerization of the actin filament can occur at both ends of actin , but occurs with much higher efficiency at the barbed-end [31] . Each actin subunit has 4 subdomains: S1 , S2 , S3 , and S4 [28] . The S2 subdomain contains a DNase-I-binding loop ( D-loop ) that can interact with a neighboring actin subunit [28] . Recently , it was shown that polymerization of F-actin at the pointed-end is slower than polymerization at the barbed-end because of an interaction between the D-loop at n-1 with a hydrophobic patch of n-2 ( at the pointed-end ) [32] . n-1n-1 Vinculin is a globular protein much smaller than actin and with 5 helical domains: domain 1 ( D1 ) , domain 2 ( D2 ) , domain 3 ( D3 ) , domain 4 ( D4 ) , and the vinculin tail domain ( Vt ) [33] . D1–D4 together form the vinculin head domains . With Vt bound to each of the vinculin head domains , vinculin is considered to be in a closed conformation . Vt contains the likely binding sites for an actin filament [34] , while D1 contains the likely binding sites for talin [35] . In its closed conformation vinculin is unable to bind both F-actin at Vt and talin at D1 [36] , and the closed conformation is often referred to as the auto-inhibited conformation [37] . D1 of the vinculin head inhibits the linkage of Vt with F-actin . Several hypotheses have been explored concerning the mechanism of vinculin activation [38]–[40] . It is clear that a vinculin conformational change is necessary to allow for binding of vinculin to both F-actin and talin [41] . A recent computational study has proposed a conformational change that could activate vinculin: D1 of the vinculin head could move away from Vt and towards its talin binding partner [42] . The movement of D1 leading to vinculin activation could result from a force-induced stretching of vinculin [40] , [43] that would result from vinculin being recruited to mechanically stressed focal adhesions [43] . The computational studies capture one explanation for reinforcement of the linkage between a cell and its surface . Other studies examine additional possibilities such as force-responsive linkage to the substrate via syndecan-4 [44] . Vinculin recruitment to focal adhesions is made possible by activation of the vinculin binding sites ( VBS ) within the talin rod [45] , [46] . The VBS are hydrophobic in nature and buried within the helical structure of talin in the absence of force [47] . Initial computational investigation [48]–[50] and later experimental investigation [45] has demonstrated the exposure of the buried hydrophobic VBS only after force induced stretch of the talin rod . Once activated , the VBS can bind D1 of vinculin . Recent computational simulation suggests the interaction of D1 with VBS is completed , again , after force-induced activation of vinculin [51] . The focal complex serves to link ECM-bound integrin to actin filaments: force-induced activation of talin and activation of vinculin are effective only if the activated vinculin can bind along the actin filament and link it to the focal adhesion . Binding of vinculin to F-actin is the final step to complete this structure . Vinculin-actin binding is necessary for focal adhesions to be mechanically resilient [52] , and the interaction is crucial to the strengthening of the focal adhesion [53] . What features of the vinculin tail and F-actin allow for this crucial interaction ? Using a combination of experimental electron microscopy and computational protein docking methods Janssen et al . [36] have addressed the Vt interaction with F-actin . They suggest two patches of basic residues on the surface of Vt link with two patches of acidic residues on the surface of F-actin . One of the actin acidic patches is on the n-2 subunit and the other is on the n subunit . Their study also suggests that full-length vinculin in its closed conformation would be unable to bind along the actin filament as D1 would clash with regions of the actin filament . It is unclear dynamically how D1 would inhibit the link to F-actin , which of the acidic patches on F-actin are more critical to linking Vt , and whether the suggested conformation of activated vinculin [42] would allow for Vt to link these acidic patches . These issues are addressed in the first section on the binding of vinculin along the actin filament . It has also been suggested that interaction of Vt with F-actin can inhibit actin polymerization [54] . During cell movement actin filaments in the lamellipodia will polymerize at their barbed-ends [31] . The polymerization is involved in membrane protrusion at the leading edge of a migration cell [55] . One line of evidence supporting the notion that Vt can inhibit actin polymerization comes from studies of the bacterial effector IpaA [54] , [56] . IpaA can stop polymerization of actin filaments of its target cell and can even cause depolymerization of the actin filaments . The direct effect of IpaA is to activate vinculin for capping of F-actin at the barbed-end . Although the mechanisms of vinculin activation for F-actin capping by IpaA are not clear , it is suggested from these studies that vinculin can cap the actin filament and prevent its polymerization . Further lines of evidence for vinculin capping of actin filaments come from other studies showing that Vt ( isolated form other vinculin residues ) can catalyze G-actin nucleation through interaction with the barbed-end of G-actin [57] . Most recently , Le Clainche et . al . [58] have explored capping of F-actin by vinculin in vitro . They used pyrenyl-labeled actin fluorescence [59] , [60] to assay the polymerization of G-actin into F-actin before and after introduction of Vt in vitro . Introduction of Vt prevents polymerization of F-actin . Their results suggest that residues 1044–1066 of Vt are critical to capping of F-actin by vinculin . Several questions arise concerning this capping of actin filaments by vinculin that the second section of this study on the capping of F-actin will address: what is the structure of F-actin capped by Vt ? What residues and surface regions of the barbed-end are critical to interaction with Vt ? How favorable or stable are these interactions between the barbed-end of F-actin and Vt ? The interaction between vinculin and actin is of importance not only to efforts aimed at understanding focal adhesion formation via talin and vinculin , but also to efforts aimed at understanding the role of vinculin in regulating actin dynamics , or the role of vinculin in other cellular processes . A study by Wilins and Lin [61] established a role of vinculin in regulating actin dynamics , and more recently Huveneers et al . suggested vinculin to be involved in stabilizing force-dependent remodeling of endothelial cell-cell adhesions [62] . This study investigates both the interaction of vinculin along the actin filament and the capping interaction of vinculin with the barbed-end of F-actin . Molecular dynamics simulations are used to probe the interaction of vinculin along the actin filament using: ( a ) a structure of only Vt interacting with actin subunits n and n-1 , ( b ) a structure of vinculin is its closed conformation , and ( c ) a structure of vinculin in its suggested activated conformation [42] . Furthermore , a similar molecular dynamics approach is used to determine the likely structure , dynamics , and energetics of the interaction between Vt and the capping end of F-actin . Finally , computational techniques are used to evaluate an additional conformational change in full-length vinculin ( beyond the suggested activation of vinculin at D1 ) and explore the possibility of an interaction between the barbed-end of the actin filaments and vinculin in this second open structure .
Using pyrenyl-labeled actin to assay F-actin polymerization Le Clainche et al [58] demonstrated in vitro that Vt can effectively cap the barbed-end of actin . Capping of the actin filaments by vinculin has been demonstrated in cells affected by IpaA [71] . In such cells the capping of F-actin serves to depolymerize the actin filaments and make the cells compliant for Shigella invasion [54] . It is unclear if capping of actin filaments could play a role at sites of focal adhesions . It is also unclear if vinculin at focal adhesions is able to cap the actin filaments . A step towards clarifying both possibilities is to understand the nature of F-actin capping by vinculin . The vinculin tail residues implicated in F-actin capping reside in the C-terminus region and have also been implicated in interaction with the lipid membrane [72] . The last 21 residues of vinculin consist of a number of charged and basic residues that are predicted to readily interact with acidic residues at the actin barbed-end or on acidic phospholipids ( Figure 4A ) . In its closed conformation vinculin head domains occlude access to most of these residues . With Vt isolated from the vinculin head domains the last 21 residues could interact with the barbed-end of F-actin either through the surface of Vt that would be occluded by vinculin head domains , the occluded surface , or through the surface of Vt already exposed to solvent , the exposed surface of Vt . From the structure of Vt it is predicted that the occluded face of Vt would better link the barbed-end given the higher density of charged residues at this surface ( Figure 4A ) . Newly added actin subunits would interact with the barbed-end of F-actin . Examination of the F-actin structure predicts that the D-loop of subunit n interacts favorably with the interface between S1 and S3 of subunit n-2 ( Figure 1 ) . Specifically , residues 283–294 , 139 , 140 , 143 , 346 , 351 , and 374 of the barbed-end are implicating in stabilizing additional actin subunits by interacting with their D-loop structures [28] . Recent high resolution imaging of the actin filament pointed-end confirms the likely interaction between the D-loop of a newly polymerized actin monomer and the interface between S1 and S3 at the barbed-end of the actin filament [32] . Capping of F-actin by CapZ and other capping proteins prevents actin polymerization by occluding access to S1 or S3 [32] . Capping of F-actin by Vt would then likely result from interaction of Vt with S1 of the barbed-end , S3 of the barbed-end , or both subunits S1 and S3 of the barbed-end ( Figure 4B ) . The interaction of Vt with F-actin is evaluated using molecular dynamics . Vt is simulated initially oriented towards the barbed-end towards either ( A ) S1 only , ( B ) S3 only , or ( C ) towards S1 and S3 . Each orientation is simulated both with the exposed face of Vt initially oriented towards the barbed-end and with the occluded face of Vt initially oriented towards F-actin ( Table S1 ) . An additional conformational change is necessary for vinculin to be able to cap actin filaments . The simulations of Vt capping of actin filaments suggest that the occluded surface of Vt forms the most likely interaction with the interface between S1 and S3 . The occluded surface is normally in contact with D4 residues . Movement of D1 away from Vt , which was shown to be sufficient for allowing vinculin to bind along actin filaments , does not result in dissociation of Vt from D4 . The occluded surface of Vt , critical to F-actin capping , requires additional conformational changes in vinculin beyond D1 movement . Le-Clainche et al [58] , [78] also suggest a second vinculin conformational change is necessary to allow for vinculin capping of the actin-filaments . The interface between D4 and Vt has been implicated elsewhere as critical to Vt activation in general . Chen et al [78] describe a pincer-like mechanism to vinculin activation in which both the interface of D1 with Vt and the interface of D4 with Vt is disrupted to allow Vt to leave the pocket formed by the vinculin head domains and link with F-acitin . In another study , Cohen et al [65] demonstrate that both interactions between Vt and D1 and interactions between Vt and D4 are critical to maintaining an auto-inhibited conformation . The simulations in this study suggests that disruption of the key interactions between Vt and D1 coupled with movement of D1 away from Vt is sufficient to allow binding of vinculin along the actin filament . The interaction between Vt and D4 could however play a critical role in regulating Vt capping of actin filaments . The separation of D1 from Vt was simulated by assuming a cooperative activation mechanism and introducing stretch of vinculin to be consistent with that mechanism [42] . The source for D4 separation from Vt is less clear . If Vt separates from D4 at the focal adhesions then perhaps the movement of the actin filaments across the developing focal adhesion can supply induce separation of Vt from D4 . If however Vt separation from D4 is particular to cells affected by Shigella [71] , then D4 separation would result from the interaction with IpaA . Whatever the source allowing fro D4 separation , it is likely that D4 would separate after D1 separation . In either scenario , the magnitude of force that would be needed to induce a second conformational change is telling of how likely it is that a conformation shift would occur . Estimating the in vivo magnitude of force accurately is not possible with molecular dynamics . The time-scale of computationally feasible molecular dynamics simulations is orders of magnitude faster than the in vivo time-scale . Nevertheless , the conformational changes we suggest here can be informative . The interaction between vinculin and actin was explored in this paper using molecular dynamics simulations in three sections: first , the interaction of vinculin along the actin filament was investigated , then , the interaction of Vt with the barbed-end of the actin filament , and finally , the possible interaction between an open II vinculin conformation and the actin filament . Simulation of the interaction along the actin filament confirmed that although Vt can bind along the actin filament , full-length vinculin in its closed conformation is inhibited from binding along the filament . The open I conformation , previously suggested as a conformation of activated vinculin , was able to bind along the actin filament as Vt had , confirming that it is likely the structure of activated vinculin . Simulation of Vt interacting with the barbed-end of F-actin confirmed that Vt could indeed prevent polymerization of the actin filament . Vt binds to the barbed-end of F-actin similar to other capping proteins [74] and can prevent the association of a new actin subunit with S1 and S3 . Simulation of vinculin conformational changes beyond D1 separation and formation of the open I conformation revealed the possibility of an open II conformation . With the open II conformation vinculin could cap actin filaments , even at the focal adhesion . In general , evaluating binding modes between a large filament such as actin and a large protein such as vinculin is a computationally demanding endeavor . In addition to the computational challenge posed by the size of the proteins , the simulation time needed to allow for a binding event to occur can be prohibitive . In the context of these computational challenges , this study has simulated the binding events with the following strategy: ( 1 ) reduce the size of the binding proteins by limiting the number of actin monomers to include in the simulations , ( 2 ) reduce the simulation time by including an initial nudge of vinculin towards actin , and ( 3 ) reduce simulation time by placing vinculin within 15 Å of actin . This has allowed for simulation of binding , but the study is limited by not having repeat binding events to capture the true scholastic nature of the binding . An interesting scenario arises when considering the linkage of vinculin to actin at focal adhesions: vinculin is also linking to talin at focal adhesions , and so how are the three to be relatively oriented ? In what order would the two binding events occur ? The results from this study are insufficient to answer those questions accurately and additional simulations would be required to predict the binding mode of a talin-vinculin-actin complex . The results here suggest that steric limitations between talin and actin should govern the exact order of binding events , or exact binding modes that would be adopted . Putting together the results from all of these simulations , we can predict three potential regions of an actin filament that would bind vinculin ( Figure 10 ) . Consistently , the acidic residues in S1 – vinculin binding site A – were shown to be critical for an interaction between Vt and F-actin . These residues stabilized both the binding of vinculin along the actin filament and they were involved in stabilizing Vt capping of the actin filament . The surface between S1 and S3 of the barbed end – vinculin binding site B – was also consistently shown to stabilize Vt . Hydrophobic residues in this region would form hydrophobic cores with non-polar residues from Vt . Both basic and acidic residues in this region would form salt-bridges with their counterparts on Vt . The interactions between Vt and S1 that were highlighted by our simulations had previously been suggested to be involved in binding of vinculin along the actin filament , and the interactions between S1 and S3 were previously shown to be significant for capping of the actin filament . However , the third region on the surface of the actin filament that is suggested here to be involved in vinculin binding is novel: the residues in S3 of subunit n-2 – vinculin-binding site C . These residues can interact with charged residues from D1 of vinculin and in doing so contribute to further stabilizing the vinculin-actin linkage . With the presence of three binding regions we can predict that vinculin will differentially bind to each of the binding sites depending on the intensity of mechanical stress on the focal adhesion . It is possible that binding site A would interact with vinculin during vinculin activation , and would completely bind vinculin after activation . This interaction would require the least level of mechanical stress . Binding site C would bind vinculin after D1 is separated from Vt and can link to it . This would potentially require some level of mechanical stress . And binding of binding site B to vinculin would occur after transition of vinculin to the open II conformation . This would require the most level of mechanical stress . The exact binding interface between vinculin and F-actin , therefore , would be a function of the level of mechanical stress at the focal adhesion . Vinculin would be a variable switch at the focal adhesion , increasing its level of activation and F-actin binding depending on the level of mechanical stress at the focal adhesion . Previously , it was shown that formation of the vinculin open I conformation allows for the complete insertion of talin VBS in to D1 of vinculin [51] . We can now expand on those results and further state that after complete linking of vinculin to talin via D1 , and linking along the actin filament via Vt in the open I conformation , any additional forces from further movement or stress of the actin filament could induce an open II conformation ( Figure 11 ) . The formation of an open II conformation would then allow vinculin to remain linked to F-actin even as it moves . In this way vinculin could act both as a molecular clutch and as a variable switch at the focal adhesion . The predictions from this study are especially relevant to understanding focal adhesion structures . Focal adhesions play a role in numerous cell types and are especially involved in the processes of cell migration . The predictions from this study contribute towards understanding molecular mechanisms of cell migration via the focal adhesions . The question that remains unanswered after our simulations and analysis is whether the actin filaments can be capped at the focal adhesion . The simulations of vinculin in an open II conformation with the barbed-end suggested that F-actin can be capped , but what role would this play at the focal adhesion ? The understanding that vinculin is a variable switch is a testable and valuable prediction from the molecular dynamics simulation , but the capping of actin filaments at focal adhesions is really an unanswered question that is posed by our molecular dynamics simulations . Further investigations – both computational and experimental – are sought to address this question .
PDB ID 1ST6 was used to build a structure of full-length vinculin [66] . The missing proline rich linker region ( residues 843–877 ) was created via homology modeling using the SWISSMODEL toolkit [82] , as previously described [42] . The proline rich linker region is suggested to be flexible and therefore its structure not resolved . Inclusion of a homology model for the linker region in this study is justified as the linker region is not suggested to play a key role in the binding events , and the simulations should not be affected by inclusion of the linker homology model . Vt was built as residues 895–1066 from the full-length vinculin model . The structure of open full-length vinculin was taken from previously published simulation [42] . PDB ID 3LUE [66] was used to build a structure of F-actin . The 3LUE structure has α-actinin [83] CH domains bound to F-actin . The CH-domains are removed and only 3 of the F-actin subunits are used to build the F-actin structure . Complex of Vt bound to F-actin was build using Janssen et al [36] to orient Vt towards the two binding pockets along F-actin . Structure of full-length was build using the same Vt orientation but including the vinculin head domain residues . Vinculin was translated to be at least 15 Å away from F-actin . For simulation of Vt interaction with the barbed end , Vt was oriented either with the exposed or the occluded surface oriented towards Vt . Three arrangements of Vt with the barbed-end were simulated: Vt oriented towards S1 of the barbed-end of F-actin , Vt oriented towards S3 of the barbed-end of F-actin , Vt oriented towards both S1 and S3 of the barbed end . Vt was placed within 10 Å of the barbed-end in these structures . The initial orientation and setup of each simulation was random with respect to the exact pose of vinculin relative to actin; the distance from actin was imposed to be 10 Å and the face of vinculin ( exposed or occluded surface ) closest to actin was controlled , but the exact orientation pose of vinculin was chosen at random . It is possible the orientation pose of vinculin directly impacted the final binding mode , but given that the final binding orientation and pose of vinculin bound to actin was government by the mechanics of interaction , the final binding mode can be seen as more representative . For simulation with vinculin in the closed or the open II conformation , additional vinculin head residues are included with maintaining the Vt orientation . Each system was solvated with 12 Å of padding at each end of the simulation box . Simulations were carried out using the NAMD Scalable Molecular Dynamics program [84] , using an explicit solvent representation . Periodic boundary conditions were used along with a Langevin piston Nose-Hoover [85] mechanism for pressure control at 1 Atm . Constant temperature of 310K was maintained using a Langevin damping coefficient of 5/ps . Rigid bonds were enforced between hydrogen atoms and their bound larger atoms [86] . The CHARMm 27 force fields were used [87] , [88] . Simulation timesteps of 2 fs were used for all molecular dynamics . Each configuration was first minimized for 1000 steps using the conjugate gradient and line search algorithm implemented in NAMD [84] . Following minimization each configuration is simulated for at least 15 ns or until equilibration . All simulation results were visualized and analyzed using VMD [89] . For simulations of binding along F-actin the Vt , closed vinculin , and open I vinculin were initially nudged towards F-actin for less than 1 ns prior to simulation for 15 ns using a constant velocity pull on the center of mass of vinculin in a direction towards the center of mass of actin . Use of the nudge reduces the entropic barrier to binding . For simulation of Vt capping actin filaments no initial nudge is used and instead Vt is placed 5 Å closer to the barbed-end . Vt is smaller than full-length vinculin and can have faster translation , thus binding occurred even without an initial nudging force . The simulations are performed one time per setup . Additional simulations would allow for a statistical estimation of reproducibility , however , given computational limitations to running multiple 15 ns simulations , the present study is limited to a single simulation per setup . Umbrella sampling of D4 separation from Vt was carried out using GROMACS [90] . The umbrella sampling approach allows estimation of a free energy path along a reaction coordinate by estimation of the free energy difference between subsections of the path . The reaction coordinate was defined as the distance between the center of mass of D4 and the center of mass of Vt . Residues in D1 16 , 51 , 81 , and 115 were constrained with 1000 KJ/mol*nm2 to maintain an open I conformation throughout the simulation . Residues 926 , 958 , 988 , and 1031 of Vt were defined as the pull group and constrained along the reaction coordinate away from residues 730 , 760 , 794 , and 824 of D4 . An umbrella potential of 1000 KJ/mol*nm2 was used with a reference step of 0 . 2 Å in order to maximize umbrella overlap , allowing for an accurate estimation of the free energy path . The final potential of mean force was calculated using Grossfield's WHAM code [91] . | The interface between a cell and its substrate is strengthened by the formation of focal adhesions . In this study molecular dynamics simulations are used to explore the connectivity of one focal adhesion forming protein , vinculin , and the cytoskeletal filament , F-actin . The simulations demonstrate: ( 1 ) that vinculin can link along F-actin at these focal adhesions when it adopts an open conformation , ( 2 ) that the vinculin tail ( Vt ) can bind F-actin at its barbed-end preventing actin polymerization , ( 3 ) that vinculin can adopt two open conformations , and ( 4 ) that the second open conformation is necessary for vinculin to cap the actin filament . The results suggest that vinculin can act as a variable switch , changing its shape and the nature of its interaction with F-actin depending on the level of stress seen at a focal adhesion . Under the highest stress vinculin would adopt the open II conformation and link anywhere on F-actin , even its barbed-end . Under less stress vinculin could adopt the open I conformation and bind along F-actin . And under minimal stress vinculin could adopt its closed conformation . This variability allows for vinculin to truly function as the cell's mechanical reinforcing agent . | [
"Abstract",
"Introduction",
"Results/Discussion",
"Methods"
] | [
"bioengineering",
"biological",
"systems",
"engineering",
"biology",
"computational",
"biology",
"engineering"
] | 2013 | The Interaction of Vinculin with Actin |
Chronic chagasic cardiomyopathy ( CCC ) is observed in 30% to 50% of the individuals infected by Trypanosoma cruzi and heart failure is the important cause of death among patients in the chronic phase of Chagas disease . Although some studies have elucidated the role of adaptive immune responses involving T and B lymphocytes in cardiac pathogenesis , the role of innate immunity receptors such as Toll-like receptors ( TLRs ) and Nod-like receptors ( NLRs ) in CCC pathophysiology has not yet been determined . In this study , we evaluated the association among innate immune receptors ( TLR1-9 and nucleotide-binding domain-like receptor protein 3/NLRP3 ) , its adapter molecules ( Myd88 , TRIF , ASC and caspase-1 ) and cytokines ( IL-1β , IL-6 , IL-12 , IL-18 , IL-23 , TNF-α , and IFN-β ) with clinical manifestation , digestive and cardiac function in patients with different clinical forms of chronic Chagas disease . The TLR8 mRNA expression levels were enhanced in the peripheral blood mononuclear cells ( PBMC ) from digestive and cardiodigestive patients compared to indeterminate and cardiac patients . Furthermore , mRNA expression of IFN-β ( cytokine produced after TLR8 activation ) was higher in digestive and cardiodigestive patients when compared to indeterminate . Moreover , there was a positive correlation between TLR8 and IFN-β mRNA expression with sigmoid and rectum size . Cardiac and cardiodigestive patients presented higher TLR2 , IL-12 and TNF-α mRNA expression than indeterminate and digestive patients . Moreover , cardiac patients also expressed higher levels of NLRP3 , ASC and IL-1β mRNAs than indeterminate patients . In addition , we showed a negative correlation among TLR2 , IL-1β , IL-12 and TNF-α levels with left ventricular ejection fraction , and positive correlation between NLRP3 with cardiothoracic index , and TLR2 , IL-1β and IL-12 with left ventricular mass index . Together , our data suggest that high expression of innate immune receptors in cardiac and digestive patients may induce an enhancement of cytokine expression and participate of cardiac and digestive dysfunction .
Chagas disease is caused by the Trypanosoma cruzi ( T . cruzi ) parasite , and affects about 6 million to 7 million people in Latin America; moreover , 1 . 2 million people have chronic chagasic cardiomyopathy ( CCC ) , which is the main cause of 12 , 000 deaths annually [1 , 2] . In the chronic phase of Chagas disease approximately 30 to 50% of patients develop cardiac disease , 5 to 15% develop digestive disease and 2 to 10% develop cardiodigestive form [2–4] . Neuronal depopulation and changes in cardiac and myenteric plexus conduction systems are fundamental for the pathophysiology of CCC , megaesophagus and megacolon . Cardiac conduction system may be diffusely affected from the sinus node to the distal third of His bundle . Chagasic patients with CCC have right bundle branch block ( present in 13 to 35% of patients ) as an electrocardiographic alteration most frequently suggestive of Chagas' disease , often associated with anterosuperior left bundle branch block [5] . Ventricular extrasystole occurs ( 15% to 55% of patients ) usually isolated , but when complex or associated with other electrocardiographic alterations are correlate with left ventricular systolic and diastolic function , and diastolic diameter [6–8] . The severity of ventricular arrhythmia is often correlated with the degree of left ventricular dysfunction , although some patients with CCC and ventricular tachycardia or ventricular atrial block have preserved global ventricular function [4 , 6] . Sudden death is more frequent in males ( mainly between 30 and 50 years of age ) . Other electrocardiographic changes observed in Chagas' disease are represented by low voltage of the QRS complex , notches and abnormal thickenings , low amplitude or absence of R wave in precordial derivations [9 , 10] . Inflammatory process involves fibrosing and progressive chronic myocarditis is also the key substrate for impairment of the conduction system in Chagas disease [5] . Inflammatory cytokines ( IL-12 , IFN-γ and TNF-α ) , nitric oxide , autoantibodies , CD8+ T lymphocyte are possible correlated with neuronal depopulation [11–15] . Cardiac disease has been correlated with immunological unbalance . Patients with CCC have high production of inflammatory cytokines such as IFN-γ , TNF-α , IL-1β and nitric oxide ( NO ) which are involved with myocarditis , fibrosis and myocardial hypertrophy . In contrast , asymptomatic patients produce high levels of IL-10 which support control of the inflammatory mechanism in the heart [16–19] . The cardiac form of Chagas disease is related to an increase of T helper ( Th ) type 1 cells and a decrease of Th2 , Th9 , Th17 , Th22 and regulatory T cell response [11 , 16] . In fact , exacerbated inflammatory process in cardiac patients has been associated to enhancing the risk of stroke and death [11] . Autoantibodies and CD8+ T lymphocytes have also been related to CCC immunopathogenic mechanism [20–22] . TLRs and NLRs are families of pattern recognition receptors ( PRRs ) located in the plasma membrane , endosomes and cytosol , and are mainly expressed by professional antigen presenting cells ( APC ) , endothelial cells and fibroblasts . PRRs are responsible for recognizing different chemical structures highly conserved in microorganisms known as Pathogen-Associated Molecular Patterns ( PAMPs ) . Signaling through TLRs and NLRs induce the transcription of genes involved in inflammatory response , and its role in the experimental T . cruzi infection has been investigated . T . cruzi contains a variety of ligands such as glycosylphosphatidylinositol ( GPI ) anchors of mucin-like glycoproteins , glycoinositolphospholipid ( GIPL ) and nucleic acids which activate different PRRs [23–26] . In fact , a deficiency of Myd88 , TLR4 , TLR7 and TLR9 lead mice to being more susceptible to T . cruzi infection [27–29] . However , TLR2 signaling and NF-κβ activation induce pro-IL-1β production , which triggers cardiomyocyte hypertrophy in T . cruzi infected rats [30] . Chagasic patients with a decrease in signal transduction upon ligation of TLR2 or TLR4 to their respective ligand may exhibit low NF-κβ activation and have a low risk of developing CCC [31] . NLRs were extensively characterized as PRRs for bacterial and viral infection [32–34] , and their role in recognizing intracellular parasites has been studied . Knockout mice for NOD1 are more susceptible to T . cruzi infection . Bone marrow-derived macrophages from NOD1 knockout mice show a reduction of products dependent on NF-kB activation and fail to control the infection in the presence of IFN-γ [35] . NLRP3 inflammasome signaling activates apoptosis-associated speck–like protein containing a caspase recruitment domain ( ASC ) and caspase-1 , thereby inducing the cleavage of pro-IL-1β and pro-IL-18 in their active forms [36–38] . ASC inflammasomes are critical determinants of host resistance to infection with T . cruzi . NLRP3-/- , ASC-/- and caspase-1-/- mice exhibit a higher mortality , cardiac parasitism , and myocarditis than wide type mice [39 , 40] . However , T . cruzi NLRs agonists are not known . Although several studies have elucidated the role of TLRs and NLRs in experimental infection by T . cruzi , the role in human CCC pathophysiology has not yet been determined . The activation of TLRs and NLRs is important in directing adaptive responses , thus resulting in macrophage activation which are important cells involved in heart disease [31 , 41] . In this study , we have described an increase in several innate components such as NLRP3 , ASC , TLR2 , IL-1β , IL-12 and TNF-α associated with the pathophysiology of CCC in humans . Moreover , the digestive form of chronic Chagas disease was correlated to high TLR8 and IFN-β mRNA expression . A better understanding of immunological mechanisms involved in CCC may lead to reduced morbidity and mortality associated with the cardiac form of the disease .
The population was composed of 65 individuals aged between 18 and 79 years old from an endemic area of Chagas disease in Rio Grande do Norte , Northeast , Brazil , as described previously [11] . The individuals were selected using two different serological methods ( Chagatest , recombinant ELISA and HAI , and indirect immunofluorescence assay ) in accordance with recommendations of the World Health Organization and the Brazilian Consensus of Chagas Disease II [42] . Western blot ( TESAcruzi® , BioMérieux , Brazil ) confirmatory sorological test was performed [43] . Informed consent was obtained from the participants and the study was approved by the Research Ethics Committee of the State University of Rio Grande do Norte ( UERN ) under protocol number 027 . 201 , and a Certificate of National System of Ethics in Research ( CAEE—SISNEP ) with protocol number 0021 . 0 . 428 . 000–11 . The study was performed according to human experimental guidelines of the Brazilian Ministry of Health and the Helsinki Declaration . Individuals with confirmed positive serology to Chagas disease were clinically evaluated including electrocardiogram ( ECG ) mapping and chest X-ray , 2D-echocardiogram ( ECHO ) and 24h Holter examination . Chagasic patients ( Table 1 ) were classified as indeterminate ( n = 18 ) , cardiac ( n = 17 ) , digestive ( n = 15 ) and cardiodigestive ( n = 15 ) clinical forms , according to the World Health Organization and Brazilian Consensus of Chagas Disease [42] . Uninfected healthy individuals ( n = 15 ) were used as controls . Clinical evaluations were performed in all chagasic patients as previously described [3] . First , plain posteroanterior and lateral chest radiography were performed to evaluate the cardiothoracic index , and which was considered abnormal if attaining a value >0 . 5 [3] . Esophageal contrast radiography was performed in right anterior oblique position using barium sulfate ( Bariogel® , Cristália Laboratory , Brazil ) classifying the esophagus changes into four groups [44] . Contrasted colon radiographs were performed in the supine , ventral and right lateral position [45] using barium sulfate solution ( Bariogel ® , Cristália Laboratory , Brazil ) via the rectum without prior bowel preparation or double contrast use . The sigmoid was classified into four grades ( zero to three ) according to Silva et al . [46] modified by Andrade and coworkers [3] . Radiographic examinations were performed using radiology equipment with X-ray penetration to deep parts ( VMI® , Brazil ) . Electrocardiographic alterations were determined using a portable EP3 2008 electrocardiograph ( Dixtal , Brazil ) with three channels and 12-lead; electrocardiographic recording was based on the Minnesota Code modified , adapted for Chagas disease [10] . Next , conventional , parasternal , supra sternal , apical , subcostal transthoracic echocardiogram and its variations were performed in all patients to calculate cardiac dimension and volumes in accordance with the recommendations of the American Society of Echocardiography [47] and using echocardiography with color flow mapping performed in standard views ( General Electric Healthcare , USA ) . The left ventricular ejection fraction ( LVEF ) was calculated according to the modified Simpson's rule ( biplane method ) [47] . The left ventricular mass index ( LVMI ) was calculate by the formula LVMI = heart mass ( g ) / patient's body surface ( m2 ) . Patients with cardiomegaly , electrocardiographic or echocardiographic alterations suggestive of Chagas disease underwent electrocardiographic monitoring for 24 hours ( 24-Holter ) using a Cardiolight Digital Recorder ( Cardios , São Paulo , Brazil ) . Innate immune receptors ( TLR1 , TLR2 , TLR3 , TLR4 , TLR5 , TLR6 , TLR7 , TLR8 , TLR9 and NLRP3 ) , signaling molecules ( Myd88 , TRIF , ASC , Caspase-1 ) and cytokine ( IL-1β , IL-6 , IL- 12 , IL-18 and TNF-α ) mRNA expression were detected by Real-Time PCR ( qPCR ) in peripheral blood mononuclear cells ( PBMC ) obtained from chagasic patients . Total RNA was obtained using Trizol reagent ( Invitrogen™ , Carlsbad , CA , USA ) and SV Total RNA Isolation System ( Promega , Madison , WI , USA ) with DNase treatment step . cDNA synthesis was performed with the High Capacity cDNA Reverse Transcription kit ( Applied Biosystems , USA ) using the Eppendorf Mastercycler gradient set ( Eppendorf , USA ) . The qPCR reactions were performed using SYBR Green ( Applied Biosystems , USA ) supported by 7500 Fast Real time thermocycler ( Applied Biosystems , Warrington , USA ) . The reactions were performed in 96 well plates ( MicroAmp® , Applied Biosystems , USA ) and the standard PCR conditions were as follows: 50°C ( 2 min ) and 95°C ( 10 min ) followed by 40 cycles of 94°C ( 30 s ) , variable annealing primer temperature ( Table 2 ) ( 30 s ) , and 72°C ( 1 min ) . Specific primers ( Table 2 ) were obtained by the Primer Express software ( Applied Biosystems , USA ) . The mRNA expression levels of the innate immune receptors , adapter molecules and cytokines were determined using the mean Ct values from triplicate measurements to calculate the relative expression levels of the target genes in the Chagas disease patients compared to healthy controls , and were normalized to the housekeeping gene β-actin using the 2–ΔΔCt formula . Cytokine quantification was performed in sera from indeterminate ( n = 18 ) , cardiac ( n = 17 ) , cardiodigestive ( n = 15 ) and digestive ( n = 15 ) chagasic patients . Uninfected individuals were used as a control ( n = 15 ) . The ELISA sets were IL-1β , IL-12 ( p70 ) and TNF-α ( BD OptEIATM , BD Bioscience ) , and procedures were performed according to the manufactures`instructions . Optical densities were measure at 450ηm . Data are reported as mean ± standard deviation ( SD ) . Kolmogorov-Smirnov test was used to verify parametric or non-parametric data distribution . The mRNA expression levels were compared using the Kruskal-Wallis test . Correlations among left ventricular ejection fraction , esophagus and colon dilation , innate immune receptors and cytokines were performed using the Spearman test . Differences were considered significant when p <0 . 05 . Our analyses were performed using PRISM 5 . 0 software ( GraphPad , CA , USA ) .
Chagasic patients ( n = 65 ) were classified as indeterminate ( n = 18 ) , cardiac ( n = 17 ) , digestive ( n = 15 ) and cardiodigestive ( n = 15 ) clinical forms . Chest X-ray demonstrated cardiomegaly in approximately 10% of cardiac and cardiodigestive patients . Electrocardiographic changes were not always associated with cardiac symptoms . Three cardiac patients had right bundle branch block , two also had anterosuperior divisional block . Four cardiodigestive patients presented right bundle branch block , two also presented anterosuperior divisional block . All ventricular atrial blocks were first degree , except for one patient with ventricular atrial blockade who received pacemaker implantation . The echocardiogram showed similar diastolic diameters and left ventricular mass index in indeterminate , cardiac , digestive and cardiodigestive chagasic patients ( Table 1 ) . In an attempt to elucidate the inflammatory mechanism involved in CCC development we analyzed the mRNA expression of innate immune receptors in chagasic individuals grouped according to clinical forms as indeterminate , cardiac , digestive and cardiodigestive ( Table 1 ) . Patients with different clinical manifestations of Chagas disease showed similar expression of TLR1 , TLR3 , TLR4 , TLR5 , TLR6 , TLR7 and TLR9 mRNA ( Fig 1 ) . Interestingly , cardiac and cardiodigestive patients presented higher TLR2 mRNA expression than indeterminate and digestive patients ( Fig 1B ) . Furthermore , cardiodigestive patients presented higher Myd88 mRNA expression than indeterminate and cardiac patients ( Fig 2A ) . Cardiac patients showed higher mRNA expression of IL-12 and TNF-α transcripts ( cytokines produced upon TLR activation ) than indeterminate patients ( Fig 2B and 2C ) . We observed similar expression of TRIF , IL-6 , IL-23 and IFN-α in chagasic patients with different clinical manifestations of Chagas disease ( Fig 2D–2G ) . Moreover , there was higher production of inflammatory cytokines ( TNF-α and IL-12 ) induced by the TLRs activation in sera in cardiac patients than in indeterminate and uninfected controls ( Fig 3A and 3B ) . However , no significant difference was observed between the levels of IL-1β between the different groups of patients ( Fig 3C ) . Together , these data indicate that TLR2 expression in cardiac patients may induce an enhancement of IL-12 and TNF-α expression and correlate to cardiac dysfunction . The mRNA expression of TLR8 was enhanced in digestive and cardiodigestive patients compared to indeterminate and cardiac patients . Furthermore , the mRNA expression of IFN-β ( cytokine produced after TLR8 activation ) was higher in digestive and cardiodigestive patients when compared with indeterminate ( Fig 2H ) . In attempt to evaluate the TLR8 and IFN-β participation in the development of the digestive form of Chagas disease , we analyzed the correlation between the mRNA expression of TLR8 and IFN-β with the rectum and sigmoid size , resulting in a positive correlation observed between TLR80020and IFN-β and rectum and sigmoid size ( Fig 4A–4D ) . We subsequently analyzed the expression of NLRP3 inflammossome , its signaling molecules ( ASC and caspase-1 ) and cytokines produced after its activation ( such as IL-1β and IL-18 ) . Cardiac patients showed relevantly higher mRNA expression of NLRP3 , ASC and IL-1β than indeterminate patients ( Fig 5A–5C ) . Similar levels of caspase-1 and IL-18 mRNA expression were observed in patients with different clinical forms of chronic Chagas disease ( Fig 5D and 5E ) . We then posteriorly analyzed the correlation between the mRNA expression of TLR2 , NLRP3 , IL-1β , IL-12 and TNF-α with the left ventricular ejection fraction ( LVEF ) and cardiothoracic index ( CI ) . We found a negative correlation among NLRP3 , TLR2 , IL-12 and IL-1β mRNA expression with LVEF ( Fig 6A–6D ) , and positive correlation between NLRP3 with CI ( Fig 7A ) . No correlation was observed between TLR2 , IL-1β and IL-12 with CI ( Fig 7B–7D ) , and between NLRP3 with left ventricular mass index ( LVMI ) ( Fig 8A ) . We also observed a positive correlation between LVMI with TLR2 , IL-1β and IL-12 ( Fig 8B–8D ) .
Pathophysiological mechanisms involved in the development of chronic chagasic cardiomyopathy ( CCC ) have been studied in chagasic patients and several immunopathogenic mechanisms involving the participation of adaptive immune response such as CD4+ T helper response [11 , 16–18 , 48 , 49] , CD8+ T cells [50–53] , and autoantibodies production [20 , 21 , 54–56] have been elucidated . However , the role of innate immunity receptors in the CCC pathophysiology has not been elicited . In this study , we have assessed the expression of Toll-like Receptors and Nod-like Receptors , their adapter molecules and induced cytokines in cardiac patients , and compared them to indeterminate , digestive and cardiodigestive clinical forms of the disease . We initially analyzed the mRNA expression of TLRs in PBMCs obtained from the same chagasic patients described in previous study [11] . Patients who showed digestive and cardiodigestive clinical forms presented higher TLR8 mRNA expression when compared to cardiac and indeterminate patients . Human TLR8 recognizes single-stranded RNAs from RNA viruses , as well as detecting RNAs from bacteria in endosomes of dendritic cells [57] . Patients with the digestive clinical form are mainly characterized by the presence of a megaesophagus and megacolon which are caused by the destruction of intramural autonomic ganglia [58] . Gastrointestinal dysfunction can change the feed flow and is associated with bowel inflammatory lesions which distort epithelium gastrointestinal homeostasis , which in turn could allow bacteria penetration and TLR8 activation . This phenomenon could be associated with the development of digestive pathology in chronic chagasic patients . In fact , patients with ulcerative colitis have higher TLR8 mRNA in colon biopsies than healthy subjects , probably due to bacterial RNA of gut microbiota resulting from microbiota dysbiosis [59 , 60] . Intestinal inflammation intensity in the ulcerative colitis is positively correlated with TLR8 and inflammatory cytokines such as IL-6 and TNF-α [60] . Susceptibility to Crohn's disease ( another intestinal inflammatory disease ) has also been associated to high TLR8 levels [61 , 62] . TLR8 activation induces pro-inflammatory cytokine production such as IL-1β , IFN-α , IFN-β , TNF-α , IL-6 , and IL-12 in PBMCs , monocytes and dendritic cells in patients [57 , 63] . Furthermore , in this study we observed higher mRNA expression of TLR8 and IFN-β in digestive and cardiodigestive patients when compared to indeterminate patients . Pathophysiologic alterations in the digestive system during Chagas disease result from the destruction of the enteric nervous system , mainly Auerbach's myenteric plexus . The inflammatory process around the neurons leads to degenerative phenomena , thereby reducing nervous cell numbers and leading to the development of megacolon and megaesophagus [64–66] . Cardiac and cardiodigestive patients showed higher TLR2 mRNA expression than indeterminate and digestive patients . On the other hand , patients with different clinical manifestations of Chagas disease showed similar mRNA expression levels of TLR1 , TLR3 , TLR4 , TLR5 , TLR6 , TLR7 and TLR9 . Literature data has demonstrated that T . cruzi-infected individuals who have indeterminate clinical form of Chagas disease are heterozygous for the MAL/TIRAP S180L variant that leads to a decrease in signal transduction upon ligation of TLR2 or TLR4 , probably leading to reduced inflammatory response in the heart [31 , 67] . Thus , low TLR2 and TLR4 signaling have been associated with a lower risk of developing CCC . TLR2 and TLR4 activations in dendritic cells and macrophages conducing Myd88 and TRIF signaling activate NF-kB and lead to the production of pro-inflammatory cytokines such as IL-6 , IL12 and TNF-α [57] . CCC development has been correlated to immunological imbalance involving high IFN-γ and TNF-α production associated with low IL-10 and IL-17 secretion [11 , 16–18 , 49 , 68] . In our study we also observed that cardiac patients showed higher mRNA expression of NLRP3 , ASC , CASPASE-1 , IL-1β , IL-12 and TNF-α than indeterminate patients . Furthermore , a negative correlation among TLR2 , NLRP3 , IL-1β and TNF-β mRNA expression with LVEF , and positive correlation of NLRP3 mRNA expression with CI was observed in chagasic patients . NLRP3 inflammasome and apoptosis-associated speck–like protein containing a caspase recruitment domain ( ASC ) activates caspase-1 in experimental T . cruzi infection , and induce the production of active IL-1β and IL-18 [39] . Pro-inflammatory cytokines ( IL-1β , IL-6 and TNF-β ) regulate cell death of inflammatory tissues , modify vascular endothelial permeability , recruit blood cells to inflamed tissues , and induce the production of acute-phase proteins [69] . NLRP3 inflammasome is activated by prokaryotic RNA and different agents that trigger damage-associated molecular patterns ( DAMPs ) such as UVB irradiation [70] pore-forming toxins [71] , urate crystals and silica [72] . Moreover , several host-derived molecules indicative of damage activate the NLRP3 inflammasome , including reactive oxygen species [73] , extracellular ATP adenosine [74] , uric acid [75 , 76] and hyaluronan [77] which are released by injured cells . Thus , the NLRP3 activation mechanism observed during experimental T . cruzi infection [39] by an possible unknown parasite ligand may act together with DAMPs generated by injured cells [69] , thereby participating in the immunophysiological mechanisms involved in CCC development in chronic chagasic patients . Pro-inflammatory cytokines such as TNF-α , TNF-β , IL-1α , IL1β , IL-6 , IFN-α , IFN-γ and IL-8 induce acute phase protein productions which can opsonize parasites , activate complements , recruit immune cells and induce enzymes which degrade the extra cellular matrix . Acute phase proteins such as C-reactive protein , Serum amyloid A , Serum amyloid P component , Complement factors , Mannan-binding lectin , Fibrinogen , prothrombin , Plasminogen , Alpha 2-macroglobulin , Ferritin , Ceruloplasmin , Haptoglobin , Alpha 1-antitrypsin and α1-antichymotrypsin enhance the inflammatory process [78] . Increased levels of acute phase proteins are associated with increased risk for cardiovascular events in healthy individuals and coronary heart disease patients [79] . C-reactive protein levels are non-specific markers of systemic inflammatory processes , which reflect a vascular inflammation state and they are associated with cardiovascular damage [80] . High C-reactive protein levels have been described in non-chagasic cardiomyopathy [81] and also during CCC [82–85] . Thus , NLRP3 and TLR2 recognize parasite antigens and molecules associated with cell damage , leading to an inflammatory process with cytokines and other inflammatory mediator production that might participate of CCC development . CCC development is initiated by the presence of the parasite causing cardiomyocyte destruction due to parasite multiplication and inflammation [86] . However , CCC development depends on the parasite’s genetics [87–89] and the genetic background of the patients [67 , 90–92] which can induce different cardiac tropism patterns by the parasite and influence the immune response [87 , 93 , 94] . CCC involves persistent myocarditis , development of conduction disturbances , dysautonomia , cardiomegaly , fibrosis , ventricular wall thinning , microvascular damage , increased platelet activity , microthrombi , myocytolysis , myocardial fibrosis and death [4 , 95 , 96] . Persistent myocarditis is responsible for progressive neuronal damage , microcirculatory alterations , heart matrix deformations and consequent organ failure [97] . In this context the important inflammatory process in the myocardium which is responsible for CCC development seems to also be maintained by the innate immunity receptor activation such as TLR2 and NLRP3 , which induce the production of inflammatory cytokines ( IL-1β , IL-12 and TNF-α ) , thus amplifying the described inflammatory mechanisms and involving elements of adaptive immunity components such as T CD4 and CD8 lymphocytes and antibodies [57] . Our findings suggest that a high TLR2 and NLRP3 expression in chagasic cardiac patients may induce an enhancement of IL-1β , IL-12 , and TNF-α , thereby increasing cardiac inflammation and contributing to the heart dysfunction . One limitation of this study concerns on the widely clinical presentation of patients with CCC , according to the extent of myocardial damage and the relative small sample size . Untreated patients samples are very rare and difficult to obtain but are essential for the understanding of immunological mechanisms of Chagas disease pathophysiology . A better knowledge of the immune response involved in CCC development , the main factor correlated to death related to Chagas disease , may also contribute to reducing mortality and morbidity . The present study generated important data about the disease pathophysiology understanding , suggesting that distinct pattern recognition receptors may contribute differentially to the development of clinical forms of Chagas disease . | Chronic chagasic cardiomyopathy ( CCC ) is the main cause of death during Trypanosoma cruzi ( T . cruzi ) infection in patients . Individuals with CCC have high production of inflammatory mediators such as IFN-γ , TNF-α , IL-1β and nitric oxide ( NO ) which are involved with myocarditis , fibrosis , and myocardial hypertrophy . Yet the role of innate immunity receptors in CCC pathophysiology has not been addressed . Activation of TLRs and NLRs is fundamental to activate the innate immune system and also to modulate adaptive responses . Herein we have evaluated the association between innate immune receptors and innate cytokines with the clinical manifestation , and with the cardiac function in patients with different clinical forms of chronic Chagas disease . Our data suggest that high TLR2 and NLRP3 expression in cardiac patients may induce an enhancement of proinflammatory cytokine expression such as IL-1β , IL-12 , TNF-α and participate of the pathophysiology of CCC . A better understanding of immunological mechanisms involved in CCC may lead to reduced morbidity and mortality associated with the most lethal clinical manifestation of Chagas Disease . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"innate",
"immune",
"system",
"medicine",
"and",
"health",
"sciences",
"cardiomyopathies",
"immune",
"physiology",
"cytokines",
"pathology",
"and",
"laboratory",
"medicine",
"immunology",
"tropical",
"diseases",
"parasitic",
"diseases",
"parasitic",
"protozoans",
"develop... | 2018 | Innate immune receptors over expression correlate with chronic chagasic cardiomyopathy and digestive damage in patients |
A main challenge of modern biology is to understand how specific constellations of genes are activated to differentiate cells and give rise to distinct tissues . This study focuses on elucidating how gene expression is initiated in the notochord , an axial structure that provides support and patterning signals to embryos of humans and all other chordates . Although numerous notochord genes have been identified , the regulatory DNAs that orchestrate development and propel evolution of this structure by eliciting notochord gene expression remain mostly uncharted , and the information on their configuration and recurrence is still quite fragmentary . Here we used the simple chordate Ciona for a systematic analysis of notochord cis-regulatory modules ( CRMs ) , and investigated their composition , architectural constraints , predictive ability and evolutionary conservation . We found that most Ciona notochord CRMs relied upon variable combinations of binding sites for the transcription factors Brachyury and/or Foxa2 , which can act either synergistically or independently from one another . Notably , one of these CRMs contains a Brachyury binding site juxtaposed to an ( AC ) microsatellite , an unusual arrangement also found in Brachyury-bound regulatory regions in mouse . In contrast , different subsets of CRMs relied upon binding sites for transcription factors of widely diverse families . Surprisingly , we found that neither intra-genomic nor interspecific conservation of binding sites were reliably predictive hallmarks of notochord CRMs . We propose that rather than obeying a rigid sequence-based cis-regulatory code , most notochord CRMs are rather unique . Yet , this study uncovered essential elements recurrently used by divergent chordates as basic building blocks for notochord CRMs .
Cis-regulatory modules ( CRMs ) , or enhancers , are genomic DNA regions that dictate location , timing and rate at which one or more genes are expressed [1] . These regions have variable length and contain a flexible number of binding sites for transcription factors that function as either activators or repressors [2] . Point mutations in one or more of the functional binding sites within a CRM can alter its spatial and temporal properties , or cause its partial or complete inactivation . Recent estimates suggest that the human genome contains hundreds of thousands of CRMs that are believed to be mainly responsible for the developmental and functional complexity of different cells , tissues , and organs [3] . Notably , mutations and deletions of human enhancers have been associated with developmental defects , disease , and cancer [4–6] . However , in the human genome , as well as in several others , CRMs can be located up to thousands of kilobases away from the genes that they control and are brought closer to their target promoters after being bound by specialized proteins that bend the DNA [7] . Furthermore , CRMs can be located within introns and/or other untranslated regions [8] , or can be grouped into synergistically acting clusters called super-enhancers [9] . The crucial roles of CRMs , their complexity and their elusive nature , render a cis-regulatory code a highly desirable tool that would greatly simplify the genome-wide identification of CRMs with related properties . Studies aimed at identifying tissue-specific cis-regulatory codes have focused on genome-wide searches of clusters of known transcription factor binding sites [10] and on interspecific conservation of clusters of binding sites and/or larger non-coding sequences [11] . Nevertheless , recent research suggests that conserved clusters of binding sites are often non-functional [12] and that even evolutionarily ultraconserved genomic regions do not necessarily possess cis-regulatory activity [13] . The aim of the present study was to determine the structure and the functional binding sites of CRMs that shared comparable cis-regulatory activity and were presumably co-regulated , and to look for elements that could define a tissue-specific cis-regulatory code . We centered our analysis on CRMs active in the notochord , the most distinctive of chordate synapomorphies [14 , 15] . In all chordates , the notochord is the main source of support for the developing embryo and an essential patterning center for many of its structures and organs [16] . In vertebrates , the notochord is replaced by the vertebral column and its remnants form the nuclei pulposi of the intervertebral discs [17] . For the present study we used as a model system the tunicate Ciona , an invertebrate chordate that couples a compact , fully annotated genome with ease of transgenesis and tractable notochord [18 , 19] . According to phylogenomics data , tunicates are the invertebrate chordates most closely related to vertebrates [20] , and thus provide an opportunity to reconstruct the genetic circuitry and the evolutionary origins of the notochord through the identification of cis-regulatory sequences that enable gene expression in this structure [21–23] . We began this analysis with the characterization of fourteen notochord CRMs from Ciona . After isolating the minimal sequences necessary for their function , we tested whether these minimal sequences could be used to predict related notochord CRMs . We also evaluated the evolutionary conservation of CRM sequences between two Ciona species , C . intestinalis and C . savignyi , and compared the structure of the Ciona notochord CRMs to fully characterized notochord CRMs from other chordates , including mouse and zebrafish . Rather than a sensu stricto cis-regulatory code , this study elucidated various combinations of functional transcription factor binding sites that function in a context-dependent fashion . These binding sites are often poorly conserved interspecifically , and therefore would have been missed by conservation-based methods of enhancer detection . However , despite the intraspecific and interspecific variability in their composition and function , binding sites for Brachyury and Foxa2 emerged as recurrent hallmarks of notochord CRMs from highly divergent chordates .
We identified fourteen CRMs that can induce gene expression in the Ciona notochord . To avoid sequence and/or positional biases , all but one of the notochord CRMs ( Fig 1 ) were isolated through testing of random genomic regions ( S1 Table ) . Minimal notochord enhancers spanning 80–547 bp were subsequently identified through sequence-unbiased truncation analyses , involving in vivo testing of ~200 constructs ( S1 , S2 and S3 Figs ) . Lastly , we assessed the effects of site-directed mutations targeting either known putative transcription factor ( TF ) binding sites or uncharacterized sequences . The results of these studies are condensed in Fig 1 . We found that the majority of the CRMs ( 9/14 , 64 . 3% ) require binding sites for the TFs Ciona Brachyury ( Ci-Bra ) and/or Ci-FoxA-a ( Foxa2/fkh/HNF3beta ortholog; hereinafter Ci-Fox ) ; in contrast , binding sites for TFs of widely different families were responsible for the function of the remaining five notochord CRMs . This analysis also revealed unexpected characteristics of these regulatory elements . For instance , enrichment for a particular binding site was not a reliable predictor of either functionality or cooperativity ( e . g . , all Ci-Fox sites in Ci-CRM70 are dispensable; Figs 1 and S1 ) . In some instances , only one of the multiple copies/variants of a given TF binding site was required for notochord gene expression ( e . g . , only one of the seven Ci-Bra sites in Ci-CRM99 is necessary; Figs 1 and S3 ) . Furthermore , even CRMs necessitating the same types of binding sites could function differently: a Myb-like site worked individually in one CRM ( Ci-C6ST-like7 ) , and in combination with a related Myb-like site in another ( Ci-CRM76 ) ( Figs 1 and S1 ) . We had previously described a notochord CRM , associated with the gene Ci-tune , activated by synergistic Ci-Bra and Ci-Fox binding sites [24] . In this study , we found that Ci-CRM96 relies on the same type of synergism ( Fig 2A ) , and although the sequences of the Ci-Bra and Ci-Fox sites differ between these two CRMs , their spacing is comparable ( 48 bp in Ci-CRM96 , 46 bp in Ci-tune ) . In contrast , the multiple Ci-Bra and Ci-Fox sites in Ci-CRM24 act redundantly , as individual mutations ( e . g . , Fox1 and Bra4 , Fig 2F ) are not detrimental to notochord staining ( Fig 2F–2I ) , and reduction/loss of notochord staining is only obtained through compound mutations ( Fig 2F , 2J , 2K and 2L ) . Unlike the previous CRMs , Ci-CRM112 is devoid of Ci-Bra sites ( Fig 2M ) . In this case , putative homeodomain ( HD ) and activator protein 1 ( AP1 ) sites appear to work cooperatively with a Ci-Fox site , since all single mutations decrease notochord staining ( Fig 2M–2Q ) , and simultaneous mutations of the functional Ci-Fox site and either the HD or AP1 sequences result in loss of staining ( Figs 2M , 2R , 2S and S2 ) . Six CRMs rely on individual Ci-Bra binding sites ( Figs 1 , S1 and S3 ) . Counterintuitively , the sequences of indispensable Ci-Bra sites differ for each Ci-Bra-dependent CRM , and sites with identical core sequences may be necessary in one context , but not in another ( e . g . , the TTGCAC sites in Ci-CRM109 and Ci-Fkbp9; S1 and S3 Figs ) . To uncover the molecular foundations of such differences , we assessed the roles of sequences directly adjacent to the necessary Ci-Bra binding sites . For Ci-CRM66 , which lies within an intron of Ci-Ephrin3 , we found that mutation of a single Ci-Bra binding site drastically decreased , but did not abolish , notochord staining ( Figs 3A , 3E , 3J and S3 ) . Linker-scanning mutagenesis revealed that the most detrimental mutations were those affecting an ( AC ) 6 microsatellite [25] directly abutting the TCACAC Ci-Bra site ( Fig 3B ) . Mutation of the first two ( AC ) pairs ( Fig 3C ) caused a sharp drop in notochord expression ( Fig 3H and 3J ) , as did a mutation that caused a “frame-shift” of the microsatellite sequence ( Fig 3B and 3F ) , suggesting that uninterrupted periodicity between the Ci-Bra binding site and this sequence may be required for the function of this CRM . The number of intact repeats also influenced activity ( Fig 3B ) , and the mutation of the entire microsatellite abolished notochord expression ( Fig 3C , 3I and 3J ) . Notably , ChIP-chip studies of genomic targets of Brachyury in differentiating mouse embryonic stem cells showed that this TF often binds ( AC ) repeats [26] . The Ciona intestinalis genome contains only nine copies of an ( AC ) ≥6 microsatellite abutting a TCACAC Ci-Bra binding site; however , despite their reported occupancy by Ci-Bra in early embryos [27] , none of the remaining eight regions directed notochord gene expression ( S2 Table ) . We also searched the sequences of the remaining five CRMs that rely on single Ci-Bra binding sites for clues on the mechanisms that might create the appropriate context for their function . Even though mouse Brachyury was initially found to bind the palindromic sequence T ( G/C ) ACACCTAGGTGTGA [28] , it was later shown that TNNCAC core half-sites are efficiently bound by Brachyury proteins from mouse and other organisms , including Ciona [29–32] . Our results confirm that a palindromic organization is not required; instead , we observed that 50% of the required Ci-Bra sites matched either the TNNCACCTAM or the CTAMGTGNNA consensus ( core sites underlined ) ( Fig 3K ) . Consequently , we selectively mutated the adjacent nucleotides while leaving the TNNCAC cores intact and found that in the case of Ci-CRM109 and Ci-CRM99 disruption of the CTAM sequence had the same effect as the mutation of the cores ( Fig 3L–3S ) . Similar results were obtained through the mutation of this stretch in the Ci-ABCC10 CRM [33] . In contrast , mutation of the CTAM sequence within Ci-CRM86 left notochord staining unaffected ( Fig 3T–3W ) and a CTAM-containing Ci-Bra binding site within Ci-CRM9 was found to be dispensable ( S3 Fig ) . We conclude that the CTAM extension is not entirely predictive of whether a CRM will necessitate a single Ci-Bra site , and the binding sites that possess it are not always necessary . It is also conceivable that a fraction of the binding sites that we tentatively attributed to Ci-Bra might be interchangeably or exclusively utilized by Ci-Tbx2/3 , the only other T-box protein present in the Ciona notochord , which acts as a mediator of Ci-Bra [34] . The sequences flanking the core TNNCAC site might therefore be required for binding specificity of either T-box factor , Ci-Bra or Ci-Tbx2/3 . In the last group of five minimal CRMs , the sequences required for notochord expression were neither Ci-Bra nor Ci-Fox binding sites ( Fig 1 ) , but instead resembled sites for bHLH ( Ci-CRM26 ) , Klf/Sp1 ( Ci-CRM90 ) , and Myb-like factors ( Ci-CRM70 , Ci-CRM76 and Ci-C6ST-like7 ) ( S1 Fig ) . These results are consistent with previous reports of notochord-expressed bHLH , Klf6 and Klf15 TFs [35–37] , and of a Myb-related gene in Ciona [38] . The requirement for two short Myb-like sites in Ci-CRM76 ( Fig 1 ) led us to hypothesize that its activity might require a specific architecture . Accordingly , we found that while reversing the orientation of one of the Myb-like sites ( abbreviated as “M” ) , M2-2 , had no effect , transposing the order of the two required Myb-like sites , M1-5 and M2-2 , largely decreased notochord staining ( S4 Fig ) . Furthermore , increasing the spacing between M1-5 and M2-2 ( 4 bp ) to that of the dispensable sites , M2-1 and M1-4 ( 8 bp ) , caused an even more substantial reduction of reporter gene expression in the notochord ( S4 Fig ) . Nevertheless , seven genomic regions containing Myb-like sites with the identical composition , orientation and spacing as Ci-CRM76 , all of which mapped near notochord genes , did not yield detectable notochord expression when tested in vivo ( S3 Table ) . Additional sequence inspection identified non-microsatellite repeats in various CRMs . Combinations of recurring motifs and/or evolutionarily conserved TF binding sites have guided the identification of CRMs active in the Ciona muscle [21 , 39–42] and central nervous system ( CNS ) [41 , 43] , as well as in various tissues/embryonic territories of Drosophila [10 , 44 , 45] and in the zebrafish notochord [46] . For these reasons , we sought to investigate whether these repeats could aid in the prediction of novel notochord CRMs in Ciona intestinalis . We noticed that Ci-CRM90 features two nearly identical 73-bp sequence blocks , each containing two copies of a smaller 20-bp repeat; moreover , a sequence motif related to the 20-bp repeat was found in Ci-CRM9 ( S4 Fig ) . Ci-CRM26 contains a 19-bp tandem repeat , whose first copy overlaps with the E-box required for activity . The exact sequences of both of these repeats are unique in the Ciona intestinalis genome; however , shorter variations of the Ci-CRM26 repeat are seen in four other notochord CRMs ( S4 Fig ) . To assess the predictive ability of functional binding sites and motifs , we tested 36 genomic fragments containing arrangements of binding sites and/or motifs identical or similar to those found in the Ci-CRMs ( Fig 1 ) . We only detected notochord expression in one construct ( S3 Table , S4 Table ) : the short motif found in Ci-CRM26 , which occurs ~3 , 017 times in the Ciona intestinalis genome , led us to the identification of a novel notochord CRM within the Ci-Noto2 locus ( S4 Fig and S4 Table ) . We also tested whether interspecific sequence homology could improve the prediction of notochord CRMs , since evolutionary conservation is widely used to pinpoint Ciona cis-regulatory regions ( e . g . , [47–49] ) . The CRMs presented here were isolated using a conservation-independent approach , but when we retrospectively assessed this parameter , we observed surprising interspecific variability among their sequences . Indeed , many of these Ciona intestinalis CRMs display limited conservation , if any , with Ciona savignyi ( S4 Fig ) . In addition , even though some binding sites , such as the Ci-Fox and E-box sites of Ci-CRM76 , are perfectly conserved between the two Ciona species , neither is required for activity ( S4 Fig ) ; this suggests that even interspecifically conserved notochord TF binding sites are not reliable indicators of functionality . These results concur with studies in Drosophila that suggest that clustered binding sites within CRMs might be retained over evolution for reasons other than selection or functional necessity [12] . In sum , the unexpected variety and flexibility of the mechanisms that we have described here limited our ability to predict notochord CRMs from sequence alone . Yet , although our results seem to question the existence of a straightforward notochord cis-regulatory code , this study uncovered recurring grammatical elements shared by notochord CRMs . In particular , Brachyury and Foxa2 binding sites emerge as the basic building blocks of most Ciona notochord CRMs ( Fig 4A ) , and these results are consistent with findings in other chordates . In fact , Brachyury binding sites have been found to be critical for the function of notochord in different animals ( e . g . [29 , 50] ) , and our previous studies in Ciona show that they can act either individually or cooperatively [33 , 34 , 53] . Their association with ( AC ) microsatellites in Ci-CRM66 and in the mouse genome [26] might represent a recurring feature of a distinct class of notochord CRMs ( Fig 4A ) . Foxa2 sites are required in notochord CRMs from zebrafish and mice [46 , 54] , although they are rarely sufficient to initiate expression when in single copy , and often necessitate additional sequences [46 , 58 , 61] whose identity appears to be lineage-specific ( Fig 4A and 4C ) . These observations and our previous results [33] reflect the reported pioneer chromatin-opening ability of Fox proteins [62] , which may not able to activate gene expression per se but are required to increase the accessibility of CRMs to other transcription factors , such as Brachyury and/or other notochord-specific activators . The basic cis-regulatory repertoire that we have uncovered was likely expanded via vertebrate-specific evolutionary events; such events include the notochord deployment of additional TFs , such as homeobox and Hox proteins and their co-factors , which are remarkably underrepresented in the tunicate notochord , [63] along with the duplication and consequent divergence of regulatory regions .
Adult Ciona intestinalis were purchased from Marine Research and Educational Products ( M-REP; Carlsbad , CA ) and kept in an aquarium in recirculating artificial sea water at 17–18°C . Culturing and electroporations were carried out as previously described [64] . After electroporation , transgenic embryos were fixed in 0 . 2% glutaraldehyde and stained at 37°C with 5-bromo-4-chloro-3-indolyl-β-D-galactopyranoside ( X-gal ) [64] . Stained embryos were washed in 500 μL PBST ( 1X PBS , 0 . 1% Tween 20 ) , post-fixed in 300–500 μL of 4% paraformaldehyde in PBST , and stored at 4°C . To determine the comparative activities of wild-type and mutated constructs , the proportions of X-gal stained embryos exhibiting notochord staining were determined from at least three independent experiments . Data presented in graphs represent average values , with error bars denoting the standard deviation . Genomic fragments for enhancer discovery and analyses were cloned into the pFBΔSP6 plasmid , which contains the LacZ reporter gene [64] . After the initial characterization of each notochord CRM , subsequent deletions and mutations were made either by utilizing unique restriction enzyme sites or by Polymerase Chain Reaction ( PCR ) , using the smallest active DNA fragment as a template . A list of the oligonucleotides employed for PCR amplifications and the restriction sites used for cloning the most relevant constructs is provided in S5 Table . For the predictions of notochord CRMs , suitable genomic regions were first identified by searching either the Ciona genome or a database of validated Ciona notochord genes for transcription factor binding sites , motifs or other sequence signatures present in notochord CRMs , using the GUFEE program [24] . Our database of Ciona notochord genes contained the sequences of the putative genomic loci of 300 notochord genes . We manually annotated the gene models from expression data present in the ANISEED database [38] and from our results . The sequences included in the database were extracted from the UCSC genome browser ( Ciona intestinalis version 1 ) by Dr . John R . Edwards ( Washington University , St . Louis ) . | Transcription factors control the spatial and temporal expression of a multitude of genes by binding their cis-regulatory modules ( CRMs ) . In this study , we investigated the architecture and composition of CRMs that direct gene expression in the notochord , a structure necessary for the support and patterning of the embryonic body plan of all chordates . We used the simple chordate Ciona to carry out a comparative study of notochord CRMs and we identified the sequences necessary for their function . These sequences , in turn , highlighted the existence of multiple mechanisms that enable gene expression in the notochord . Surprisingly , combinations of binding sites identical to those found in active CRMs were not necessarily able to direct notochord gene expression and were often poorly conserved among cogener species . These results challenge the concept of a notochord-specific cis-regulatory “code” , and outline the limitations of methods for CRM identification that rely upon interspecific conservation of non-coding sequences . Nevertheless , a broad comparison of the structure of the Ciona CRMs with that of the notochord CRMs characterized thus far from all chordates outlines the existence of essential evolutionarily conserved building blocks , such as binding sites for the transcription factors Brachyury and Foxa2 , that are shared by subsets of these regulatory modules . | [
"Abstract",
"Introduction",
"Results",
"and",
"Discussion",
"Materials",
"and",
"Methods"
] | [] | 2015 | Brachyury, Foxa2 and the cis-Regulatory Origins of the Notochord |
The fecal egg count reduction test ( FECRT ) is recommended to monitor drug efficacy against soil-transmitted helminths ( STHs ) in public health . However , the impact of factors inherent to study design ( sample size and detection limit of the fecal egg count ( FEC ) method ) and host-parasite interactions ( mean baseline FEC and aggregation of FEC across host population ) on the reliability of FECRT is poorly understood . A simulation study was performed in which FECRT was assessed under varying conditions of the aforementioned factors . Classification trees were built to explore critical values for these factors required to obtain conclusive FECRT results . The outcome of this analysis was subsequently validated on five efficacy trials across Africa , Asia , and Latin America . Unsatisfactory ( <85 . 0% ) sensitivity and specificity results to detect reduced efficacy were found if sample sizes were small ( <10 ) or if sample sizes were moderate ( 10–49 ) combined with highly aggregated FEC ( k<0 . 25 ) . FECRT remained inconclusive under any evaluated condition for drug efficacies ranging from 87 . 5% to 92 . 5% for a reduced-efficacy-threshold of 90% and from 92 . 5% to 97 . 5% for a threshold of 95% . The most discriminatory study design required 200 subjects independent of STH status ( including subjects who are not excreting eggs ) . For this sample size , the detection limit of the FEC method and the level of aggregation of the FEC did not affect the interpretation of the FECRT . Only for a threshold of 90% , mean baseline FEC <150 eggs per gram of stool led to a reduced discriminatory power . This study confirms that the interpretation of FECRT is affected by a complex interplay of factors inherent to both study design and host-parasite interactions . The results also highlight that revision of the current World Health Organization guidelines to monitor drug efficacy is indicated . We , therefore , propose novel guidelines to support future monitoring programs .
Infections with the soil-transmitted helminths ( STHs ) , namely Ascaris lumbricoides , Trichuris trichiura and hookworm ( Necator americanus and Ancylostoma duodenale ) are among the most common infectious diseases in children of tropical countries causing malnutrition , growth stunting , intellectual retardation , and cognitive deficits [1] . Currently , the large-scale administration of benzimidazole drugs ( i . e . , albendazole and mebendazole ) is the most widely used method to control morbidity due to STH infections , and a scale-up of these large-scale treatment programs is underway in Africa , Asia , and Latin America ( donation of 400 million tablets of albendazole by GlaxoSmithKline and 200 million tablets of mebendazole by Johnson & Johnson ) . Rather than aiming to achieve eradication , these control programs are focused on reducing infection intensity and transmission potential , and hence reduce morbidity [2] . However , due to the scarcity of alternative anthelmintics , it is imperative that monitoring systems are designed to detect any change in drug efficacy due to emerging resistance of the parasites against benzimidazoles [3]–[7] . At present , the fecal egg count reduction test ( FECRT ) is recommended to monitor anthelmintic efficacy against STH in animal [8] and public health [9] . Guidelines on how to conduct a FECRT in public health were published by the World Health Organization ( WHO ) in the late 1990s [10] , providing recommendations on sample size ( ∼200 infected subjects ) , stool sampling ( two stool samples of two different days both before and after administration of drugs ) , the detection limit of the method to quantify the number of eggs ( Kato-Katz thick smear with a detection limit of 24 eggs per gram of stool ( EPG ) ) and thresholds defining reduced efficacy ( FECRT <70% for A . lumbricoides and FECRT <50% for T . trichiura and hookworm ) . However , the current guidelines have some important weaknesses . At first , the level of understanding of the effects of the factors inherent both to study design ( sample size , stool sampling and the fecal egg count ( FEC ) method ) and host-parasite interactions ( level of egg excretion and level aggregation of STH infections across host populations ) to support these guidelines is poor . In veterinary sciences , there is empirical evidence that low FEC may thwart interpretation of FECRT results , particularly when sample size is small and/or detection limit of the FEC method is low [11] , [12] . As a consequence of this , it is most likely that performing a FECRT across the three STHs , would require a different study design , and this solely due to the differences in fecundity ( A . lumbricoides≫hookworm>T . trichiura [1] ) . Another important issue of the current guidelines is the additional technical and financial resources that are required to monitor anthelmintic efficacy . Based on the cost assessment of the Kato-Katz thick smear for STH diagnosis by Speich and colleagues in epidemiological surveys in an African setting [13] , it can be deduced that the re-examination already would require US$ 3 . 46 per subject . Therefore , any effort to reduce the cost and the complexity of a surveillance system is desirable . Finally , recent efficacy trials performed in seven countries across Africa , Asia , and Latin America questioned the validity of the thresholds for reduced efficacy [9] , as a single dose of albendazole revealed to be highly efficacious against both A . lumbricoides ( FECRT >99% ) and hookworm ( FECRT >90% ) . As a consequence of this , it was proposed to adopt the current thresholds of reduced efficacy to <95% and <90% for A . lumbricoides and hookworm , respectively . The aim of the present study was to assess the impact of sample size , detection limit of the FEC method , level of egg excretion , and aggregation of FEC on the interpretation of the FECRT . To this end , data were generated using a statistical simulation and analyzed using tree based-models . The outcome of these trees was subsequently validated on five efficacy trials previously conducted in Africa , Asia , and Latin America . From the results , we propose cost-effective study designs to successfully monitor anthelmintic drug efficacy in future anthelmintic treatment programs .
Data were generated by Monte Carlo simulation as previously described by Dobson et al . ( 2009 ) [14] and which was extended by varying sample size , detection limit of the FEC method , level of eggs excreted , and level of aggregation of eggs across the hosts . To fully understand this simulation , the various steps will be explained in more detail . First , the distribution of parasites within the host population before administration of drugs was defined by a negative binomial distribution . This distribution is determined by two parameters: the mean level of egg excretion across subjects ( mean pre-drug administration ( pre-DA ) FEC ) and the level of aggregation of FEC across subjects ( k ) . Low values of k indicate that only few subjects are excreting the majority of eggs , where high values indicate that egg counts are more normally distributed across the host population . From this pre-defined distribution , a number of individual subjects were randomly drawn representing the sample size . An example is given in Table 1 , where the outcome of such a random sample is shown for a mean pre-DA FEC = 250 EPG , k = 1 and sample size = 6 . The pre-DA FEC observed , however , will be different from the ‘true’ pre-DA FEC due to the variation ( i . e . stochasticity ) introduced by sampling eggs associated with the FEC method . This component of variation was simulated using a Poisson distribution defined by the expected number of eggs counted ( = ‘true’ host FEC/detection limit ) . In Table 1 , the expected number of eggs to be counted when using a FEC method with a detection limit of 24 EPG ( in casu the standard Kato-Katz thick smear ) for subject A with a ‘true’ subject FEC of 796 EPG equaled 33 . 2 eggs ( 796/24 ) . A random sample was then drawn from this pre-defined Poisson distribution , and for this sample 35 eggs were observed , which was multiplied by 24 ( detection limit ) to obtain an observed pre-DA FEC of 840 EPG . This procedure was repeated for each of the six subjects . In order to simulate a TDE of 50% , the ‘true’ pre-DA FECs were multiplied by 0 . 5 ( 1-TDE ) . The observed FEC after the administration of the drug ( post-DA FEC ) was generated as described above for the pre-DA FEC . Subsequently , the FECRT was calculated as described in the formula below , resulting in an observed reduction of 58 . 6% for the example provided in Table 1 . It is important to note that only one sample is examined per subject and that all subjects are included in the calculation of the FECRT , even those for whom the observed pre-DA FEC equaled zero . Finally , the entire process was iterated 500 times , to obtain 500 estimates of FECRT for this pre-defined parasite population , sample size , detection limit , and TDE . The parasite-host population parameter values chosen for mean pre-DA FEC ( 50 , 100 , 150 , 200 , 250 , 500 , 750 , and 1000 EPG ) and k ( 0 . 01 , 0 . 025 , 0 . 05 , 0 . 075 , 0 . 1 , 0 . 25 , 0 . 5 , 0 . 75 , 1 , 1 . 5 , and 2 ) were based on previously conducted studies where STH were quantified [9] , [15] , [16] . The values for the sample size were 6 , 10 , 15 , 20 , 25 , 50 , 75 , 100 , 125 , 150 , 175 , and 200 , covering a large range of applied sample sizes to determine drug efficacy against STH [16] . The values for the detection limit represented those of four currently used FEC methods both in human and veterinary parasitology: FLOTAC ( detection limit = 1 and 2 EPG ) [17] , FECPAK ( detection limit = 5 and 10 EPG ) ( http://www . fecpak . com ) , Kato-Katz thick smear ( detection limit = 12 and 24 EPG ) [18] and McMaster ( detection limit = 25 , 33 . 3 , and 50 EPG ) [19] . The FEC methods used in veterinary medicine were included in this analysis because they have recently been validated for the diagnosis of STH in public health ( FLOTAC [20] and McMaster [21] ) . In addition , the inclusion of each of these additional assays allowed assessing the impact of detection limit in greater depth . The TDE was set on 50 , 60 , 70 , 80 , 82 . 5 , 85 , 87 . 5 , 90 , 92 . 5 , 95 , 97 . 5 , and 99% , resulting in 114 , 048 combinations ( 8 ( mean pre-DA FEC ) ×11 ( k ) , 9 ( detection limit ) ×12 ( sample size ) ×12 ( TDE ) ) that were each iterated 500 times . The impact of the various factors on the sensitivity and specificity of the FECRT was evaluated . Every TDE that was less than 90 or 95% was considered as a truly reduced efficacy and as truly efficacious if different . Both thresholds have been recently suggested for hookworm and A . lumbricoides , respectively [9] . The current threshold for T . trichiura ( below 50% ) was not included , because its remains to be elucidated [22] . A combination of evaluated factors ( 500 iterations ) was considered to be “sensitive” ( i . e . true test positive ) when a FECRT could be calculated ( observed mean pre-DA FEC >0 ) and a truly reduced efficacy ( TDE <90% or <95% ) was correctly detected in at least 95% of the iterations or “insensitive” ( i . e . , false negative ) otherwise . A combination of evaluated factors was considered to be “specific” ( true test negative ) when a FECRT could be calculated ( observed mean pre-DA FEC >0 ) and TDE ≥90 or ≥95% was correctly detected in at least 95% of the iterations or “non-specific” ( false positive ) otherwise . In the example provided in Table 1 , more than 95% of the 500 iterations yielded a FECRT below the defined thresholds , therefore , the FECRT for the combination of a mean pre-DA FEC = 250 EPG , k = 1 , detection limit = 24 EPG , and a sample size = 6 was considered ‘sensitive’ to detect the reduced efficacy of 50% . For this combination , the specificity ( correctly determine susceptibility when the STH are drug-susceptible ) cannot be evaluated as the TDE was below the thresholds for reduced efficacy of both 90 and 95% . Subsequently , tree-based models ( classification trees ) were built in R using the packages ‘rpart’ and ‘randomforest” ( version 2 . 10 . 0 , 2009 , The R Foundation for Statistical Computing ) with both sensitivity and specificity as a binary outcome variable ( outcome values are either 0 or 1 ) and the parasite-host population ( mean pre-DA FEC and k ) , the sample size , the detection limit , and the TDE as continuous predictor variables [23] . The sample sizes across the different trials were predicted by the classification trees ( predicted sample size ) and subsequently compared with those estimated by exact inference on the raw data of five previously conducted efficacy trials ( required sample size ) . These trials evaluated the efficacy of a single dose albendazole ( 400 mg ) against A . lumbricoides ( four out of five trials ) , T . trichiura ( three out of five trials ) and hookworm infections ( all five trials ) in school children in three countries in Africa ( Cameroon , Ethiopia and Tanzania ) , one country in Asia ( Cambodia ) and one Latin American country ( Brazil ) [9] . These trails were selected for two reasons . First , they were standardized in terms of the follow-up ( between 14 and 30 days after the administration of drug ) , the detection technique ( the McMaster egg counting method , detection limit = 50 EPG ) and statistical analysis ( see formula above ) . Second , the prevalence of STH before the drug administration exceeded 20% in each of these trials , and hence meeting the criteria to implement preventive chemotherapy programs [24] . For this validation all subjects screened at baseline were included ( subjects might be falsely classified as non-infected due to the lack of sensitivity of the McMaster FEC method ) . However , subjects with a baseline FEC of 0 EPG were not treated nor re-examined at follow-up . To include these subjects it was assumed that the FEC at follow-up of these non-infected subjects ( falsely/truly ) also equaled zero after drug administration . In addition to this , a number of infected subjects did not provide a stool sample at follow-up . These subjects were replaced by a random sample of subjects for which complete data were available . The sample size , prevalence , mean pre-DA FEC , the aggregation of the FEC ( k = ( arithmetic mean FEC ) 2/ ( variance FEC - arithmetic mean FEC ) ) and the FECRT observed in these trials are summarized in Table 2 . For the validation of the statistical methods , these values observed for FECRT , mean pre-DA FEC and k are considered to be ‘true’ values . The overall protocol of this multi-country study was approved by the ethics committee of the Faculty of Medicine , Ghent University ( no . B67020084254 ) and was followed by a separate local ethical approval for each study site . For Brazil , approval was obtained from the institutional review board ( IBR ) from Centro de Pesquisas René Rachou ( no . 21/2008 ) , for Cambodia from the national ethic commitee for health research , for Cameroon from the national ethics committee ( no . 072/CNE/DNM08 ) , for Ethiopia from the ethical review board of Jimma University , for India from the IBR of the Christian Medical College ( no . 6541 ) , for Tanzania ( no . 20 ) from the Zanzibar Health Research Council and the Ministry of Health and Social Welfare , for Vietnam by the Ministry of Health of Vietnam . An informed consent form was signed by the parents of all subjects included in the trials . This clinical trial is registered under the ClinicalTrials . gov , identifier NCT01087099 . The predicted sample sizes were deduced from the results of the classification trees ( Figures 1 , 2 , S2 and S3 ) and are shown in Figure 3 . For example , the predicted sample size to correctly diagnosis a reduced efficacy against A . lumbricoides in the Brazilian trial ( FECRT = 100% , mean pre-DA FEC = 1 , 353 EPG and k = 0 . 063 ) ranged from 50 to 200 for both thresholds defining reduced efficacy . When none of the combinations resulted in a reliable diagnosis , the predicted sample size was set at >200 , as this was the largest sample size examined in the classification trees . This was for example the case for the efficacy against T . trichiura ( FECRT = 92 . 4% , mean pre-DA FEC = 110 EPG and k = 0 . 065 ) in the Ethiopian trial , and this for both thresholds . The required sample size based on the raw data of the different trials was estimated by bootstrap analysis ( re-sampling with replacement and 10 , 000 iterations ) , as at present no formulae are available to calculate sample size for the correct diagnosis of reduced efficacy . In this analysis , different sample sizes were analyzed in order to determine the smallest sample size for which FECRT could be calculated ( mean pre-DA FEC >0 ) and a truly reduced efficacy ( TDE <90% or <95% ) was correctly detected in at least 95% of the iterations or when a TDE ≥90 or ≥95% was correctly detected in at least 95% of the iterations . The outcome of the bootstrap analysis for Brazilian trial against A . lumbricoides described above is illustrated in Figure S1 . As the sample size increase , the probability of correctly detecting a reduced efficacy increased . The required sample size based on this trial was 17 for both thresholds . In 24 cases the agreement between the required and the predicted sample size was assessed ( two thresholds ( 90 and 95% ) ×12 FECRT ( four for A . lumbricoides , three for T . trichiura and five for hookworm ) . There was an agreement between the required and the predicted sample size , if the exact required sample size fell within the predicted sample size interval . For cases where the required sample size did not fall within the predicted sample size interval , it was assessed whether the required sample size was overestimated ( required sample size < lower limit of the predicted sample size interval ) or underestimated ( required sample size > upper limit of the predicted sample size ) .
The classification-trees for the specificity to detect efficacy ≥90% and sensitivity to detect reduced efficacy <90% are provided in Figure 1 and 2 , respectively . The terminal nodes are colored green when the specificity/sensitivity was reliable ( ≥85% ) , and red if different . The specificity was evaluated in the 47 , 520 combinations were the TDE was ≥90% and was affected with decreasing importance ( increasing number of bifurcations from the root node ) by TDE , sample size and aggregation of the FEC ( k ) . The detection limit and mean pre-DA FEC did not considerably influence the specificity , since these parameters did not result in any bifurcation across the classification tree . From the red-green color code to define a reliable specificity , it can be deducted that false positive conclusions concerning reduced efficacy were drawn when the TDE was between 90 and 92 . 5% ( specificity = 0% , n = 9 , 504 ) . For a TDE ≥92 . 5% , reliable specificity results depended on the sample size . For small sample sizes ( <50 subjects ) , reliable conclusions could only be drawn when lowly aggregated FEC ( k≥0 . 05 ) were combined with TDE ≥95% and a sample size of at least 10 subjects ( specificity = 89 . 8% , n = 7 , 776 ) . For large sample sizes ( ≥50 ) , specificity was always high , regardless of the aggregation of the FEC and TDE ( specificity = 96 . 0% , n = 22 , 176 ) . The sensitivity was evaluated in the remaining 66 , 528 combinations where the TDE did not exceed 90% . The most important factor affecting the sensitivity was the sample size , followed by both TDE and aggregation of the FEC and finally the mean pre-DA FEC . The detection limit did not considerably influence the sensitivity . For sample sizes <50 , reduced efficacies were only correctly diagnosed when lowly aggregated FEC ( k≥0 . 075 ) were combined with a TDE <82 . 5% and a sample size of ≥10 ( sensitivity = 90 . 3% , n = 11 , 520 ) . For sample sizes ≥50 , the diagnosis of reduced efficacy depended on the TDE . For TDE between 85 . 0% and 87 . 5% , satisfactory sensitive results were only found when mean pre-DA FEC were high ( ≥150 EPG ) ( sensitivity = 88 . 3% , n = 4 , 158 ) . For TDE <85 . 0% , sensitivity was high ( = 98 . 1 , n = 33 , 264 ) , regardless of the mean pre-DA FEC . The combinations that result in a reliable detection of a normal or reduced efficacy ( sensitivity and specificity >85% ) and their respective TDE limits for which the FECRT cannot reliably provide a correct diagnosis ( ‘grey’ zone ) are summarized in Figure 3 . All combinations resulted in a reliable classification of efficacy status , except for small sample sizes ( <10 ) and moderate sample sizes ( 10–49 ) combined with highly aggregated FEC ( k <0 . 075 ) . The TDE limits for sensitivity ( TDE = 82 . 5% ) and specificity ( TDE = 95% ) that were least discriminatory occurred for moderate sample sizes ( 10–49 ) combined with low aggregated FEC ( k≥0 . 075 ) . Best discrimination TDE limits for sensitivity and specificity were 87 . 5% and 92 . 5% , respectively , and occurred for large sample sizes ( 50–200 ) combined with high mean pre-DA FEC ( ≥150 EPG ) . For the reduced efficacy threshold of 95% , only the combinations , which result in a reliable classification of efficacy status and their TDE limits in which FECRT results are unreliable , are reported ( Figure 3 ) . The classification trees of the specificity and sensitivity are provided in Figures S2 and S3 , respectively . Compared to a reduced efficacy defined as TDE <90% , there were three important differences . First , the pre-DA FEC did not affect the diagnosis of reduced efficacy . Second , the detection limit had a considerable impact on the interpretation of FECRT . A detection limit ≥15 did not always allow a reliable FECRT , particularly for sample sizes <50 subjects . Finally , there was a difference in the critical value ( s ) for sample size ( 10 and 50 for 90% threshold vs . 10 , 25 , and 50 for the 95% threshold ) and aggregation of the FEC ( 0 . 075 vs . 0 . 05 and 0 . 25 ) . The least discriminatory TDE limits for which conclusions were doubtful was found when moderate sample sizes ( 10–24 ) were combined with a high detection limit ( <15 EPG ) and lowly aggregated FEC ≥0 . 25 . For these combinations TDE limits were 87 . 5% to 97 . 5% for sensitivity and specificity , respectively . Best discrimination ( TDE 92 . 5% and 97 . 5% for sensitivity and specificity respectively ) was observed for moderate sample sizes ( 25–49 ) combined with high detection limits ( <15 EPG ) and for large sample sizes ( ≥50 subjects ) regardless of the detection limit . The predicted and the required sample sizes across the different trials for both a normal and a reduced efficacy <90 and <95% are provided in Table 2 . Overall , there was an agreement between the predicted and the required sample size in eight out of 24 cases ( highlighted in bold ) . In the 16 remaining cases , the required sample size fell out of the interval of predicted sample size . Yet , the required sample was only underestimated in three cases ( highlighted in italic ) . In the remaining 13 cases , the required sample size was overestimated ( not highlighted ) . There was a slight variation in agreement between the required and predicted sample size across the two thresholds defining reduced efficacy . For a threshold of 90% , there was an agreement between the required and the predicted sample size in five cases , whereas this was only observed in three cases for a threshold of 95% . Moreover , two out of the three cases for which the required sample size was underestimated were found for the latter threshold .
In the present study , the most applied test to evaluate anthelmintic drug efficacy against A . lumbricoides and hookworm in public health , was virtually performed under varying conditions of sample size , detection limit of the FEC method , level of excretion , and aggregation of eggs within the host population . Subsequently , tree-based models were built to assess the impact of these factors on the specificity and the sensitivity to detect normal or reduced efficacy . Finally , the outcomes of these models were validated on different efficacy trials done in Africa , Asia and Latin America . The present study provides novel insights into three aspects of FECRT . The first important finding is that a successful interpretation of the FECRT is not always possible and that this is not always due to factors inherent to the design of a study , but can also be caused by factors inherent to host-parasite interactions ( e . g . , level of excretion and aggregation of eggs within the host population ) . For a threshold of 90% ( hookworm ) , unreliable FECRT results were obtained when sample sizes were small or when moderate sample sizes were combined with highly aggregated FEC . For a threshold of 95% ( A . lumbricoides ) , diagnostic performance was poor when sample sizes are small and when moderate sample sizes were combined with highly aggregated FEC and/or with FEC methods with a low detection . Second , our results highlight that the interval of TDE for which the FECRT remains inconclusive ( so called ‘grey’ zone , Figure 3 ) is unexpectedly small , ranging from 87 . 5% to 92 . 5% for hookworm and from 92 . 5% to 97 . 5% for A . lumbricoides . Third , the study design with the greatest discriminatory power to classify drug efficacy requires examination of 50 to 200 subjects , for both hookworm and A . lumbricoides . For this interval of sample sizes , there were no additional requirements on the detection limit of the FEC method and the level of aggregation of the FEC did considerably influence the interpretation of the FECRT . Only for hookworm , mean pre-DA FEC <150 EPG led to a less reliable interpretation of FECRT , as the ‘grey’ zone ranged from 85 . 0% to 92 . 5% . Overall , our findings contrast sharply with the recommendations provided by WHO [10] , explained by two main reasons . First , our analysis indicates that including subjects with any STH status ( absence ( true or falsely ) or presence of eggs in stool ) will not affect the final interpretation of the FECRT , yet this allows a dramatic reduction in the required sample size . For example , monitoring drug efficacy in a low risk-population ( STH prevalence = 20% ) would require screening 1 , 000 subjects ( in order to obtain a sample size of 200 infected subjects ) according to WHO guidelines , whereas according to our findings only 50 to 200 subjects are required . This , however , remains a large interval of possible sample sizes , which requires further refinement . The outcome of the efficacy used to validate the tree-based models , indicated that a minimum of 200 subjects are recommend , as this sample size allowed for a reliable detection of normal or reduced efficacy in all 14 trials where the FECRT fell outside the ‘grey’ zone . Second , our results do not support the need for four fecal samples per subject ( two before and two after administration of drugs ) , and hence will further reduce the costs to implement a monitoring system . This is mainly based on the fact that in the present simulation of FECRT based on two stool samples per subject ( one before and one after administration of drugs ) , and hence partially ignoring any variation in FEC due to differences in FEC across days , did not result in an underestimation of the required sample size . Moreover , the detection limit of the FEC method revealed to be less critical than anticipated , highlighting the importance of the feasibility of the FEC method used . Recently , Kato-Katz thick smear , FLOTAC , and McMaster egg counting methods have been compared for their feasibility in diagnosing STH [13] , [25] . Of these three methods , McMaster egg counting method was considered the most feasible , as the procedure does not include centrifugation steps ( vs . FLOTAC ) and allows quantifying all STH in one single reading ( vs . Kato-Katz thick smear ) . Based on these studies assessing the cost of these diagnostic methods , it is estimated that the average time for preparing , reading and examining one stool sample is roughly 5 min for McMaster egg counting method , 10 min for Kato-Katz thick smear , and 26 min for FLOTAC [13] , [25] . The combination of different statistical procedures ( Monte Carlo simulation and tree-based models ) , allowed for a cost-reduced data generation providing a decision support framework rather than a descriptive analysis . At present , both approaches are increasingly applied in various aspects of both public [26] and animal health [27] . However , this statistical approach to evaluate FECRT has limitations that must be acknowledged . First , it is assumed that worm abundance is adequately reflected by FECs , yet it remains unclear whether this holds true for STH infections in humans , particularly for hookworms . For this STH , a density dependent fecundity - female worms that survived the anthelmintic treatment produce relatively more eggs - has been described in dogs ( Ancylostoma caninum ) [28] . These density dependent effects imply a reduced drug efficacy for subjects with higher pre-intervention FEC , but this has not yet been observed in human trials [9] . Secondly , the generation of the observed FEC did not consider additional variation caused by properties of the detection technique beyond the detection limit , which impedes a straightforward extrapolation of the findings across FEC methods . Both the specific density of the flotation solution ( large difference in mass of parasite eggs ) [29] and the inclusion of a centrifugation step ( increasing FEC when included ) have an important impact on the FEC obtained by various FEC methods [30] . For Kato-Katz thick-smear , the templates used to substitute the calibrated weight of examined stool by a calibrated volume introduce an additional variation [25] , [31] . Additionally , differences in processing samples across investigators or laboratories should not be neglected [25] , [31]–[33] . As a consequence , it will become necessary to quantitatively validate the ability of both old and novel techniques to determine true FECRT rather than simply compare their ability to correctly diagnose the presence or absence of infections [12] . Thirdly , this simulation did not include any STH populations defined by a mean baseline FEC <50 EPG and/or a k <0 . 01 . Although this kind of populations are to be expected after a successful implementation of preventive chemotherapy programs , the simulation still represents a significant part of the populations at risk of STH infections . This is in particular when the target to administer anthelmintics to at least 75% of the population at risk by 2010 set by World Health Assembly Resolution 54 . 19 in 2001 , was not met ( coverage was <20% in 2008 ) [34] . Moreover , it is most likely that by then the endpoints of these programs will shift from ‘reducing morbidity’ to ‘eradicating’ of STH infections , which will demand a shift in study design and efficacy indicators of monitoring programs of anthelmintic efficacy . In conclusion , this study points out that the final interpretation of the FECRT was affected by a complex interplay of factors inherent to both study design and host-parasite interaction . The results also indicate that current WHO guidelines need to be revised . Based on the current study and the outcome of previously assessed efficacy trials [9] , we propose to include a minimum of 200 subjects independent of STH status ( subjects who are not excreting eggs can also be included ) and to examine two stool samples per subject ( one at baseline and one at follow-up ) . In this set-up , the choice of FEC method is not critical and arithmetic-mean based FECR <95% for A . lumbricoides and <90% for hookworms can be used as indicators for reduced efficacy and potential presence of drug resistance against albendazole . | The reduction in number of eggs excreted in stools after drug administration is a primary parameter to monitor the efficacy of drugs against parasitic worms . Guidelines on how to perform such a fecal egg count reduction test ( FECRT ) are provided by the World Health Organization . However , it remains unclear to which extent these guidelines are cost-effective . We , therefore , performed a simulation study in which the FECRT was performed under varying conditions to determine the critical values for sample size , the detection limit of the fecal egg count ( FEC ) method , mean baseline FEC , and variation of FEC across host population that allow for conclusive FECRT results . The results revealed that a reliable monitoring system demands a sample size of 200 subjects and that in some cases FECRT results may be thwarted by low mean baseline FEC . For this sample size , the detection of the FEC method or the variation of FEC across the host population did not affect the FECRT results . Our findings underscore that the current guidelines are not cost-effective , demanding too much financial and technical resources . We , therefore , propose novel guidelines to support future monitoring programs . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"medicine",
"infectious",
"diseases",
"public",
"health",
"and",
"epidemiology",
"clinical",
"research",
"design",
"drugs",
"and",
"devices",
"global",
"health",
"public",
"health"
] | 2011 | Novel Insights in the Fecal Egg Count Reduction Test for Monitoring Drug Efficacy against Soil-Transmitted Helminths in Large-Scale Treatment Programs |
The mitochondrial ( mt ) genome is present in many copies in human cells , and intra-individual variation in mtDNA sequences is known as heteroplasmy . Recent studies found that heteroplasmies are highly tissue-specific , site-specific , and allele-specific , however the functional implications have not been explored . This study investigates variation in mtDNA copy numbers ( mtCN ) in 12 different tissues obtained at autopsy from 152 individuals ( ranging in age from 3 days to 96 years ) . Three different methods to estimate mtCN were compared: shotgun sequencing ( in 4 tissues ) , capture-enriched sequencing ( in 12 tissues ) and droplet digital PCR ( ddPCR , in 2 tissues ) . The highest precision in mtCN estimation was achieved using shotgun sequencing data . However , capture-enrichment data provide reliable estimates of relative ( albeit not absolute ) mtCNs . Comparisons of mtCN from different tissues of the same individual revealed that mtCNs in different tissues are , with few exceptions , uncorrelated . Hence , each tissue of an individual seems to regulate mtCN in a tissue-related rather than an individual-dependent manner . Skeletal muscle ( SM ) samples showed an age-related decrease in mtCN that was especially pronounced in males , while there was an age-related increase in mtCN for liver ( LIV ) samples . MtCN in SM samples was significantly negatively correlated with both the total number of heteroplasmic sites and with minor allele frequency ( MAF ) at two heteroplasmic sites , 408 and 16327 . Heteroplasmies at both sites are highly specific for SM , accumulate with aging and are part of functional elements that regulate mtDNA replication . These data support the hypothesis that selection acting on these heteroplasmic sites is reducing mtCN in SM of older individuals .
Mitochondria are the central cellular structures for energy production . While most human mt proteins are encoded in the nuclear genome and imported to the mitochondrion following translation , mitochondria also contain their own genome [1] . In addition to tRNA and rRNA genes , the mt genome harbors 13 genes encoding proteins of the respiratory chain . The human mt genome is usually present in several copies per mitochondrion and the total number of mtDNA copies per cell varies greatly between different individuals and different tissues of the same individual [2 , 3] . For example , myocardial muscle cells contain on average more than 6 , 000 copies per cell [4] , while leukocytes have around 350 mtDNA copies per cell [5] . Replication and degradation of mtDNA are coupled in order to keep mtDNA levels constant in a cell [3] . Changes in mtCN have been correlated with several diseases and some factors that influence mtCN have been described [3 , 6] . These are for example replication regulating proteins such as the single-stranded-DNA-binding protein mtSSB , which stimulates mitochondrial DNA polymerase ɣ and therefore increases replication rate [6] . In addition , overexpression of mitochondrial RNA polymerase and its mitochondrial transcription factor A ( TFAM ) enhance replication by increased synthesis of primers for replication [6] . A reduction of TFAM in heterozygous knockout mice showed a strong reduction in mtCN and a total knockout resulted in a total lack of mtDNA and lethality [7] . However , it is still not known how inter-individual differences arise in mtCN , to what extent such inter-individual differences in mtCN are correlated across different tissues of an individual , and what effect these natural differences might have on healthy individuals . Several studies have examined changes in mtCN with age; while some detected a decrease of copy number with increasing age in healthy individuals [8–10] , others failed to identify age-dependent changes in copy number [4 , 11] or showed an increase for certain tissues [12] . Thus , there is currently much uncertainty concerning the role of age and other factors on mtCN . Copy number is usually estimated by quantitative PCR methods [4 , 10] , although some studies have utilized shotgun sequencing data [9 , 13–15] . However , comparisons of different methods for estimating mtCN are rarely done [9] and moreover most studies have focused on a single tissue . Here , we analyze mtCN in 12 different tissues that were obtained at autopsy from 152 individuals . We compared three different methods for estimating mtCN: ddPCR ( in 2 tissues ) ; shotgun sequencing ( in 4 tissues ) ; and capture-enriched sequencing ( in all 12 tissues ) . Furthermore , we examined the influence of various features such as age , haplogroup and sex on mtCN . We also inquired whether mtCN can be linked to mtDNA heteroplasmy , i . e . intra-individual variability in the mtDNA sequence . Several studies revealed that heteroplasmy increases in healthy human individuals with aging [16–19] . In addition , it has been shown that there is a tissue-specific accumulation of heteroplasmy at defined sites [15 , 17–21] , which might hint to either positive selection on specific sites [15 , 19] or a lack of negative selection due to relaxation of functional constraints [22] . All samples examined for mtCN in the present study were investigated previously for heteroplasmy [19] . As the most common heteroplasmic sites are found exclusively in the control region of the mitochondrial genome , which is essential for replication and transcription initiation and regulation [23 , 24] , we hypothesized that heteroplasmic mutations in the control region could have an influence on mtCN , and indeed we report here that an increase in the minor allele frequency ( MAF ) at two sites that are frequently heteroplasmic in skeletal muscle is associated with an age-related decrease in mtCN .
Twelve different tissues ( blood: BL; cerebellum: CEL; cerebrum: CER; cortex: CO; kidney: KI; large intestine: LI; liver: LIV myocardial muscle: MM; ovary: OV; small intestine: SI; skin: SK; and skeletal muscle: SM ) were obtained at autopsy from 152 human individuals ( 85 males , 67 females ) of mostly European origin and DNA was extracted as previously described [19] . The collection of samples and the experimental procedures were approved by the Ethics Commissions of the Rheinische Friedrich Wilhelm University Medical Faculty and of the University of Leipzig Medical Faculty ( Approval numbers: Rheinische Friedrich Wilhelm University: 121/11 , University of Leipzig: 349-11-07112011 and 305-15-24082015 ) . Libraries for sequencing were prepared as previously described [19] . Shotgun sequencing ( without mtDNA enrichment ) was performed on an Illumina HiSeq platform using 95 base pair , paired-end reads . Two runs were performed: run 1 consisted of pooled libraries from four tissues ( with all samples from BL ( 148 samples ) , SM ( 152 ) , LIV ( 152 ) and MM ( 150 ) ) and run 2 consisted of the same libraries from two tissues , BL and SM . To check reproducibility of the results with independent libraries , we prepared a new subset of libraries from the BL and SM DNA extracts from each of 26 individuals of mixed age , sex and mtCN and performed shotgun sequencing on an Illumina MiSeq platform using 150 base pair , paired-end reads ( referred to as shotgun sequencing run 3 ) . Reads of at least 50 bp length were mapped against the human reference genome 19 ( hg19 ) using BWA with default settings [25] and without filtering for high mapping quality . All reads that aligned successfully , i . e . both mates of a pair mapped to the same chromosome and within a distance to each other of at most 3 standard deviations of the insert size distribution , were analyzed and the length of their insert sizes summed up per chromosome . For each chromosome the number of mappable bases was determined by removing any poly-N stretch in the hg19 ( > 5 consecutive Ns ) and counting the number of remaining bases . The number of aligned bases of each chromosome was divided by the number of mappable bases on this chromosome to obtain the coverage per chromosome and sample . Samples were subsequently removed from the data set when either of the following criteria was fulfilled: 1 ) the number of bases aligning to mtDNA was ≤ 60% of the total mtDNA length; 2 ) the logarithmic normalized SD of the autosomal coverage was identified by a Grubbs test [26] as an outlier . For each tissue the standard deviation of the coverage of each autosome was determined over all remaining samples . Chromosomes that were identified as an outlier by a Grubbs test based on their SD in 50% of the tissues were not considered in the mtCN calculation . The mtCN was determined by the following formula [9]: mtCN=2*mtDNAcoverage122∑i=122autosomalcoverage ( 1 ) In order to account for the difference in absolute mtCN per tissue , the SDs were normalized by the mean autosomal coverage of a sample . In a previous study , Illumina sequencing had been performed after capture-enrichment for mtDNA [19] . The capture-enriched sequencing data from samples from BL ( 139 ) , CEL ( 150 ) , CER ( 143 ) , CO ( 152 ) , KI ( 151 ) , LI ( 150 ) , LIV ( 151 ) , MM ( 149 ) , OV ( 47 ) , SI ( 150 ) , SK ( 152 ) and SM ( 150 ) , were processed and mtCN estimation carried out as described above for shotgun sequencing data . The mtCN from BL ( 150 samples ) and SM ( 152 ) samples was determined using droplet digital PCR ( ddPCR ) , a quantitative PCR method that allows an absolute measurement of the number of target DNA molecules [27] . A 20 μl PCR mix was prepared and dispersed into up to 20 , 000 droplets . After amplification , droplets that included template DNA were identified by fluorescent dyes and counted . Specific primers ( S1 Table ) for amplification of mt and nDNA regions were modified according to [10] . For determination of mtCN in BL , mtDNA and nDNA were amplified in the same multiplex reaction using HEX- and FAM-labeled probes for n and mtDNA , respectively . Reactions were performed in ddPCR Supermix including primers , probes and 1 ng template according to manufacturer’s instructions . Due to the high mtCN in SM , determination of nDNA and mtDNA in SM samples was performed in two separate reactions with 10 ng template for nDNA and 20 pg template for mtDNA and either nuclear or mt specific primers in EvaGreen Supermix . To control for differences between DNA preparations , at least 12 DNA samples per tissue type were extracted twice and the mtCN of different preparations was compared . MtCN was calculated after correction for dilution factors by mtCN=2*mtDNAcountsnDNAcounts ( 2 ) For each sample the arithmetic mean and the standard deviation ( SD ) of mtCN was calculated over all replicates . To account for the separate ddPCR reactions for nDNA and mtDNA measurement per SM sample , the SDs were averaged over the individual SDs of the SM samples . To compare the SDs between BL samples and SM samples , the SD was normalized by the mean mtCN of each sample . Both the arithmetic mean and the SD were subsequently averaged over all samples of a tissue type . In order to estimate the likelihood that recent mtDNA insertions into the nuclear genome ( NUMTs ) could artificially increase the mtCN estimates from sequencing data , we analyzed the length distribution of identified NUMTs from a NUMT database [28] and compared it to the insert size distribution in our data set . When paired-end reads overlapped , the length of the merged read was defined as the insert size [29] . If reads were too far apart to be merged , the length of the region flanked by the two reads was defined as insert size . A read from a NUMT can only be aligned falsely with high quality to mtDNA when the read does not contain a flanking nuclear DNA signature of at least a few high-quality bases . For each insert size , all NUMTs longer than the insert were extracted from the published list as those inserts could potentially fully fall in a NUMT region . The probability of false alignment to the human mtDNA sequence was then estimated , taking into account the number of putative read insert positions within a NUMT , the number of NUMTs in the human genome that were larger than each read insert size , and the average coverage of all autosomes in each sequencing run , as follows: probabilityp=∑i=1Mi*∑j=1Nj*cl , withi=insertsize , M=maximalinsertsize , j=numtlength , N=max . NUMTlengthc=averagecoverageautosomesl=lengthofthemappablegenome ( 3 ) The mtCN values for each tissue were log-transformed to produce a normal distribution of the residuals for comparison with other parameters , e . g . age , sex , haplogroup , etc . An underlying normal distribution of the residuals is a requirement for many statistical tests ( including the linear regression models used here ) and was therefore formally tested for every tissue using a Shapiro-Wilk test . In order to fulfill the requirement of a normal distribution , one LIV-sample ( LIV253 ) with very high mtCN had to be excluded from the shotgun sequencing data sets prior to further investigations . Linear regression and Pearson correlation analysis as implemented in R [30] was performed and the resulting p-values were corrected for multiple testing using the Benjamini-Hochberg method [31] . Partial regression analysis was used to investigate the combined effect of age and heteroplasmy levels on mtCN [32] . When performing linear regression analysis between MAF of specific heteroplasmies and mtCN , only heteroplasmies that were present in at least 10 individuals were considered . We tested for correlations of mtCN with single heteroplasmic sites , pairs of sites occurring in the same individual , and triplets of heteroplasmic sites , using linear regression and Pearson correlation analysis as described above .
For all analyses , tissue samples that had been collected at autopsy from 152 individuals and 12 different tissues in a previous study [19] were used . MtCN was estimated using shotgun sequencing ( in BL , MM , LIV and SM ) , capture-enrichment sequencing ( in BL , CEL , CER , CO , KI , LI , LIV , MM , OV , SI , SK and SM ) , and ddPCR ( in BL and SM ) . The first two approaches are based on sequencing data and mtCN is estimated by dividing the number of bases that map to the mtDNA by the number of bases mapped to the nuclear genome . In order to estimate reproducibility , variation in coverage across the different autosomes was measured and chromosomes with high variation ( as identified by the Grubbs outlier test ) were removed . Chromosomes 16 and 19 were excluded from the mtCN calculation in the shotgun sequencing data as the SD of their coverage was enhanced in at least 50% of the tissues according to the Grubbs test ( S1 Fig ) . In addition , for all 12 tissues from all individuals , mtCNs were estimated using capture-enriched sequencing data . Here , sequencing libraries had been enriched for mtDNA prior to sequencing [19] , and so mtCN estimates will be elevated in the capture-enriched data; however , relative mtCN estimates might be the same , and so allow for comparisons between individuals . According to the Grubbs outlier test , SD values for the coverage of chromosomes 1 , 16 and 19 were enhanced and therefore these chromosomes were not included in mtCN calculations from capture-enriched sequences ( S1 Fig ) . MtDNA insertions into the nuclear genome ( NUMTs ) could potentially artificially increase mtCNs calculated from sequencing data if reads deriving from recently-inserted NUMTs are falsely aligned to mtDNA . With currently used alignment tools , NUMTS that have been in the nuclear genome for a long time are identified as they have accumulated mutations that distinguish them from authentic mtDNA sequences . Only NUMTs that have inserted into the nuclear genome rather recently have not had enough time for distinguishing mutations to occur . As we are therefore not able to directly identify reads arising from recent NUMTs , we instead estimate the probability of incorrectly attributing a NUMT read to the mtDNA genome , using the length distribution of a previously published list of identified NUMTs [28] . For each read insert length , all NUMTs longer than that length were identified as NUMTS from which reads could be generated that could be fully placed within the NUMT . The chance that a read from a NUMT was falsely aligned to the mtDNA genome , given the observed read distribution , was ≤0 . 06% , and therefore NUMTS were considered to have a negligible impact on mtCN estimation . We next evaluated the reproducibility of the estimated mtCNs for the two sequencing methods . To estimate standard deviations ( SDs ) of the methods , intra-individual differences in the chromosomal coverage were evaluated , and the resulting SD estimates were averaged over each tissue and normalized by the average mtCN for the tissue . The shotgun sequencing method returned normalized SD values for chromosomal coverage of <3 . 2% , while the SDs of mtCN estimates from capture-enrichment were between 8 . 6 and 18 . 2% ( average: 12 . 6% ) ( Table 1 ) , indicating more variation in chromosomal coverage of the capture-enriched data ( S1 Fig ) . As both shotgun and capture-enrichment data were available for four tissues ( BL , SM , LIV , and MM ) , we further investigated the reliability of mtCN estimates from capture-enrichment by testing if mtCN estimates from both methods were correlated . While there was a greater enrichment in mtCN for BL than for the other tissues , for all four tissues there was a significant correlation between mtCN values estimated via shotgun sequencing vs . capture-enriched sequencing ( Fig 1A ) . Hence , mtCN estimates based on capture-enrichment sequencing data were considered suitable and used for further investigation into factors influencing mtCNs . In a third approach , mtCN was estimated by ddPCR in SM and BL , with each sample measured at least 3 times and the SD calculated over all replicates per sample . Using ddPCR the normalized SD was 10–13% for replicates ( Table 1 ) . The estimated mtCN values from ddPCR vs . shotgun sequencing of the same samples were significantly correlated ( BL samples: Pearson correlation p<0 . 001 , R2 = 0 . 75; SM samples: p = p<0 . 001 , R2 = 0 . 64 , Fig 1B ) and the mean mtCN values were quite similar ( BL: mean mtCN from shotgun sequencing = 257 and from ddPCR = 260; SM: mean mtCN from shotgun sequencing = 2788 and from ddPCR = 2744; Table 1 ) . To further evaluate the error rates in ddPCR vs . shotgun sequencing , shotgun sequencing was repeated with the same libraries for samples from BL and SM and mtCN and the average SD of the replicates was estimated . The regression coefficient R2 was 0 . 92 for BL and 0 . 99 for SM ( S2 Fig ) and the average SD was ≤8 . 5% and therefore lower than the SD for ddPCR . However , the shotgun sequencing replication was carried out on the same libraries , and hence did not include variation from library preparation or other experimental procedures . To account for this , new libraries were prepared from BL and SM from 26 individuals , sequenced on the MiSeq platform , and the mtCNs were compared ( S2 Fig ) . For both tissues mtCNs could be determined with high reproducibility from shotgun sequencing of an independent library , with R2 values of 0 . 88 and 0 . 98 for BL and SM samples respectively . As with the ddPCR experiments , the SD of the replicates for mtCN in SM was lower than that for BL ( 0 . 044 vs . 0 . 103 for SM and BL , respectively ) , indicating a higher reproducibility for the tissue with higher mtCN ( Table 1 ) . Remarkably , two samples with extremely high mtCN were identified . One SM sample from a 71 year old male who died of cardiac arrest showed a 137-fold higher mtCN ( shotgun sequencing ) than the average of all other SM samples ( S3 Fig ) and a more than 60-fold higher mtCN than the sample with the 2nd highest mtCN . Results from capture-enrichment ( 32-fold higher than the average , 14-fold higher than the second highest mtCN ) and ddPCR ( 80-fold higher than average , 30-fold higher mtCN than the second highest ) were similar . In addition , one LIV sample from a 72 year old male who died of multiple organ failure showed a 20-fold higher mtCN than the tissue average in shotgun sequencing ( 8-fold in capture-enrichment ) . The mtCN estimates for other tissues in these two individuals are all within the average range for that tissue , indicating that this extreme increase in mtCN is tissue-specific . In other tissues , the highest mtCN value was maximum 4-fold higher than the average in shotgun sequencing ( 6-fold in capture enrichment ) , showing that extremely high mtCNs are rare in the data set . For each method , we tested if mtCN estimates were correlated between pairs of tissues within an individual ( Fig 2 ) . All p-values were corrected for multiple testing using Benjamini-Hochberg correction [31] . In the capture-enriched data , positive correlations were identified between SI and LI ( r = 0 . 31 , p<0 . 001 ) , between CO and CER ( r = 0 . 33 , p<0 . 001 ) and between KI and CER , CO , SI and SM . In addition , mtCN in SK was negatively correlated with that in SI , LI and in LIV . In the shotgun sequencing data there was a negative correlation of mtCN in SM vs . LIV ( Pearson’s r = -0 . 23 , p = 0 . 042 , Fig 2 ) , but this was not observed in the capture-enriched data ( Fig 2 ) . Overall , we did not observe any regular pattern in the intra-individual variation in mtCN among different tissues ( S4 Fig ) . We also investigated associations between mtCN and age , sex , or haplogroup . There was a strong age-related decrease of mtCN in SM ( r = -0 . 25 , p = 0 . 016 in shotgun sequencing , Fig 3A , S2 Table ) . Remarkably , this association holds only for males ( r = -0 . 35 , p = 0 . 008 in males compared to r = -0 . 15 , p = 0 . 442 in females; S2 Table ) . Similar correlation coefficients were obtained for mtCN estimates from capture sequencing and ddPCR , although the p-values were not significant after correction for multiple testing ( S2 Table ) . In LIV , mtCN shows an increase with age that is significant in capture enrichment data ( r = 0 . 27 , p = 0 . 022 ) and approaches significance in shotgun sequence data ( r = 0 . 20 , p = 0 . 088; S2 Table ) . This increase was mainly due to a subset of individuals ( Fig 3A ) with higher mtCNs . When looking at samples with mtCN≥4 , 500 , these showed a stronger correlation of mtCN with age ( r = 0 . 64 , p = 0 . 016 ) than the group mtCN≤4 , 500 ( r = 0 . 19 , p = 0 . 048 ) . No solely sex-related associations with mtCN were identified for any tissue ( Fig 3B , S2 Table ) and there were no significant associations between mtCN and major haplogroup ( Fig 3C and S3 Table ) . As all samples in this study were previously analyzed for heteroplasmy [19] , we tested for associations between mtCN and the total number of heteroplasmic positions as well as the MAF at specific heteroplasmic positions . SM exhibited a highly significant negative correlation between mtCN and the total number of heteroplasmies for all three methods ( r = -0 . 27 to -0 . 31 , all p≤0 . 002; Fig 4 and S4 Table ) . BL and LIV also exhibited significant negative correlations between the total number of heteroplasmies and mtCN , but not for all of the methods ( S4 Table ) . As the number of heteroplasmic sites has been shown to significantly increase with age [16 , 19] , we analyzed the influence of age and the total number of heteroplasmic sites on mtCN in SM by partial regression . MtCN was best explained by age ( p<0 . 05 ) , although the number of heteroplasmies also showed a slight correlation with mtCN that was borderline significant for shotgun sequencing data ( p = 0 . 049 , Fig 4 ) . In LIV , changes in mtCN were significantly associated with age ( p<0 . 05 , Fig 4 ) , whereas BL showed a higher partial regression of mtCN with heteroplasmy than with age ( Fig 4 ) . Overall , these results indicate that correlations of mtCN with the total number of heteroplasmies are explained mainly by age rather than the actual increase in the number of heteroplasmic sites . With respect to associations between mtCN and the MAF at single heteroplasmic sites , analyses were limited to positions that were heteroplasmic in at least ten individuals for a tissue ( S5 Table ) . After correction for multiple testing , sites 408 and 16327 each showed a significant association between mtCN and MAF in SM for shotgun sequence data ( p = 0 . 014 for both sites , Fig 5 and S6 Table ) . While the consensus allele at position 408 is a T in all tissues and individuals in this study , an A allele is observed as a heteroplasmy ( MAF = 2–28 . 2% ) in 65/152 individuals in SM . The consensus allele at position 16327 is a C in all but one individual in this study , but heteroplasmy involving a T ( MAF = 2–20% ) occurs in SM in 46/152 individuals . Other alleles were not found at either position in these samples [19] . As the MAF at these positions also increases with age [19] , we applied partial regression to test whether age or heteroplasmy was more strongly associated with mtCN . While age was again significantly correlated with mtCN ( p = 0 . 001 for both sites ) , the MAF at both heteroplasmic sites also showed significant associations with mtCN ( p = 0 . 022 ( 408 ) , p = 0 . 02 ( 16327 ) , Fig 5 ) . The total number of heteroplasmies , however was not significantly associated with mtCN ( p>0 . 05 ) . This indicates that changes in mtCN in SM are associated with age , with the MAF at positions 16327 and 408 also explaining some of the change in mtCN . As the aging-dependent decline in mtCN was significant in males but not in females , we asked whether there were sex-dependent differences in the correlation between mtCN and MAF , too . After correction for multiple testing , neither MAF at site 408 nor at site 16327 was significantly correlated with mtCN in males or females alone ( males: 408: p = 0 . 64 , 16327: p = 0 . 64; females: 408: p = 0 . 4 , 16327: p = 0 . 4 ) . No stronger correlations were found when testing the MAF of pairs or tripletons of heteroplasmic sites , indicating that heteroplasmy at two or three specific sites was not more associated with variation in mtCN than heteroplasmy at single sites . In sum , mtCN in SM differed dramatically from other tissues in exhibiting highly significant age , sex and heteroplasmy-dependent associations .
In this study , mtCN from different tissues was estimated using three different methods . Exact measurements of mtCN can be complicated in some tissues as intra-individual differences in mtCN of a single tissue can occur . These differences arise from mosaic-like structures as those found especially in myocardial muscle [4] and one should therefore focus investigations of mtCN on rather homogenous tissues . In ddPCR and shotgun sequencing experiments , SM samples showed smaller SDs for multiple measurements than BL samples . This could indicate that the precision of the two methods is increasing with higher mtCNs as small absolute measurement errors have less impact on the final results . In general , the shotgun sequencing method returned normalized SD values that were smaller than those for ddPCR . The high variance between measurements in ddPCR might arise from the several pipetting steps for mtDNA- or nDNA-specific primers as well as additional dilution steps between nDNA and mtDNA-determination . In Illumina shotgun sequencing , the only step in which mtDNA and nDNA compete is the loading of DNA molecules into the nanowells of the flowcell . The distribution of reads across chromosomes indicates that for most chromosomes this happens randomly without any bias for certain sequences . Chromosomes that had to be excluded from the data set showed a large SD across all samples due to the assembly of reads with lower mapping quality . Since shotgun sequencing exhibited high reproducibility of mtCN with two independently prepared libraries and moreover allows high-throughput analysis of large sets of samples , this is the preferred method for mtCN determination . We assumed that capture-enrichment linearly increases the amount of mtDNA without saturation effects when estimating mtCN from sequencing after capture-enrichment . However , we observed that the enrichment process is non-linear , resulting in a stronger enrichment of mtDNA in samples with smaller mtCN . Samples with very high mtCN , like the SM-sample with a 100-fold increase in mtCN , already have a very high proportion of mtDNA prior to enrichment and will therefore be underestimated after enrichment . Samples with lower mtCN , like BL samples , have a low starting ratio that is strongly increased in the enrichment process leading to overestimated mtCNs after enrichment . Hence , enrichment most strongly influences the tails of the mtCN distribution; however , the overall high correlations between mtCN estimates from capture-enriched vs . shotgun data indicates that mtCN estimates by the former method do reliably reflect relative mtCN values . We identified two samples from different individuals that had 20-fold ( LIV ) and a ≈100-fold ( SM ) higher mtCN compared to the tissue average . Similar results were obtained with the different methods for mtCN determination , indicating that these high mtCN are not method-dependent errors . To our knowledge no such increases in mtCN have been described before . A strong increase of mtCN has been described when large parts of the coding region of the mtDNA are deleted [33] and steep increases in copy number occur during cell differentiation , such as an 1100-fold increase in mtCN between day 5 and 6 of differentiation of pluripotent embryonic stem cells [34] . In addition , external factors such as stress [35] or diseases such as cancer , diabetes or HIV ( reviewed in [36] ) can increase mtCN . Cai et al . reported a stress-induced increase in mtCN in liver , but a decrease in SM in mice . Interestingly , mtCN in those two tissues was negatively correlated in our shotgun sequencing data . As we do not have any information on the stress levels of the individuals in our sample set , no further comparison of stress levels with mtCN is possible . However , previous studies found that stress increased mtCN by only around 2-fold [35] , which cannot explain the samples with very high mtCN found here . An excess of replication has been associated with several mtDNA regulatory proteins , such as TFAM [6] or PGC-1 , a common transcriptional co-activator of nuclear receptors [37] . The individuals we identified did not suffer from any diagnosed disease that might impact mtCN , and these extreme mtCNs occurred in just one tissue in each individual . Heteroplasmy , haplogroup or age do not explain this extreme increase in mtCN . As all enzymes required for mtDNA replication control , like TFAM or PGC-1 , are encoded in the nucleus [24] , we speculate that mutations in nuclear-encoded genes might trigger this increase; however an environmental cause cannot be excluded . Following mtCN estimation , several parameters were tested for possible correlations with mtCN using linear regression and a Pearson correlation test . The data from capture-enrichment returned the fewest significant results , which probably arises from Benjamini-Hochberg corrections for multiple testing as for capture-enrichment there were 11 or 12 tissues ( ovarian tissue data could only be analyzed in females ) analyzed compared to four in shotgun sequencing and two in the ddPCR . As the F- and r-values , which were not corrected for multiple tests , were in the same range for all three methods , we conclude that the most striking correlations were identified regardless of the method used . To our knowledge , few studies have investigated differences in relative mtCN between different tissues of an individual; however , one previous study did investigate mtCN in samples from brain , SM and MM from 50 individuals [11] . This study identified correlations of mtCN between most tissues . In our study , different tissues of the same individual showed a high variation in mtCN , not only for the total number of copies , but also for the relative number with respect to the tissue average . Positive correlations were mainly found for similar tissues , such as different parts of the brain or for large and small intestine . Interestingly , mtCN in skin had a strong negative correlation with mtCN in both intestinal tissues ( as well as liver ) . Skin and intestine are surface tissues with exposure to different microbiomes [38] , and hence different environmental circumstances might be influencing mtCN . The variation in mtCN between different tissues of an individual is largely uncorrelated ( Fig 2 ) . Thus , it is not the case that individuals tend to have generally high or low mtCN values across all tissues , but rather mtCN varies in a tissue-specific fashion . This is in good agreement with a recent study investigating stress-induced changes in mtCN [35] , which showed that stress increased mtCN in liver , but decreased mtCN in muscle in mice . Therefore , external factors can produce tissue-specific changes in mtCN . Specific haplogroups have been suggested to influence mtCN . For example , a comparison of cybrid cells from haplogroups H and Uk found that cells of haplogroup Uk exhibited lower mtCN than those of haplogroup H [39] . In our present study , no individual belonged to haplogroup Uk . Another study comparing haplogroups H and J found that the haplogroup J defining mutation increased TFAM binding and as a consequence mtDNA replication and mtCN [40] . In our data set no significant differences in mtCN were detected for different haplogroups , including haplogroup J . However , the relatively small sample size for each haplogroup in our study ( e . g . , only 14/152 individuals with haplogroup J ) means that we lack power to identify potential haplogroup-related effects on mtCN , and further studies with larger sample sizes are warranted . Age is another factor that could influence mtCN; for example , previous studies have described age-dependent decreases in mtCN for BL in individuals older than 50 years [8 , 10] . Although in our study all three methods exhibited a negative correlation between age and mtCN in BL for individuals over 50 , only for ddPCR did the correlation approach significance ( p = 0 . 09 , S2 Table ) . In contrast , the age-dependent decrease in SM is very striking in our data , especially in males . This result seems to contradict other studies of mtCN in the same tissue [4 , 41] . However , the study of Barthelemy et al . investigated samples from younger individuals ( 2–45 years , n = 16 ) , and no information on the sex distribution of the individuals was given in either study . The strong decrease in mtCN with age in males may be explained by the composition of muscle tissues . Males tend to have a higher ratio of fast-contracting muscle fibers ( type IIB ) over slow fibers ( type I ) [42] . Slow muscle fibers are characterized by a high activity , strong coupling of the electron chain and therefore high oxygen capacity . To account for aging effects , slow fibers reduce the coupling of the electron chain , resulting in low reactive oxygen species and stable mitochondrial function in old age [43] . Fast contracting muscle fibers , on the other hand , have a short longevity and are susceptible to aging . They show an advancing transformation to slow-contracting muscle fibers during aging which is accompanied by a reduction in mtDNA content [42] . Due to these differences in muscle composition , effects on mtCN during aging might be stronger in males than in females . Finally , we investigated the influence of heteroplasmy on mtCN . The total number of heteroplasmic sites was strongly negatively correlated with mtCN in SM in our data . As the number of heteroplasmic sites is also strongly age-related [16 , 19] and mtCN in SM is decreasing with age , the major predictor of mtCN remains unclear after linear regression analysis . We tried to account for this by partial regression , in which one factor is set constant while the effect of the other is investigated . For the total number of heteroplasmies , we found that age had a more significant correlation with mtCN , indicating that a non-specific enrichment of heteroplasmic sites is not the major effector for changes in mtCN . We then investigated potential associations between the MAF at specific heteroplasmic positions and mtCN . Previous studies have found that sequence variations in the polycytosine tract at positions 16180–16195 , and in a TFAM binding motif at position 295 in the mtDNA control region , are associated with changes in mtCN [5 , 40] . Another study on saliva from women that suffered from major depressive disorder showed that heteroplasmy at site 513 was significantly correlated with changes of mtCN [44] . In our study , we also found indications for associations between the MAF at common low-frequency heteroplasmic sites ( 408 and 16327 ) and mtCN in SM . The MAF at both positions has been shown to be highly correlated with age [18–21 , 45] . Heteroplasmy at position 16327 is highly specific for SM as it was not found in more than four individuals in any other tissue in our previous study [19] . Heteroplasmy at position 408 is even more common in SM , but was also found in some individuals in CER ( 12/142 ) and CO ( 13/152 , S5 Table ) with very low MAF ( ≤1 . 1% ) in the previous study [19] . Together with site 189 , site 408 was reported to be associated with aging-dependent decrease mtCN in 52 females from family units [46] . However , the authors of this study stated that the modest changes in MAF that they observed at those sites might not be the major determinant of mitochondrial dysfunctions during aging [46] . In our data set , we found that the correlation of MAF at site 408 with mtCN in either males or females alone was not significant after correction for age . The reduced sample size ( 84 males and 67 females instead of 151 in the entire sample set ) might reduce the power to detect significant effects in the data sets . Site 408 is adjacent to the transcription start site within the light strand promoter ( S7 Fig ) , which initiates production of an RNA primer for heavy strand replication [24] , while 16327 is located within the TAS-region ( S7 Fig ) , which regulates D-loop formation . The ratio of D-loops over double stranded DNA has been shown to be crucial for mtCN control [47] and parts of the TAS region , including site 16327 , have been identified as a putative binding site for unknown proteins [48] . As the occurrence of heteroplasmy arises in an allele-specific way , we hypothesize that heteroplasmy at positions 408 or 16327 might lead to changes in replication regulation . The fact that MAF at 408 was not correlated with mtCN in CER and CO might be explained by the much lower MAF in these tissues . Intra-individual deleterious mutations in mtDNA may persist at a low level due to a lack of negative selection [22 , 49] as they need to reach a critical frequency threshold in order to impact mitochondrial function [50] . If the negative selection pressure is relaxed in one tissue relative to others , due to the metabolic needs of that tissue ( especially during aging ) then tissue-specific deleterious heteroplasmies could potentially increase in frequency due to this relaxation of functional constraints [51–53] . Alternatively , low-frequency mutations might also rise in frequency via positive selection , if they are advantageous for that specific tissue [15 , 19] , e . g . if they influence replication rates [16 , 54] . The results of the present study do not bear on the reason ( s ) for the tissue-specific and age-related increases in heteroplasmy . However , reduced mitochondrial activity does lead to reduced accumulation of reactive oxygen species and protects mitochondria against aging [55] . We hypothesize that positive selection for a heteroplasmic allele could arise via a slight decrease in mtCN that results in a slowdown of mitochondrial metabolism . This , in turn , would reduce the production of aggressive reactive oxygen species and hence provide a better cellular fitness during aging , in accordance with the “survival of the slowest” hypothesis [56] . Both alternative minor alleles at these two positions ( 408A and 16327T ) are fixed in other haplogroups than those investigated here [57] . For example , 16327T is one of the alleles defining haplogroups C and U1B . One individual in our data set was identified as haplogroup U1B , but did not show any substantial difference in mtCN compared to other individuals ( S7 Table ) . While one individual does not allow relevant interpretation of haplogroup-dependent changes in mtCN , it does suggest that overall the 16327T allele is not highly deleterious . However , functional differences between haplogroups as well the functional impact of reduced mtCN in aging SM remain to be elucidated . In conclusion , we found that SM exhibits several interesting properties with respect to mtCN and heteroplasmy , including age-related decreases in mtCN in males , and decreases in mtCN associated with increasing MAF at two heteroplasmic sites in the control region . Several additional questions remain . For example , why do the minor alleles at tissue-specific heteroplasmic sites remain at rather low levels and do not reach higher and therefore presumably more detrimental frequencies in the cell during aging ? One effect of aging is the reduction of mitochondrial fusion and fission events that allow an exchange of genetic material between mitochondria in young individuals [58] . It has been proposed that reduced fusion and fission prevent cells from spreading detrimental mitochondria throughout the cell [59] and therefore decelerate aging effects . This might also explain the persistence of low frequency of heteroplasmy in aging individuals . Several sites in the control region show site-specific and allele-specific heteroplasmy in specific tissues , such as position 60 and 94 in LIV and KI or position 204 and 564 in MM [19] , yet in this study only the heteroplasmy at sites 408 and 16327 in SM showed a correlation with mtCN . Overall , elucidating the functional consequences of tissue-specific heteroplasmy remains a major challenge for further investigation . | The total number of mitochondrial genomes in a human cell differs between individuals and between the tissues of a single individual; however the factors that influence this variation remain unknown . We estimated mtDNA copy number ( mtCN ) in 12 different tissues of 152 individuals applying three different methods , and found age-related variation for two tissues: mtCN in skeletal muscle is negatively correlated with age ( especially in males ) while mtCN in liver is positively correlated with age . Overall , mtCNs of different tissues within an individual are mainly independent of each other , indicating that tissue-specific rather than individual-specific processes largely influence mtCN . Heteroplasmy refers to intra-individual differences in the sequence of the mtDNA genome and heteroplasmic mutations accumulate during aging . Linear and partial regression analyses of mtCN with heteroplasmy ( determined in a previous study of these same samples ) revealed that the decrease of mtCN in skeletal muscle is mainly correlated with an increasing total number of heteroplasmic sites , and with increasing minor allele frequency at two sites ( 408 and 16327 ) , that are heteroplasmic almost exclusively in skeletal muscle . As both sites are part of functional elements required for regulation of mtDNA replication , we suggest that selection may be acting via increasing heteroplasmy to reduce mtCN during aging . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"sequencing",
"techniques",
"mitochondrial",
"dna",
"population",
"genetics",
"computational",
"biology",
"cloning",
"regression",
"analysis",
"mathematics",
"chromosome",
"mapping",
"statistics",
"(mathematics)",
"dna",
"replication",
"genome",
"analysis",
"shotgun",
"sequ... | 2016 | Age-Related and Heteroplasmy-Related Variation in Human mtDNA Copy Number |
Laboratory-based studies have highlighted that pooling stool and urine samples can reduce costs and diagnostic burden without a negative impact on the ability to estimate the intensity of soil-transmitted helminth ( STH , Ascaris lumbricoides , Trichuris trichiura and hookworms ) and schistosome infections ( Schistosoma mansoni and S . haematobium ) . In this study , we compare individual and pooled stool examination strategies in a programmatic setting . Stool samples were collected from 2 , 650 children in 53 primary schools in Amhara Regional State , Ethiopia , during the national mapping of STHs and schistosome infections . Eggs of STHs and S . mansoni were quantified in both individual and pooled samples ( pools were made from 10 individual samples ) using a single Kato-Katz smear . A pooled diagnostic examination strategy provided comparable estimates of infection intensity with higher fecal egg count ( expressed in eggs per gram of stool ( EPG ) ) than those based on individual strategy ( Ascaris: 45 . 1 EPG vs . 93 . 9 , p = 0 . 03; Trichuris: 1 . 8 EPG vs . 2 . 1 EPG , p = 0 . 95; hookworms: 17 . 5 EPG vs . 28 . 5 EPG , p = 0 . 18; S . mansoni: 1 . 6 EPG vs . 3 . 4 EPG , p = 0 . 02 ) , but had lower sensitivity ( Ascaris: 90 . 0% vs . 55 . 0%; Trichuris: 91 . 7% vs . 16 . 7%; hookworms: 92 . 6% vs . 61 . 8%; S . mansoni: 100% vs . 51 . 7% , p < 0 . 001 ) . A pooled approach resulted in a ~70% reduction in time required for sample testing , but reduced total operational costs by only ~11% . A pooled approach holds promise for the rapid assessment of intensity of helminth infections in a programmatic setting , but it is not major cost-saving strategy . Further investigation is required to determine when and how pooling can be utilized . Such work should also include validation of statistical methods to estimate prevalence based on pooling samples . Finally , the comparison of operational costs across different scenarios of national program management will help determine whether pooling is indeed worthwhile considering .
Neglected tropical diseases ( NTDs ) are a group of parasitic , bacterial and viral infections that pose an important burden on public health , particularly in tropical and sub-tropical countries [1] . In 2015 , this group of 17 diseases resulted in approximately 26 million disability-adjusted life years ( DALYs ) and are considered an issue of global importance hindering progress towards the Sustainable Development Goals [2] . Soil-transmitted helminthiasis ( STH ) and schistosomiasis ( SCH ) are two of seven NTDs that are amenable to control through regular mass drug administration ( MDA ) [3] . Millions of people are infected worldwide , with each disease attributable for more than 10% of the overall NTD burden ( schistosomiasis: 11%; soil-transmitted helminthiasis: 14% ) [4] . Both are targeted primarily through school-based treatment programs , during which anthelmintic drugs ( albendazole or mebendazole for STH and praziquantel for SCH ) are administered to all school age children [5] . Fueled by the London Declaration on NTDs ( January 2012; [6] ) , the global treatment coverage of school-aged children has increased since 2011 for SCH ( 2011: <20% vs . 2016: 53 . 7% ) and STH ( 2011: ~30% vs . 2016: 69 . 5% [7–9] , with the ultimate goal of treating at least 75% of school-aged children in all endemic countries by 2020 [10] . The pharmaceutical industry donates anthelmintic drugs at-scale ( albendazole: GlaxoSmithKline , mebendazole: Johnson & Johnson , praziquantel: Merck KGaA ) [11] . However , MDA programs require substantial political and financial investments from endemic countries [12] . Additionally , there are costs for periodically assessing the epidemiology of the diseases . Prior to treatment , nationwide epidemiological surveys are used to target treatment appropriately . Periodic follow-up surveys are required to measure progress and determine whether scaling-down of MDA is justified [13] . Ethiopia published its first National Master Plan for NTDs in 2012 ( 2012–2015 ) , outlining plans to scale-up MDA efforts for eight priority NTDs . For STH and SCH , Ethiopia mobilized financial resources through a range of partners to support its nationwide baseline surveys . Given the substantial cost of such surveys , there is a need to identify cost-effective mapping approaches to further drive country ownership . The examination of a pooled stool sample strategy ( ten individual samples ) rather than using individual samples is a potential cost-saving strategy . Evidence from veterinary medicine shows that pooling can reduce diagnostic burden and costs , without having a negative impact on estimating the intensity of helminth infections [14] . Pooling has been evaluated for the assessment of STH and SCH in humans [15 , 16] , highlighting that a pooled approach holds promise for rapid assessment of infection , but lacks diagnostic sensitivity . However , previous studies have focused on a small-scale ( number of samples: 116–840 ) , confined geographical area ( Ethiopia: Jimma Town [15 , 16] and Amibara District [17]; Côte d’ Ivoire: Azaguié health district [18] ) , moderate and high transmission areas ( STH prevalence ~50% [15 , 16]; SCH prevalence ~25% [15 , 16] and ~50% [15 , 16] ) , and provided limited information on operational costs . In the current study , we compared an individual and pooled examination strategy for the detection and quantification of soil-transmitted helminth and Schistosoma mansoni infections in 2 , 650 children in 53 primary schools across 35 woredas of Amhara Regional State in Ethiopia . In addition , we compared the time for sample testing for a subset of the samples , and the total operational costs for both strategies .
This study was embedded in the national mapping of STH ( caused by Ascaris lumbricoides , Trichuris trichiura and the hookworms , Necator americanus and Ancylostoma duodenale ) and SCH ( caused by S . mansoni and S . haematobium ) in Ethiopia . The study protocol was reviewed and approved by the Scientific and Ethical Review Office of the Ethiopian Public Health Institute ( ref . no . : SERO-128-4-2005 ) . The Regional Health and Education Bureau were informed about the survey . During the survey , school directors , teachers , and students were informed of its purpose , including its benefits , potential risks and operational procedures . Participation in the study was entirely voluntary . Written informed consent was obtained from the directors of all participating schools , and verbal consent was obtained from all subjects . Subjects who provided a stool sample were given a single dose of mebendazole 500 mg and subjects excreting eggs of S . mansoni or S . haematobium were provided a single dose of praziquantel ( 40 mg/kg of body weight ) . This study was conducted in Amhara Regional State in the North of Ethiopia ( 9° - 14° N and 36° - 40° E ) . Amhara consists of 10 zones and 157 woredas and is divided into three major ecological zones: the highlands ( >2 , 300 m above sea level [asl] ) , midlands ( 1 , 500 to 2 , 300 m asl ) and the lowlands ( <1 , 500 m asl ) . The annual mean temperature is between 15°C and 21°C . The mean annual rainfall is 1 , 165 mm , with the highest rainfall from June to September . The region’s population is 17 . 2 million of which 2 . 4 million ( 14% ) are 10 to 14 years [19] . This study was part of a school-based cross-sectional study to map the distribution of STH and SCH in Amhara Regional State and was conducted from February to March 2015 . Ten schools were randomly selected in each district . From this list , five schools were purposively selected based on ( i ) reports of schistosomiasis , ( ii ) the presence of water bodies close to the schools and ( iii ) practices of irrigation and fishing in the community . Finally , 50 grade-five pupils ( 9 to 14 years; 25 girls and 25 boys ) were randomly selected . In schools with fewer than 25 boys or girls in the appropriate grades , children ( 9 to 14 years ) from lower grades ( grade four ) or higher grade ( grade six ) were included . Each subject was asked to provide a stool sample of approximately 3 g in order to examine samples individually and to subsequently make pools of 10 individual samples . All laboratory procedures were performed in the nearest District Health Facility . At the collection site , the stool samples were immediately stored in a cool box ( at 4°C ) to avoid development of hookworm eggs . On average , samples were kept in cool box for 2hrs prior to processing . At the District Health Facility , samples were processed individually using a single Kato-Katz thick smear , as described elsewhere [20] . Subsequently , stool samples were combined into pools of 10 samples . The procedure of pooling was based on the methodology described by Kure et al . , 2015 [15] and is illustrated in S1 Fig . In summary , 50 individual samples per school were placed in 5 rows of 10 samples . From each individual , 1g of stool was transferred into a new pre-labeled stool cup and thoroughly mixed with a wooden spatula until the color of the mixture became uniform . Finally , the pools were processed applying a single Kato-Katz thick smear . The sensitivity and specificity ( based on faecal egg counts ( FEC ) expressed in eggs per gram of stool ( EPG ) ) of the pooled examination strategy was determined . The sensitivity was calculated using the combined results of both strategies as the diagnostic ‘gold’ standard , against which the sensitivity of the different individual approaches were compared . Therefore , the specificity of both strategies was set at 100% , as indicated by the morphology of the eggs . The sensitivity was determined at the level of the pools and at the level of schools by comparing against the combined strategies . Differences in sensitivity between examination strategies were assessed by a permutation test taking into account the dependency of results within samples ( 10 , 000 iterations ) [21] . The variation in sensitivity of a pooled examination over different levels of egg excretion was explored for each of the four helminths . The classification of the levels of egg excretion were based on the 33th and 66th quantile ( q33 and q66 ) of the mean of the corresponding individual FECs , resulting in 3 levels of egg excretion ( level 1: mean FECs ≤ q33; level 2: q33 < mean FECs ≤ q66; level 3: mean FECs > q66 ) . Differences in sensitivity between levels was assessed by a permutation test taking into account the dependency of results within samples ( 5 , 000 iterations ) . Tukey’s method was applied for multiple comparisons [21] . The agreement between FEC obtained by examining a pooled sample with the mean FEC of the corresponding individual FECs was evaluated by a permutation test ( 5 , 000 iterations ) based on Pearson correlation coefficient and differences in FEC . For the assessment of correlation , FECs of the pooled examination strategy and the mean FECs were log transformed . For the assessment of the difference , no transformation was applied . A comparison of time taken to prepare and examine a subset of the samples ( n = 2 , 450 ) was conducted . The steps were ( i ) the preparation of Kato-Katz thick smears , ( ii ) the pooling of stool , and ( iii ) the examination of the Kato-Katz thick smear . Given that timing of the preparation for each individual Kato-Katz thick smear would slow down the workflow , we recorded the total time to make batches of 10 Kato-Katz thick smears . The preparation of pools and the examination of a Kato-Katz thick smear were recorded on an individual basis . The mean time and corresponding standard deviation was calculated for preparing and reading of the individual and pools across the five survey teams . The total operational cost to map soil-transmitted helminth and S . mansoni infections were estimated for both strategies . The operational costs were assumed to depend on ( i ) human resources utilised , ( ii ) the number of schools that could be screened in one day , and ( iii ) the time during which no activities linked to the survey could be performed . In the present study , five field teams were involved , each consisting of three laboratory technicians and one nurse supported by one vehicle . Our experience from previous surveys suggested that number of schools that can be screened by one team per day ranges from one to three depending on the schools’ accessibility . The teams obtained the permission of the Health and Education Office to operate within each district . Teams were on the road during the entire survey , but were not able to work over weekends as both schools and Health and Education Offices were closed . We estimated the operational cost for one team to be on the road for twelve weeks across different scenarios of school accessibility . We first calculated the cost for one day of work at each school ( e . g . driving to schools , sample collection and processing samples ) , administration ( e . g . obtaining the permission to conduct the survey ) , travel , and days-off ( a day without survey related activities ) across the three levels of school accessibility . These costs included the expenses for materials , salary , transport , and fees to facilitate the work at the schools and data entry . The costs for the materials included equipment , supplies and reagents , based on an itemized cost assessment considering the cost per unit , the usage over a one year period , the life expectancy ( in years ) , and the number of samples that can be processed per day . For simplicity , the number of samples that can be processed per day was fixed at 100 . This assumption implies that the cost per sample will not increase or decrease when fewer or more samples per day are screened . It was estimated that the cost of materials for 50 individuals or pooled samples equaled US$ 4 . For more details on the itemized cost assessment see S1 Table . The daily salary equaled US$ 13 . 7 for each team member . When samples were individually processed , a team consisted of three laboratory technicians and one nurse . When a pooled examination strategy was applied , we assumed two technicians were required rather than three . Data entry clerks were paid US $1 . 1 per 100 data records entered . The daily cost of car rental ( including driver ) was US$ 66 . 0 and the cost for fuel ( 20 L ) was estimated at US$ 15 . 8 . Two schoolteachers were paid US$ 7 . 2 each to support the survey team in informing the students about the survey , facilitating the selection of students , and sample collection . We determined the number of days required for each of the four activities ( work at school , administration , travel , and days off ) within a period of 12 weeks ( 84 days ) for three scenarios of school accessibility . S2 Fig illustrates the activities over a 12-week period when school accessibility was low . In this scenario , there are 45 days of work at school , 21 days off , and 9 days for either travel or administration . S3 and S4 Figs illustrate the activities over a 12-week period when school accessibility is moderate and high , respectively . In these scenarios , the total number of five schools per district remained unchanged , and hence there are days that fewer schools per day are visited . For example , in the scenario of moderate school accessibility , two schools per day will be surveyed in the first two days , followed by one day where only one school is visited . In that case , the daily cost for a poorly accessible school was used . Finally , the number of days and the cost per days were then multiplied for each activity separately , to obtain the total operational cost . We performed a one-way sensitivity analysis in which the cost for materials , salary , fee for school teachers , data entry , rent of car and fuel transport was increased and decreased by 10% . This univariate sensitivity analysis was applied for both individual and pooled examination strategies and the three scenarios of school accessibility , separately .
Eggs of STH or S . mansoni were found in 354 out of the 2 , 650 ( 13 . 4% ) subjects . The predominant helminth species was hookworm ( 6 . 2% , n = 113 ) , followed by A . lumbricoides ( 4 . 3% , n = 164 ) and S . mansoni ( 2 . 6% , n = 70 ) . The least prevalent was T . trichiura ( 0 . 8% , n = 22 ) . The overall mean FEC was 45 . 1 EPG for A . lumbricoides , 17 . 5 EPG for hookworm , 1 . 6 EPG for S . mansoni and 1 . 8 EPG for T . trichiura . At the level of the schools , at least one parasite was found in 41 out of 53 ( 77 . 4% ) schools , and as illustrated in S2 Table , both prevalence and intensity of infections ranged widely across the schools ( T . trichiura: 2 . 0–26 . 0% , 0 . 5–55 . 2 EPG; S . mansoni: 2 . 0–36 . 0% , 0 . 5–35 . 5 EPG; A . lumbricoides: 2 . 0–50 . 0% , 0 . 5–1 , 168 . 8 EPG; and hookworms: 2 . 0–78 . 0% , 0 . 5–535 . 2 EPG ) . Tables 1 and 2 summarize the sensitivity at the level of the pools and the schools , respectively . The sensitivity was set at 100% when at least one parasite egg was detected in a sample or school for both strategies and the sensitivity of both strategies were compared against this target . Generally , the examination of pooled samples resulted in a significantly lower sensitivity for each of the four helminths . At the level of the pools , the sensitivity for a pooled examination strategy ranged from 16 . 7% for T . trichiura to 61 . 8% for hookworms , whereas for an individual examination strategy the sensitivity was at least 90% for all helminths . When determined at the school level , the sensitivity remained roughly unchanged , for a pooled examination strategy it ranged from 22 . 2% for T . trichiura to 75 . 0% for S . mansoni and was significantly lower than the sensitivity of individual examination strategy ( sensitivity >89% ) . As illustrated in Fig 1 , the probability of detecting helminth eggs increased with higher levels of egg excretion for A . lumbricoides ( level 1: 40 . 7% vs . level 2: 42 . 3% vs . level 3: 81 . 5% ) , hookworms ( level 1: 52 . 0% vs . level 2: 50 . 0% vs . level 3: 82 . 6% ) and S . mansoni ( level 1: 30 . 8% vs . level 2: 50 . 0% vs . level 3: 80 . 0% ) . The pair-wise comparison revealed only a significant difference between levels 1 and 3 ( p = 0 . 006 ) , and levels 2 and 3 ( p = 0 . 012 ) for A . lumbricoides . Since there were only 12 cases of T . trichiura , variation in sensitivity across the levels of egg excretion was not assessed . Overall , there was a significant positive correlation between the mean FECs of individual samples and the FECs of the pooled samples for A . lumbricoides ( 0 . 68 , p <0 . 001 ) , hookworm ( 0 . 65 , p <0 . 001 ) , and S . mansoni ( 0 . 75 , p <0 . 001 ) ( Fig 2 ) . Given the low number of cases at the pooled level ( n = 12 ) , the correlation in FECs between examination strategies was not determined for T . trichiura . The mean FEC based on a pooled examination strategy were generally higher than those based on an individual strategy ( Table 3 ) . A significant difference between FECs derived from individual and pooled examination strategies was observed for A . lumbricoides , ( FECindividual = 45 . 1 EPG vs . FECpooled = 93 . 9 EPG , p = 0 . 03 ) and S . mansoni ( FECindividual = 1 . 6 EPG vs . FECpooled = 3 . 4 EPG , p = 0 . 02 ) . The total time to prepare and read 2 , 450 stool samples individually equaled 198h 16min , compared to 53h 50min when samples were processed in pools; a reduction of 72 . 8% . Considerable inter-team variation was observed in time required to prepare and screen samples ( Table 4 ) . The mean time for the preparation of ten individual Kato-Katz thick smears and pools of ten individual stool samples ranged from 10 . 3 min to 31 . 7 min , and from 2 . 4 min to 8 . 8 min , respectively . The reading of the Kato-Katz thick smears ranged from 2 . 2 min to 4 . 4 min for the examination of individuals , and 2 . 7 min to 8 . 2 min for the examination of pools . Across the teams the reduction in time when using a pooled strategy ranged from 50 . 1% to 82 . 0% . The cost per day of the four activities when samples were individually examined and when schools are poorly accessible are summarized in Table 5 . The estimated daily costs were US$ 155 . 6 for work at school , US$ 136 . 6 for a day of administration and a day of travel , and US$ 128 . 7 for a day-off . The differences in daily cost per activity can be explained by differences in usage of material ( only required for work at school ) , payment of fees to school teachers ( only applicable for work at school ) , and the amount of fuel ( less fuel required on days off ) . The cost breakdown for one day of work at school across the three levels of school accessibility when samples are individually examined is summarized in Table 6 . The daily cost increases from US 155 . 6 for poorly accessible schools to US$ 190 . 3 for moderately accessible schools to US$ 225 . 1 for highly accessible schools . Table 5 reports the cost breakdown for the four activities at a poorly accessible school using a pooled strategy . Table 6 reports the costs of work comparing the different levels of accessibility under a pooled strategy . These two tables show that costs are lower compared to an individual examination strategy , and these differences are due to lower usage of materials and less data entry ( one tenth of an individual examination strategy ) . Moreover , since the workload to process samples can be covered by only 2 laboratory technicians , salary costs were reduced . The number of days of work at school , administration , travel and days-off over an 84-day period are summarized in S2 Fig ( poor accessibility of schools ) , 3 ( moderate accessibility of schools ) and 4 ( high accessibility of schools ) . When schools are poorly accessible , a team will spend 45 days ( 53 . 6% ) working in schools and 9 days for administration ( 10 . 3% ) and travel ( 10 . 3% ) , and will have 21 days-off ( 25 . 0% ) . When the schools are moderately and highly accessible the number of days of working in schools equaled 36 ( 42 . 9% ) and 32 ( 38% ) , respectively . The number of days dedicated to administration and travel equaled 12 ( 14 . 3%; moderate school accessibility ) and 16 days ( 19% , high school accessibility ) , the number of days-off equaled 24 ( 28 . 6%; moderate school accessibility ) and 20 ( 23 . 8%; high school accessibility ) . The total operational costs for one team to be on the road for 12 weeks across the different levels of school accessibility are summarized in Table 7 . When samples are processed individually , the estimated operational costs were US$ 12 , 161 . 3 when schools are poorly accessible , US$ 12 , 801 . 0 when schools are moderately accessible and US$ 13 , 590 . 8 when schools are highly accessible . Applying a pooled examination strategy reduced the costs by approximately 11% , regardless of the accessibility of schools . The one-way sensitivity analysis , presented in the Figs 3 and 4 , demonstrates the main cost drivers . Overall , car hire had the largest impact on total costs , followed by salaries ( Fig 3 ) . Varying these parameters resulted in a relative change in operational costs of approximately 4% and 3% , respectively . The impact of the other parameters did not exceed 2% ( fuel: 0 . 7%– 1 . 6%; material: 0 . 0–0 . 2%; data entry: 0 . 0–0 . 0% ) . The impact of car rental costs was consistent across examination strategies and levels of school accessibility . The impact of salary variance was similar for the three levels of accessibility , but was slightly less pronounced for an individual examination strategy ( ~3 . 0% ) compared to a pooled examination strategy ( ~3 . 6% ) . Despite these differences in total operational costs , both parameters had little impact on the cost-savings effect of a pooled examination strategy ( less than 1% reduction; ( Fig 4 ) ) .
SCH and STH programs rely heavily on large-scale surveys to initiate and monitor their success [22] . Initial treatment frequency is determined by the prevalence of infection at baseline , prior to treatment , based on WHO guidelines [1] . Ongoing monitoring and evaluation primarily uses intensity of infection as an indicator for the program’s progress . Under either the individual or pooled approach , large-scale surveys constitute an important cost for governments and funders , which in resource-limited countries present major challenges . Various studies have provided insights on minimizing the operational costs for individual diagnosis of helminths , while ensuring accurate results and subsequently , correct programmatic decisions [23] . To date , the diagnostic strategy of pooling stool to reduce costs has not been fully explored for STH and SCH in humans [15–18] , as studies were based on a small number of samples collected in confined geographical areas where transmission was moderate to high . Therefore , our group tested the applicability of a pooling strategy under field conditions during the national mapping of STH and SCH in Ethiopia . We compared individual and pooled examination strategies for the detection and quantification of STH and intestinal schistosomiasis ( caused by S . mansoni ) at a scale that is unprecedented in the literature and in area were transmission was low . Finally , we compared the time for sample testing , and the total operational costs for both strategies . Overall , our findings on the diagnostic performance are in line with previous small-scale laboratory studies [15–18] , confirming that a pooled strategy provides comparable estimates of population infection intensity , but that it often fails to detect infections , particularly those that are light . At this stage , it remains premature to make any formal recommendations on a pooled approach in a programmatic setting . We evaluated a pooling approach during an early phase of a STH and SCH program ( mapping of disease ) in a low transmission area applying only one diagnostic method ( a single Kato-Katz thick smear ) and one pool size ( 10 individual samples ) . It has been shown that the sensitivity of a pooled examination strategy is a function of the number of individual samples pooled ( sensitivity inversely correlated with the number of individual samples ) and the intrinsic sensitivity of the diagnostic technique [17 , 18] . As a consequence of this , pooling 10 individual samples and testing with a single Kato-Katz thick smear , a technique with poor sensitivity [24] , may not be ideal to assess the intensity and prevalence of infections in all possible scenarios of STH and SCH epidemiology and phases of the program . Complementary studies evaluating pooling of samples in varying scenarios of endemicity , program phase , and diagnostic effort ( number of samples pooled and analytic sensitivity ) are welcomed to inform program managers on when and how to best pool samples . Given that it would be impossible to field test each of these scenarios , one could complement field studies with in silico approaches . Such an approach are best illustrated by the recent study by Lo et al . ( e . g . , reference 18 ) . In this study , field data were used to inform a micro-simulation study . This in silico study was designed to verify whether pooling held promise for drawing programmatic conclusions across scenarios of endemicity other than those observed in the field . For application of a pooled approach in assessing the prevalence of infections , it is also necessary to develop and validate statistical approaches that allow the estimation of the true underlying prevalence based on the results of a pooled examination strategy . A variety of methods have been described for this , and they differ based on how the inference is drawn ( frequentist vs . Bayesian approach ) , assumptions on the diagnostic performance ( perfect vs . imperfect diagnostic techniques ) , number of samples pooled ( fixed number vs . variable number ) and input data ( binary inputs vs . counts ) With a few exceptions , these methodologies were initially developed for diseases other than STH or SCH [25–28] . Our results on the time required for testing demonstrated that a pooling strategy reduced the time to prepare and read slides under field settings by 72 . 8% . A previous study which pooled five individual samples reported a similar reduction in time ( ~70%; Kure et al . , 2015 ) , and this highlights that the reduction in laboratory time is likely not a linear function of the number of samples pooled . Between field teams there was a large variation in reduction in laboratory time between the examination strategies , ranging from 50 . 1% to 82 . 0% . This variation can by explained by a series of factors , such as the experience of technicians , school set-up , and varying issues related to the working environment . In general , a large proportion of the time for sample testing is dedicated to reading slides: reading a slide takes more than half of the total time to test an individual stool sample ( Table 4 ) . Efforts to develop and validate easy-to-use and point-of-care technology that allows electronic imaging of slides , and subsequently automated egg counting should be further encouraged [29] . Our results indicate that , despite a reduction in sample testing time of ~73% , the pooling strategy has relatively little impact on total survey costs ( total operational costs were reduced by ~11% ) . The cost of any survey likely depends on diagnostic technique/s used and survey design [24 , 30–32] . The one-way sensitivity analysis on the different sources of costs revealed little to no variation in the relative cost-savings when a pooled examination strategy was used . Rather , our results indicated that the total operational costs were mainly impacted by logistical factors such as obtaining permission from the district offices and being constrained to the days children are at school . These factors incur additional costs for vehicle rental and survey team salaries , which affected the total operational cost for both strategies . In this regard , our observations indicated that under the different scenarios of school accessibility the teams spend between 36% and 44% of the total days off work when one and three schools are sampled per day . As recently highlighted by Turner and colleagues [33] the cost of a Kato-Katz thick smear varies considerably due to factors such as the method of collection ( processing samples on site vs . examining samples next day off-site ) , the number of sites sampled per day ( increases cost ) , the number of samples collected per site ( decreases cost ) , variation in personnel , and adjustment of microscope costs ( microscope used for other activities vs . microscope exclusively used for the STH/SCH survey ) . Given the number of schools per woreda ( n = 5 ) , subjects per school ( n = 50 ) , the operational steps in the field ( S1 and S2 Figs ) , their corresponding costs ( Table 7 ) and a survey period of 12 weeks , the estimated cost per single Kato-Katz thick smear on an individual stool sample varies from US$ 3 . 4 when 3 schools are surveyed per day ( total number of school children = 4 , 000 ) to US$ 5 . 4 when one school is surveyed per day ( total number of children = 2 , 250 ) . Under the same scenario , the cost for a single Kato-Katz thick smear when a pooled examination strategy is applied increases approximately tenfold ( US$ 48 . 1 when one school is surveyed per day: US$ 30 . 2 when 3 schools are surveyed per day ) , indicating that the way samples are examined ( individual vs . pooled ) should also be considered when costs of the Kato-Katz thick smear are estimated . These differences are explained by the low number of Kato-Katz thick smears ( 1x a single Kato-Katz thick smear is processed from one pooled sample vs . 10x a single Kato-Katz thick smears from 10 individual samples ) and the relatively low reduction in total operational cost when samples are pooled . Finally , the total operational costs were estimated for this specific survey in an Ethiopian setting , and care should be taken when extrapolating to any other national programs . Consequently , it is necessary to compare operational costs of both strategies across a variety of scenarios of national program management to determine whether and when pooling is worthwhile considering . In conclusion , we identified that a pooled strategy provided comparable results for infection intensity , but that it lacks sensitivity and therefore may perform poorly at estimating infection prevalence . A pooled examination strategy resulted in a reduction of 73% of time spent for sample testing , but this only resulted in a reduction of 11% in total operational costs . Based on these findings we conclude that a pooled examination strategy holds some promise for the rapid assessment of intensity of STHs and schistosome infections in a programmatic setting , but that does not result in a major cost-saving opportunity . To make any formal recommendations on a pooled approach , further investigation is required to determine when and how pooling can be utilized . For prevalence-based assessments , such work should also include validation of statistical methods to estimate prevalence based on a pooled examination strategy . Finally , operational costs should be compared different scenarios of national program management . | Infections with intestinal ( roundworms , whipworm and hookworms ) and blood-dwelling ( schistosomes ) worms pose a significant public health burden in developing tropical countries . To optimize control programs against these worms , large-scale surveys are required to determine the worm distribution to initiate control and to monitor the success of the programs . These large-scale surveys come at an important cost for governments , which in resource-limited countries present major challenges . During a nationwide survey in Ethiopia , we assessed whether the examination of pooled rather than individual samples could be a cost-saving strategy to assess prevalence and intensity of worm infections ( which is an indicator of worm-related morbidity and success of the control program ) . We showed that a pooled examination strategy was useful in estimating the intensity of worm infections , but that it underestimated prevalence . Examination of pooled samples significantly reduced laboratory time , but it only resulted in limited financial gain . | [
"Abstract",
"Introduction",
"Methods",
"Results",
"Discussion"
] | [
"schistosoma",
"invertebrates",
"schistosoma",
"mansoni",
"medicine",
"and",
"health",
"sciences",
"salaries",
"education",
"engineering",
"and",
"technology",
"helminths",
"sociology",
"tropical",
"diseases",
"hookworms",
"social",
"sciences",
"parasitic",
"diseases",
"a... | 2018 | Comparison of individual and pooled diagnostic examination strategies during the national mapping of soil-transmitted helminths and Schistosoma mansoni in Ethiopia |
Viewing of ambiguous stimuli can lead to bistable perception alternating between the possible percepts . During continuous presentation of ambiguous stimuli , percept changes occur as single events , whereas during intermittent presentation of ambiguous stimuli , percept changes occur at more or less regular intervals either as single events or bursts . Response patterns can be highly variable and have been reported to show systematic differences between patients with schizophrenia and healthy controls . Existing models of bistable perception often use detailed assumptions and large parameter sets which make parameter estimation challenging . Here we propose a parsimonious stochastic model that provides a link between empirical data analysis of the observed response patterns and detailed models of underlying neuronal processes . Firstly , we use a Hidden Markov Model ( HMM ) for the times between percept changes , which assumes one single state in continuous presentation and a stable and an unstable state in intermittent presentation . The HMM captures the observed differences between patients with schizophrenia and healthy controls , but remains descriptive . Therefore , we secondly propose a hierarchical Brownian model ( HBM ) , which produces similar response patterns but also provides a relation to potential underlying mechanisms . The main idea is that neuronal activity is described as an activity difference between two competing neuronal populations reflected in Brownian motions with drift . This differential activity generates switching between the two conflicting percepts and between stable and unstable states with similar mechanisms on different neuronal levels . With only a small number of parameters , the HBM can be fitted closely to a high variety of response patterns and captures group differences between healthy controls and patients with schizophrenia . At the same time , it provides a link to mechanistic models of bistable perception , linking the group differences to potential underlying mechanisms .
The phenomenon of bistable perception has fascinated researchers for a long time [1 , 2 , 3] . Recently , the description of response patterns to bistable stimuli such as the Necker Cube , Rubin’s vase or rotating spheres with switching rotation direction gained increasing interest in computational neuroscience [4 , 5 , 6 , 7 , 8] . By modeling dynamic changes of perception during viewing of one and the same stimulus , one aims at providing potential explanations for neuronal mechanisms underlying perception and perceptual changes and to identify related brain areas as well as potential dysfunctions , e . g . in schizophrenia [9 , 10] . Interestingly , the response patterns to continuously shown bistable stimuli often share common properties [7 , 11] . Typically , the distribution of intervals of constant perception , termed dominance times , is unimodal and right-skewed , and extremely short dominance times , i . e . , rapidly fluctuating precepts , are rare [12 , 13 , 14] . The dominance times under continuous stimulation are therefore often modeled as Gamma distributed [15 , 16 , 17 , 18 , 19 , 20] . The mean of dominance times can be highly variable across subjects [14 , 20] , whereas the coefficient of variation ( CV ) is often comparable [21] . In comparison to a continuous presentation , intermittent presentation of a bistable stimulus , i . e . , by repetitive interruption of stimulation for short time periods , has been observed to stabilize the percept if the interruption period is long enough , typically longer than 0 . 7 seconds [15 , 18 , 22 , 23 , 24 , 25] . In this case , dominance times get longer and can also show a certain degree of periodicity [14] . In addition , such stable phases with long dominance times during intermittent presentation can also interchange with unstable phases of rapid percept changes . Fig 1 shows examples of response patterns to continuous and intermittent presentation of a bistable stimulus from the dataset reported in [10] . Modeling studies with elaborated mathematical models have been proposed that can explain a number of properties of bistable perception like the distribution of dominance times under continuous stimulation [8 , 17 , 19 , 20 , 21 , 26] or cyclic behavior and the impact of the duration of the stimulus presentation on the dominance times in intermittent stimulation [14 , 18] . One key ingredient of these models of bistable perception is typically a competition between neuronal populations that correspond to the different percepts [14 , 17 , 18 , 20 , 27] . In order to account for stabilized perception in intermittent viewing , the use of multiple timescales for memory traces of past perception has been proposed by [14] and [18] . Many such models require a high number of parameters in order to describe the variety of response patterns . As a consequence , they can often hardly be fitted to experimental data , especially in the typical cases when only a few dozen dominance times are observed . In addition , the majority of models focus either on continuous or on intermittent viewing . Interesting models that are applicable to both cases have been proposed by [14 , 17 , 18] . The relevance of a joint description of continuous and intermittent viewing is illustrated here on a dataset including responses of patients with schizophrenia and of healthy controls to continuous and intermittent presentation of a rotating sphere with ambiguous rotation direction reported earlier in [9 , 10] . In [10] , an enhanced alternation rate for the group of patients with schizophrenia during intermittent presentation was reported . Interestingly , when we analyzed previously unpublished data recorded in the same participants during continuous presentation , the opposite could be observed [Fig 2; the data was collected during an initial training run for which the experimental procedures but not the results are described in 10] . Due to the differences in patterns and time scales between continuous and intermittent presentation , the potential neuronal mechanisms underlying the transitions between the different response properties remain unclear . Therefore , we propose here a new model for the description of response patterns to bistable perception that links the observed behavior in continuous and intermittent stimulation to potential underlying neuronal processes . First , the model should be able to describe the high variety of both , continuous and intermittent stimulation within one model framework . Second , we use a minimal number of parameters in order to allow parameter estimation and model fitting to the typically short experimental data . This can then allow the statistical investigation of differences between clinical groups . Note that strictly speaking , the term ‘dominance time’ refers to slightly different objects in continuous and intermittent viewing . While during continuous presentation , switches occur from a dominant to a suppressed percept ( percept-switch ) , dominance times during intermittent presentation consist of multiple continuous presentation periods , and switches typically occur because of different perceptual choices at the onset of the presentation ( percept-choice ) [28] . In the present model , the observed sequences of dominance times are treated as conceptually similar . This simplification allows for a parsimonious model description in both continuous and intermittent viewing but may not fully capture the relation between the perceptual processes in the two regimes . The remainder of the article is organized as follows . First , we use a simple Hidden Markov Model ( HMM ) that describes the observed perceptual processes with a few parameters . For continuous presentation , one state produces independent and identically distributed dominance times with a two-parametric distribution . For intermittent presentation , switching between stable and unstable phases requires two hidden states with short and long dominance times , respectively . The HMM has the advantage that it allows straightforward model fitting and data description with a minimal number of parameters . However , it remains descriptive and lacks relations to potential underlying mechanisms . Therefore , we link the HMM to a hypothetical underlying stochastic model . This model is termed here Hierarchical Brownian Model ( HBM ) and intends to describe aggregated underlying neuronal activity , producing the observed behavioral responses . The HBM is based on two main ideas: First , it assumes that switching between percepts results from two conflicting neuronal populations [cmp . , e . g . , 18] . In order to minimize the number of parameters , this process is reduced to a simple Brownian motion with drift that fluctuates between two thresholds , where the first passage times indicate state changes [similar to 21] . For continuous presentation , one therefore requires only two parameters , i . e . , the drift of the Brownian motion and the threshold . The distribution of the resulting first passage times—i . e . , dominance times—is then the same as in the HMM , with a simple relation between the two HBM and the two HMM parameters . Second , in order to describe intermittent presentation in the same model framework , we use a hierarchical model . The idea is to describe the switching between stable and unstable phases that is typical for intermittent presentation by using an analogous threshold crossing mechanism of conflicting neuronal populations . Specifically , we assume a second pair of neuronal populations whose corresponding Brownian motion modulates the drift and threshold of the first population pair and thus causes switching between stable and unstable phases . We give a set of model assumptions under which the HBM parameters are comparable to the HMM parameters , thus allowing both model fitting to experimental data sets and potential relation to underlying mechanisms . The parameter estimation is straightforward using maximum likelihood and the HBM can reproduce both , the unimodal distribution in the continuous presentation and the bimodal distribution of dominance times in the intermittent presentation , including also various different response patterns . Moreover , it allows the identification of specific differences between the clinical groups in [10] and relates these to the hypothesized underlying processes .
As a first step to describe the processes observed in bistable perception , we reduce data analysis to the dominance times , i . e . , the times between reported changes of the percept . As described above , the distribution of dominance times tends to be unimodal in the continuous case , while stable and unstable phases interchange in intermittent stimulation . We denote the dominance times by di , i = 1 , 2 , … , n . In the unimodal continuous case , we assume the di to be the realizations of independent and identically ( i . i . d . ) distributed random variables Di , i = 1 , 2 , … , n ( Fig 3A ) , where the Gamma or the Inverse Gaussian ( IG ) distribution are suitable two-parametric distributions [17 , 18 , 19 , 21] . For comparability with the HBM , we focus on the Inverse Gaussian distribution here . For the intermittent case , we assume a HMM with a stable and an unstable state , which are hidden and produce long and short dominance times , respectively ( Fig 3B ) . This requires two parameters for the switches between states , and two parameters for the distribution of dominance times in each state . Formally , let Y ≔ ( Yi ) i = 1 , … , n describe a Markov chain on {S , U} , where S and U denote the stable and unstable state , respectively . Let pSU = 1 − pSS and pUS = 1 − pUU denote the transition probabilities . The dominance times ( di ) i∈{1 , 2 , … , n} are assumed to be Inverse Gaussian distributed and conditionally independent given Y , with mean and standard deviation given by ( μS , σS ) for Yi = S and ( μU , σU ) for Yi = U . Note that independence of dominance times is assumed here in continuous presentation . This assumption enables straightforward parameter estimation and is in agreement with the observation that serial correlations of dominance times are typically not reported [e . g . , 29 , 30] . However , weak long-term dependencies of dominance times reported under continuous presentation [31] cannot be reproduced in the HMM . As such long-term dependence was not observed in the majority of cases in the present data set , also showing no group differences , we use here the simple assumption of independence . In addition , the two used HMM parameters are sufficient to capture the main group difference in the response properties reflected in the alternation rate . As described in the previous section , the HMM captures a high variety of response patterns both in continuous and intermittent viewing , including uni- and bimodal distributions of dominance times with alternations between stable and unstable states and a high variability across subjects . With its small number of parameters , the HMM can be fitted also to short data sections available empirically and therefore also capture differences between experimental groups . However , the HMM description remains phenomenological and does not provide insight into potential neuronal processes . Also , it cannot provide explanations for potential effects that different lengths of blank displays could have on the response patterns , as discussed for example by [14 , 22 , 24 , 25] . In addition , the HMM cannot represent the following interesting empirical observation: Before changing from stable to unstable state , the last dominance time tends to be shorter . Therefore , we introduce here a new model , called Hierarchical Brownian Model ( HBM ) , which provides a potential link between the phenomonological description of the response and potential underlying neuronal processes . The HBM assumptions can also provide hypotheses on the effects of different lengths of blank displays and naturally yields shorter dominance times before a state change to the unstable state . The HBM assumes two competing neuronal populations which indicate perception of right and left rotation , respectively . As has been proposed by various authors [14 , 18] , we implicitly assume mechanisms of self-excitation , cross-inhibition and adaptation across these neuronal populations , without explicitly modeling them in order to reduce the number of parameters and to allow for model fitting to short trials . In order to obtain a parsimonious model description , we again assume independence of dominance times by neglecting potential mechanisms of week long-term adaptation [31] . For possible model extensions compare section ‘Applicability and model extensions’ in the discussion . We use the simplified assumption that perception arises from the difference in the activity of the two populations , which is modeled here by a Brownian motion with drift [similar to 21] that fluctuates between two thresholds , where the first passage times indicate state changes . This results in two parameters for the case of continuous presentation that are directly linked to the two parametric distribution of dominance times in the HMM . Further , we describe switching between stable and unstable states in intermittent presentation by applying an analogous mechanism , which leads to a hierarchical model . We assume another hierarchical layer of neuronal populations and a corresponding Brownian motion which modulates the drift of the first population pair and thus causes switching between stable and unstable phases .
In the present article we have proposed a model framework for the description and analysis of perceptual responses to bistable stimuli . In particular , the first goal was to describe a high number of observed patterns in responses to continuous and intermittent stimulation and their differences between a group of patients with schizophrenia and healthy controls . The variety of patterns includes more or less regular dominance times during continuous stimulation and a switching between long and short dominance times , i . e . , stable and unstable states , during intermittent stimulation , with a tendency for periodically occurring percept changes . We started on a descriptive level , assuming that dominance times were generated by a simple HMM with only one state for continuous presentation and a stable and an unstable state in intermittent presentation . The HMM was sufficiently small to allow model fit to short empirical data sets and could also describe the high variety of empirically observed response patterns in continuous and intermittent presentation . Interestingly , it also revealed a high degree of reproducibility of response patterns of the same subject across different sessions . In addition , it allowed to relate observed group differences in the rate of percept alternations to HMM parameters , suggesting that especially the relative time spent in the stable state was reduced in the patients with schizophrenia . Our second goal was to relate the observed response patterns and group differences to potential underlying mechanisms and thus , to build a link to models with detailed neurophysiological assumptions [7 , 13 , 14 , 18 , 20] that may not include all types or response patterns and/or may not allow fitting to short data sets . To that end , we proposed a hierarchical model of interacting Brownian motions ( HBM ) . The HBM is based on the common assumption that the sequence of percept changes results from a competition of conflicting neuronal populations [14 , 17 , 18 , 20 , 27] . Instead of modeling these in detail , we describe the activity difference by a Brownian motion P with drift ν0 [21] between two borders ±b , where the first hitting times of the borders indicate percept changes . Roughly speaking , the drift ν0 could be considered related to the neuronal interactions within and between the populations , while the border b could be considered related to the population sizes . In order to describe responses to intermittent presentation , this mechanism is adapted in another population pair . These populations exhibit a corresponding background process B that evokes switching between stable and unstable states , similar to switching between the two percepts . In particular , B causes the perception process P to change parameters from small drift νS and large border bS in the stable state to fast drift νU and small border bU in the unstable state . The HBM could be fitted nicely to the given empirical data set , reproducing a high variety of response patterns to continuous and intermittent stimulation in healthy subjects and patients with schizophrenia . In particular , the model fit was even improved over the descriptive HMM by reproducing shorter stable dominance times before a change to the unstable state . The HBM also provided more detailed explanations for the observed group difference that patients with schizophrenia showed higher alternation rates during intermittent stimulation , while percept alternation was decreases during continuous presentation . In particular , the HBM contains additional mechanisms of switching between stable and unstable state for intermittent presentation , which is assumed inactive during continuous presentation . The HBM , similar to the HMM , suggests an increased probability of switching to the unstable state for the patients with schizophrenia and thus , a longer relative time spent in the unstable state . The HBM also provides additional potential explanations related to the borders , or assumed population sizes , suggesting a higher increase from continuous ( border b ) to stable intermittent presentation ( border bS ) in the healthy subjects . This is a first finding on the transition from continuous to intermittent presentation , which results from including both continuous and intermittent presentation in one model . These findings suggested by the HBM , which include a longer relative time spent in the unstable state for the patients with schizophrenia and a smaller population size involved in percept stabilization , are also in agreement with recent findings of [41] . They studied the learning behavior of healthy subjects of whom the degree of delusional ideation [42] had been measured . In compliance with earlier studies [for a review see 43] , they reported that subjects with larger delusion proneness made decisions on the basis of less information and were also less resilient against irrelevant information [compare also the literature about jumping to conclusions , e . g . 44 , 45] . In the present setting , the ambiguous stimulus represents a constant source of partly contradicting visual information [see also 5 , 8] . In that sense , the unstable state could be considered a state in which one is less resilient against this contradicting visual information , which yields a high rate of percept changes . The fact that the patients with schizophrenia spent more time in the unstable state is therefore highly consistent with the findings of [41] . Moreover , this finding is also compatible with current models of schizophrenia in the framework of predictive coding [46] that propose a reduced top-down influence of stored predictions . However , it goes beyond previous work by highlighting the role of a background process that controls the balance between stable and unstable states in perceptual inference . In addition , the population sizes could be considered related to the amount of information taken into consideration to create a percept . Again , consistently with [41] , we find , in the stable state , larger estimated population sizes , bS , of the perceptual populations L and R in healthy controls than in patients with schizophrenia . Also , these population sizes are typically much larger than the population sizes in the unstable state ( bS >> bU ) , which would be consistent with the notion that subjects in the unstable state need less information to change their perception . The HBM may also be used to describe dominance times resulting from other experiments with ambiguous visual stimuli studying , e . g . , motion-induced-blindness , binocular rivalry , moving plaids , the Necker Cube , orthogonal gratings or the house/face-paradoxon [e . g . 21 , 47] or also bistable auditory stimuli [48] . The HBM is , however , not designed for tristable stimuli , and transient stimulus manipulations as used in after-effect studies cannot be captured by the HBM in its current form . In different bistable settings , the HBM cannot be applied directly , but would allow for potential extensions . For example , in its present form , the HBM describes only balanced perception . However , it could be extended with respect to unbalanced bistable displays , e . g . , for different eye contrasts during binocular rivalry [49] , by choosing different drift parameters for the positive and the negative drift direction during presentation . Similarly , the drift could be chosen to vary as a function of attention [50 , 51] or as a function of long-term history ( e . g . , the cumulative history H proposed in [31] ) . In studies on mixed perception during binocular rivalry [19] , one might use an additional border to define an intermediate range for the perception process in which mixed perception is described . One should note that the HBMi in its current form is restricted to a duration of blank displays lb ≤ lp ⋅ ν0/νS . For longer blank displays , the mean drift of P during stable states , ν S * , will be negative , yielding no perception change with high probability . However , it would be possible to extend the model accordingly , assuming a temporal evolution in the drift parameters , given corresponding extended empirical observations . In addition , note that the border of the perception process is assumed to be b during continuous stimulation and bS ( or bU ) during intermittent stimulation . Therefore , an instantaneous change of intermittent to continuous presentation is not yet described . Here , we qualitatively assume that the border jumps very fast from b to bS with the onset of a blank display , while going back slowly during stimulation . A transition from continuous to intermittent presentation would therefore instantly change the response pattern , while a reverse transition would gradually reverse the change back to the one-state process . Quantitative validation and fitting of this assumption would be interesting , but requires corresponding empirical observations , in which the length of the presentation period lp is varied . This would also allow investigation of potential relations between the HBMc and HBMi parameters and thus , between the mechanisms assumed to underlie the identified group differences . Concerning the impact of the duration of the blank display lb , two aspects should be discussed . First , the HBM can theoretically reproduce a phenomenon reported earlier in [14] . Conditional that one percept has been present for a short while , the probability of a percept change rises with the blank duration lb . In the HBMi , the same is observed during the unstable state with typically short dominance times: During the unstable state the drift in the blank displays , νU , is typically larger than the drift ν0 during stimulation . Therefore , longer blank displays speed up P , thereby reducing perceptual stability . Second , one interesting potential model extension is concerned with the relationship between the length of the blank display and the alternation rate . As reported earlier by [14 , 15 , 18] , the mean dominance time in intermittent presentation has been found to be a function of the relationship between the presentation length lp ( or ‘ON’-period ) and the length of the blank display lb ( or ‘OFF’-period ) . Particularly , the dependence between lb and the alternation rate is non-monotonic , as would be implied in the HBMi , but follows an inverted U-shape [22 , 24 , 25] with a peak roughly at 0 . 4 s . Such an inverted U-shape would be possible in a model extension of the HBMi . As discussed in the Results , the drift terms νS , νU only represent the mean drift across the period of blank display , which is sufficient and parsimonious in the given data set with fixed length of blank display . However , the model would be fully consistent with the assumption that the drifts change during the ‘OFF’-period , such that the mean drifts νS ( lb ) and νU ( lb ) are functions of the length of the blank display lb . In Fig 19A these mean drifts νS , νU decrease with lb , where the stronger drifts at the beginning of the blank display could be effects of the recent stimulation . Panel B shows the resulting mean alternation rate , which has an inverted U-shape with a maximum around 0 . 4 s and shows increased stability under intermittent stimulation for lb > 0 . 7 . This increased stability is caused first by a small drift νS < ν0 in that range . Second , it is also caused by the fact that the time interval lb in which the background process B has positive drift is longer , leading to an increased probability to reach b ˜ S and thus , to stay in the stable state . Estimation of the functions νS ( lb ) and νU ( lb ) from a suitable data set with variable lengths of blank displays would be an interesting task . In summary , the proposed HBM intends to provide a link between empirical data analysis and mechanistic modeling . On the one hand , it aims at precisely describing the high variety of response patterns observed in perceptual responses to bistable stimuli . On the other hand , it aims at bridging the gap to detailed mechanistic models of bistable perception , allowing assumed processes to be fitted to short empirical data sets and thus , also the analysis of group differences . Various extension possibilities show a potential of the HBM to investigate related experimental contexts . By including both continuous and intermittent stimulation , the HBM can thus provide interesting new hypotheses on potential neuronal mechanisms of cognitive phenomena .
Here we describe the estimation procedures of the HMM parameters for continuous presentation and for intermittent presentation . We denote by d ≔ ( d1 , d2 , … , dn ) the set of dominance times modeled as realizations of random variables D = ( D1 , … , Dn ) . Here we describe the estimation procedures of the HBM parameters for continuous presentation and for intermittent presentation . Here , we derive a formula for the alternation rate ρ used in Fig 19 given the HBMi parameter set ( μ S * , σ S * , μ U * , σ U * , p S S * , p U U * ) . For each length of blank display lb and the value of νS , νU as shown in Fig 19A the mean drifts per second in the stable and the unstable state ν S * , ν U * and the mean drift of the background process ν B * are derived using Eq ( 4 ) . Then , we use Eqs ( 8 ) and ( 11 ) to derive the values ( μ S * , σ S * , μ U * , σ U * , p S S * , p U U * ) given the mean drifts per second , and we recall Eq ( 17 ) for φ S * , and analogously for φ U * We now show that the alternation rate can be described as ρ≔lim Δ → ∞ E [ N * ( Δ ) ] Δ = φ S * μ S * + φ U * μ U * , ( 18 ) where N* ( Δ ) denotes the number of perceptual changes in an interval of length Δ . We split up N * ( Δ ) = N S * ( Δ ) + N U * ( Δ ) , where N S * ( Δ ) and N U * ( Δ ) denote the number of perceptual changes in the respective stable and unstable phases in the interval of length Δ . We then show E [ N S * ( Δ ) ] Δ → Δ → ∞ φ S * μ S * , and analogously for the unstable state . To that end , let ΔS be the time spent in the stable state in a time interval of length Δ . By the Elementary Renewal Theorem [e . g . 56] it holds E [ N S * ( Δ ) ] Δ S = E [ N S * ( Δ S ) ] Δ S → Δ → ∞ 1 μ S * as Δ → ∞ naturally implies ΔS → ∞ . According to the definition of φ S * as the expected relative time spent in the stable state , we get Δ S / Δ → φ S * in probability . This yields the claim E [ N S * ( Δ ) ] Δ = Δ S Δ · E [ N S * ( Δ ) ] Δ S → Δ → ∞ φ S * μ S * . | Patients suffering from schizophrenia show specific cognitive deficits . One way to study these cognitive phenomena works with the presentation of ambiguous stimuli . During viewing of an ambiguous stimulus , perception alters spontaneously between different percepts . Percept changes occur as single events during continuous presentation , whereas during intermittent presentation , percept changes occur at regular intervals either as single events or bursts . Here we investigate perceptual responses to continuous and intermittent stimulation in healthy control subjects and patients with schizophrenia . Interestingly , the response patterns can be highly variable but show systematic group differences . We propose a model that connects these perceptual responses to underlying neuronal processes . The model mainly describes the activity difference between competing neuronal populations on different neuronal levels . In a hierarchical manner , the differential activity generates switching between the conflicting percepts as well as between states of higher and lower perceptual stability . By fitting the model directly to empirical responses , a high variety of patterns can be reproduced , and group differences between healthy controls and patients with schizophrenia can be captured . This helps to link the observed group differences to potential neuronal mechanisms , suggesting that patients with schizophrenia tend to spend more time in neuronal states of lower perceptual stability . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Materials",
"and",
"methods"
] | [
"medicine",
"and",
"health",
"sciences",
"classical",
"mechanics",
"markov",
"models",
"fluid",
"mechanics",
"social",
"sciences",
"electrophysiology",
"random",
"variables",
"neuroscience",
"perception",
"cognitive",
"psychology",
"mathematics",
"probability",
"distributio... | 2017 | A hierarchical stochastic model for bistable perception |
The basis for correctly assessing the burden of parasitic infections and the effects of interventions relies on a somewhat shaky foundation as long as we do not know how reliable the reported laboratory findings are . Thus virtual microscopy , successfully introduced as a histopathology tool , has been adapted for medical parasitology . Specimens containing parasites in tissues , stools , and blood have been digitized and made accessible as a “webmicroscope for parasitology” ( WMP ) on the Internet ( http://www . webmicroscope . net/parasitology ) . These digitized specimens can be viewed ( “navigated” both in the x-axis and the y-axis ) at the desired magnification by an unrestricted number of individuals simultaneously . For virtual microscopy of specimens containing stool parasites , it was necessary to develop the technique further in order to enable navigation in the z plane ( i . e . , “focusing” ) . Specimens were therefore scanned and photographed in two or more focal planes . The resulting digitized specimens consist of stacks of laterally “stiched” individual images covering the entire area of the sample photographed at high magnification . The digitized image information ( ∼10 GB uncompressed data per specimen ) is accessible at data transfer speeds from 2 to 10 Mb/s via a network of five image servers located in different parts of Europe . Image streaming and rapid data transfer to an ordinary personal computer makes web-based virtual microscopy similar to conventional microscopy . The potential of this novel technique in the field of medical parasitology to share identical parasitological specimens means that we can provide a “gold standard” , which can overcome several problems encountered in quality control of diagnostic parasitology . Thus , the WMP may have an impact on the reliability of data , which constitute the basis for our understanding of the vast problem of neglected tropical diseases . The WMP can be used also in the absence of a fast Internet communication . An ordinary PC , or even a laptop , may function as a local image server , e . g . , in health centers in tropical endemic areas .
The Internet has made possible high standard educational undertakings with microscopy images also in the field of diagnostic medical parasitology ( for an example , see: www . parasite-diagnosis . ch ) . However , the limitation until now has been that presentation of selected illustrations cannot replace working with a real microscope . The success of web-based virtual microscopy for histopatology [1] ( www . webmicroscope . net ) at the outset prompted us to demonstrate the histopathology of “the schistosome-infected mouse” which is presented in the beginning of this study . It soon became evident that some serious obstacles associated with education and quality control in medical parasitology can be solved using web-based microscopy , the main topic of this study . Diagnostic parasitology , essentially being equivalent to microscopical examination of stool and blood samples , is performed globally at the basic level of the health care-system . Despite the recent introduction of polymerase chain reaction ( PCR ) -based methods , which make possible parasite identification also in cases of morphological identity , the methodology has changed little during the 150 years elapsed since it was described by Davaine [2] , [3] . Fresh stool samples are studied under the microscope either as such or after the addition of Lugols solution to increase the sensitivity—and specificity—of the method . Additional concentration and staining procedures can be employed , but such procedures are usually performed at the next level , in parasitological laboratories associated with hospitals or microbiology departments of universities . Whereas routine diagnostic methods in microbiology usually depend largely on cultivation under a variety of defined conditions under which microbial growth is quantified , diagnostic parasitology is equivalent to visual identification of parasites and/or parasite-derived materials . Thus , the quality of medical parasitology at the basic level relies heavily on the individual microscopist . Several good atlases describing medically important parasites have been published . The Training Manual on Diagnosis of Intestinal Parasites based on the World Health Organization ( WHO ) Bench Aids for the Diagnosis of Intestinal Parasites [4] has had a fundamental impact by providing a reference for the morphological identification of human parasites . Also ambitious external quality assessment programs have had a definitive effect , as shown by United Kingdom National External Quality Assessment Scheme ( UKNEQAS ) with a reported scheme of eight distributions a year to 285 participants [5] . However , several problems like differentiating between parasites and non-parasite material are difficult to solve . Thus , there is a need for developing quality assessment and education in parasitological diagnostics both in endemic and in non-endemic areas , based on defined samples containing helminth eggs and protozoa as well as “parasite-like” material , which may give false positive results . Besides distribution problems and challenges related to limited supply and costs involved , variation seriously restricts the widespread use of current quality assurance programs . Sample variation may depend on transportation and handling but the major problem is uneven distribution of parasites in the samples even if obtained from the same single source . To obtain defined and identical samples for large-scale distribution to several parasitology laboratories appears to be an unrealistic goal . In the present report , by developing the Web Microscope for Parasitology ( WMP ) , we have explored the potential of virtual microscopy to reach that goal . To identify parasites in suspensions it was necessary to develop a technique that facilitates navigation in the z-plane for “focusing” . In order to evaluate the potential of virtual microscopy to provide reference material for immunofluorescence ( IF ) microscopy , slides showing the staining patterns of specific antischistosome antibodies were also included . This was primarily motivated by the variation in published illustrations of the diagnostic schistosome gut reactive immune response seen in early infection [6] .
Paraffin sections of tissues from a Schistosoma mansoni-infected mouse [6] were stained with hematoxylin and eosin using standard protocols . Stool samples fixed in Sodium Acetate Acetic Acid Formalin ( SAF ) fixative , Giemsa-stained thick and thin blood films containing malaria parasites and a smear of a leishmaniasis skin lesion were from specimens sent for microscopical examination to the parasitology laboratory at the Swedish Institute for Infectious Disease Control ( SMI; Stockholm , Sweden ) and the Microbiology Department of the National University of Léon ( UNAN; Léon , Nicaragua ) . Lugols iodine solution was used to enhance the microscopical features of protozoan cysts . Microscope slides were mounted under a cover slip for photography . For slides containing pools of several different stool parasites glycerol-gelatin was used as water-soluble mounting medium as described for insect specimens e . g . by Christie ( www . psych . ubc . ca/bchristie/Techniques/PVA . htm ) or Schauff ( www . ars . usda . gov/SP2UserFiles/ad_hoc/12754100CollectingandPreservingInsectsandMites/collpres . pdf ) . For IF microscopy , we used sera from patients suffering from recent or chronic schistosomiasis and shown to exhibit typical serum antibody reactivity with the parasite . Frozen sections of S . mansoni adult worm pairs used for routine diagnostics were obtained by perfusion of infected mice as described previously [6] . Paraffin sections of worms fixed in Bouin's fixative gave essentially similar staining results , but were preferred as they were more stable . Specimens were sent to the University of Tampere ( Tampere , Finland ) for digitization . The entire specimens on a microscope slide were scanned and photographed at high magnification ( usually with a 40× objective ) , using a motorized microscope board . Sequentially acquired image tiles were stitched together , generating large image files in the order of 10–50 GB [1] . After image compression , the generated ‘virtual slides’ were uploaded to a network of image servers ( www . webmicroscope . net/WMNetwork ) for web-based viewing [1] . The ‘virtual slides’ can be viewed from personal computers with a web browser ( MS Internet Explorer or Mozilla FireFox ) , allowing free navigation in the x and y planes at the desired magnification . Of note , in order to resemble real microscopy , a moderately fast Internet connection is required ( e . g . , 1–2 Mb/s and higher ) . Regions of special interest ( ROI ) were indicated and some annotations added . The possibility to determine the size of an object is helpful especially in determining the size of intestinal protozoan cysts , but also to distinguish between parasite material and artefacts , such as pollen . Pictures accompanying the text are screen images on the monitor seen at the Internet page ( www . webmicroscope . net/parasitology ) . Images in the text were mounted using the Adobe Photoshop program .
Ten specimens on microscope slides were digitized . They are seen in Figure 1 as “thumbnails” corresponding to part of the established website for medical parasitology ( www . webmicroscope . net/parasitology ) . In line with the current use of virtual microscopy for histopathology , we have digitized tissue samples obtained from a mouse experimentally infected with S . mansoni . We have digitized specimens containing some commonly occurring parasites . As “proof of principle” we show the appearance of malaria and leishmania parasites ( Figure 2 ) and some helminth eggs and intestinal protozoan cysts in stool samples . For the latter specimens and to mimic actual microscopy , we developed a technique making “focusing” possible . A further example of the use of virtual microscopy is to provide reference material for unusual , non-permanent or demanding parasitological assays . We show as an example , the diagnostics of acute schistosomiasis , which may depend on the demonstration of serum antibodies directed against excretory products of the intravascular worms . The IF samples showing serum antibody staining patterns typical for acute and chronic schistosomiasis , respectively on a number of sections of adult male and female schistosomes is seen in Figure 3 . The stool sample ( “stool 1” , Figure 4 ) contains both helminth eggs and intestinal protozoan cysts , which can be examined at various magnifications: Trichuris trichiura ( Figure 4B ) and an Entamoeba cyst ( Figure 4C ) . Fading of the iodine staining was evident in case of intestinal protozoa , whereas the stained appearance of T . trichiura eggs is due to endogenous material . The effect of increasing the number of layers is demonstrated . We can see that eggs in a stool sample scanned in three focal planes gives the illusion of focusing . However , the image obtained with more layers is superior , e . g . , in examining the different structures of a Taenia spp . egg at various focal levels ( Figures 4D–4G ) . Focusing of a S . mansoni egg shows the typical lateral spine ( Figures 4H and 4I ) . We observed that the “illusion of focusing” corresponds surprisingly well with the focusing in a real microscope . It also turned out that stool material covering a Taenia spp . egg does not interfere substantially with the resolution obtained in underlying structures . By adding up to 10 layers , the focusing capacity is increased further as expected , but data handling becomes somewhat slower . The use of several layers made it possible to readily identify protozoan cysts which are dispersed in the 3-dimensional space of the specimen . The resolution using the 40× objective of our current microscope equipment is very good for helminth eggs and the focusing function obtained by scanning the specimen in two or more focal planes gives the authentic feeling of “focusing” like in an ordinary microscope . For intestinal protozoan cysts the resolution appears somewhat weaker , but still acceptable . The identification of , for example , some Entamoeba spp . cysts is possible in the specimen where the number of nuclei and their chromatin distribution is obvious .
Introduction of virtual microscopy of entire parasitological specimens as a new tool brings “virtual microscopy” very close to working with a real microscope . The author has lost the privilege of selecting “typical” microscopy fields for publication . Artefacts and non-parasite material , which can be mistaken for helminth eggs and intestinal protozoan cysts , cannot be ignored in digitized whole specimens . By presenting digitized whole parasitological specimens of different types at our website , we have illustrated these points . Web microscopy for education and quality control in medical parasitology overcomes several problems basically due to the fact that parasites present in patient specimens are randomly distributed especially in suspensions . Thus , it is impossible to produce identical stool samples that are , for example , utilized to assess inter-laboratory agreement or for quality control studies ( for example see [7] ) . As a result , quality assessment programs invariably lead to discussions on false negative and false positive results obtained by the participating centres . By using web-based virtual microscopy , we can distribute an identical sample and make it accessible worldwide via a network of image servers and discussions can be focused on observations on defined structures , which are traceable . We conjecture that the basis for correctly assessing the disease burden and for measuring the effects of interventions relies on a somewhat shaky foundation as long as we do not know how reliable the reported laboratory findings are . The problem may be considerable in many developing countries . We simply do not know the fundamentals of false positive and false negative diagnostic findings , as we have seen in the case of “amebiasis” at health centers in Nicaragua [8] . Thus , the WMP may have an impact on the reliability of data , which constitute the basis for our understanding of the vast problem of neglected tropical diseases , a problem we have only started to explore [9] . By programming the motorized microscope stage to move also in the z-direction during digitization , 3–20 different focal planes were acquired for the specimens containing stool parasites . In the browser viewer the virtual slides corresponding to the focal planes can be stacked on top of each other in layers , ordered by their z-depth . By using a variable transparency when switching from one focal layer to another , an effect simulating focusing with a real microscope is created in the web browser . If two focal layers are used in the browser , the amount of image data that needs to be loaded from the server is doubled . Addition of several layers would quickly saturate most network connections . We therefore developed a novel smart-focusing technique , which uses only a single focal layer when zooming or navigating in the x and y directions . However , immediately as the user moves in the z-direction ( ‘focuses’ ) , either by ‘dragging’ a focusing slider on the web page , or pressing the focusing keys on the keyboard ( Q and A ) , all available layers are turned on and begin loading data . Even if there are multiple layers , the required image data will load quickly as only the area of the specimen currently visible to the user needs to be refreshed . When the user again moves in the x or y direction or changes zoom-level , all but the currently visible layer are turned off , which enables faster zooming and panning . With the smart-focusing method there is technically no limit on how many focal layers that can be added and stacked in the same browser window . In our tests we have found that up to 10 focal planes can be used , provided that the network connection is good ( more than 2 Mb/s ) , and that the computer from which the slides are being viewed has a capacity of at least 2 GHz or a dual core processor . With ordinary office computers , connection speeds and monitor sizes , a maximum of 5 layers currently seems to be recommendable . Several factors may influence the speed and smoothness of viewing virtual slides over a computer network . These include the bandwidth of the Internet connection , the end-user and server computer speed , as well as the size of the view area on the user's computer screen . The actual size of a virtual slide residing on the server does , however , not affect the viewing speed , since only the area currently visible on the user's screen is being processed and transmitted . Larger image files , on the other hand , require more storage space on the image server . Within parasitology , the visualization and identification of protozoa would clearly be improved by using a higher objective and camera resolution in the scanning process . As the size of the produced virtual slides thus increases , longer digitization times are needed and the storage capacity needs grow exponentially . Especially if , in addition , multiple focal layers are captured at high resolution , data storage may currently be a rate-limiting factor . It may , however , be a temporary problem since storage costs are dropping rapidly and digitization speeds continuously improve . The expanding global World Wide Web , faster communication speeds and the increase in number and capacity of personal computers , will clearly improve the usefulness of virtual microscopy . By incorporating a server into the local area network at SMI and the Karolinska Institutet ( Stockholm , Sweden ) we were able to view the virtual slides with high communication speed ( ∼100 Mb/s ) . Installation of a local server could be a solution also for hospitals or universities in developing countries with limited connections to the Internet . The technical requirements of a local server are modest and expenses could be kept low , but would clearly overcome data transfer problems . We have recently established a European Virtual Microscopy Network , which currently consists of virtual microscopy image servers and mirrored digital specimens in five European countries . This network automatically directs the user to the server with the best connection speed and according to an ongoing study , will further improve the image loading and viewing speed ( unpublished data ) . Thus we think it is realistic to expect that WMP will benefit from further technical improvements and become a useful tool for parasitological education and quality control globally and we encourage proposals for utilizing the WMP within the framework of activities outlined above . | Here , we describe a novel tool to observe parasites by virtual microscopy on the Internet . Microscopy-based identification of parasites is the basis for both diagnostics and epidemiological assessment of parasite burden globally . Yet , quality assessment of diagnostic parasitology laboratories is difficult , as delivering identical educational specimens has been impossible . In this study , a series of parasite specimens on ordinary glass slides were digitized using a recently developed microscope scanner technique . Up to 50 , 000 images captured at high magnification are digitally stitched together to form a representation of the entire glass slide . These “virtual slides” digitized at a thousand-fold magnification can hold more than 60 gigabytes of data . Handling such large amounts of data was made possible because of efficient compression techniques and a viewing system adopted from the geospatial imaging industry . Viewing the samples on the Internet very much resembles , for example , the use of Google Maps , and puts only modest requirements on the viewer's computer . In addition , we captured image stacks at different focal planes , and developed a web-based viewing system for three-dimensional navigation in the specimens . This novel technique is especially valuable for detailed visualization of large objects such as helminth eggs in stool specimens . | [
"Abstract",
"Introduction",
"Materials",
"and",
"Methods",
"Results",
"Discussion"
] | [
"infectious",
"diseases/tropical",
"and",
"travel-associated",
"diseases",
"computer",
"science/applications",
"infectious",
"diseases/neglected",
"tropical",
"diseases",
"infectious",
"diseases/helminth",
"infections",
"evidence-based",
"healthcare/methods",
"for",
"diagnostic",
... | 2008 | Web-Based Virtual Microscopy for Parasitology: A Novel Tool for Education and Quality Assurance |
Genome-wide RNA expression data provide a detailed view of an organism's biological state; hence , a dataset measuring expression variation between genetically diverse individuals ( eQTL data ) may provide important insights into the genetics of complex traits . However , with data from a relatively small number of individuals , it is difficult to distinguish true causal polymorphisms from the large number of possibilities . The problem is particularly challenging in populations with significant linkage disequilibrium , where traits are often linked to large chromosomal regions containing many genes . Here , we present a novel method , Lirnet , that automatically learns a regulatory potential for each sequence polymorphism , estimating how likely it is to have a significant effect on gene expression . This regulatory potential is defined in terms of “regulatory features”—including the function of the gene and the conservation , type , and position of genetic polymorphisms—that are available for any organism . The extent to which the different features influence the regulatory potential is learned automatically , making Lirnet readily applicable to different datasets , organisms , and feature sets . We apply Lirnet both to the human HapMap eQTL dataset and to a yeast eQTL dataset and provide statistical and biological results demonstrating that Lirnet produces significantly better regulatory programs than other recent approaches . We demonstrate in the yeast data that Lirnet can correctly suggest a specific causal sequence variation within a large , linked chromosomal region . In one example , Lirnet uncovered a novel , experimentally validated connection between Puf3—a sequence-specific RNA binding protein—and P-bodies—cytoplasmic structures that regulate translation and RNA stability—as well as the particular causative polymorphism , a SNP in Mkt1 , that induces the variation in the pathway .
The potential for using comprehensive data sets , such as RNA expression data , as a means for uncovering complex genetic traits has led to the production of eQTL data – gene expression data across a population of genetically diverse individuals – in a variety of different organisms [1]–[6] . One application of this data type is the use of subtle perturbations in the regulatory network induced by natural genetic variations to reveal the regulatory interactions and influences . These data thus provide a unique opportunity for uncovering the cell's regulatory structure , and for revealing the genetic basis for phenotypic traits . Many approaches have been developed that attempt to identify one or more genetic regions containing polymorphism ( s ) that cause a change in gene expression [1]–[3] , [7] , [8] . Some approaches [7] , [8] expand these ideas by searching for a more integrated regulatory network , where targets are viewed as affected not only by changes in genotype , but also by changes in the activity level of regulatory proteins , estimated by their mRNA levels . These methods have been used successfully to identify important regulatory relationships , including some that underlie key phenotypic traits . A key challenge in the application of these methods is that the number of candidate regulators is enormous relative to the amount of available data , making it difficult to robustly identify the correct regulator . This problem is exacerbated when multiple regulators are correlated , and therefore many regulators have similar potential to explain the expression of their targets . Unfortunately , correlated regulators are the rule , rather than the exception , both for sequence polymorphisms ( due to linkage disequilibrium ) and for regulatory genes identified by gene expression signature ( due to co-expression ) . In these cases , methods that attempt to recognize regulatory relationships are often forced to make choices that are arbitrary , misleading , or non-specific . For example , most linkage-based approaches identify only a chromosomal region , leaving a human to predict the true causal polymorphism ( s ) within the region . This approach results in a large number of hypotheses for experimental testing , especially in higher organisms , such as humans where chromosomal regions are often very large and methods of experimental validation are time and labor intensive . In this study , we propose a novel approach based on the observation that not all candidate regulators are equally likely to play a causative role in gene expression . Indeed , researchers often manually select among candidate polymorphisms , favoring those that are in conserved regions , those that produce significant changes in protein sequence , or those that lie in functionally relevant genes ( such as transcription factors or signaling proteins ) . However , it is not clear how to weight these different features , and , indeed , their relevance might vary across organisms , tissues , or even conditions . We propose a novel Bayesian method , called Lirnet , that automatically learns three model components ( Figure 1 ) : a regulatory network; a set of regulatory potentials for all candidate regulators , evaluating the likelihood that they play a causal role; and a set of regulatory priors , which define a regulator's regulatory potential in terms of its regulatory features , such as the conservation of a SNP or the annotated function of a gene ( see Methods; Figure 2 , Tables S1 , S2 , S3 ) . All three components are learned from the data , in an unbiased way , using an iterative approach: As Lirnet constructs a set of regulatory programs for the genes in the data , it learns which types of regulators are more predictive of their putative targets; it adjusts the regulatory prior to match the observed trends , and it then relearns the regulatory programs using the adjusted prior ( Figure 1 ) . Thus , the method automatically tailors itself to the regulatory interactions in a particular data set . Moreover , Lirnet can use any set of features that are likely to indicate regulatory potential , including sequence features ( such as conservation or significance of amino acid change ) that are available for many organisms . This feature , combined with Lirnet's ability to learn the importance of these features automatically , makes it especially advantageous for mammalian systems , where many forms of prior knowledge used in simple model organisms are incomplete or unavailable . Recently , there have been several approaches for identifying a causal gene in eQTL data [9] , [10] . Zhu et al [9] learned a Bayesian network from the eQTL data . They show that incorporating various other genomic data such as transcription factor binding sites ( TFBS ) and protein-protein interaction ( PPI ) data improves the quality of the learned network . They also use the network for identifying the most likely causal regulator in a genomic region . For a group of genes linked to a given region and a candidate regulator in the region , they test the overlap between the linked genes and the genes regulated by the candidate in the learned network . Suthram et al . [10] , building on earlier work of Tu et al . [11] , propose an alternative method called eQED , which also aims to select a particular regulation within a linked region . These methods define an electric-circuit model for the flow of influence in a separate PPI network and use it to select he most relevant regulator in the region . These methods , like ours , utilize domain knowledge encoded in TFBS or PPI data for identifying causal regulators . However , these methods do not incorporate any information on properties of individual SNPs , such as their conservation score ( a feature that was indeed chosen to be important in our automated analysis ) . Moreover , both methods are biased towards discovering regulatory relationships involving transcription factors . In comparison , Lirnet uses a broad range of regulatory features , enabling the identification of novel regulatory relations , as well as those involving other mechanisms such as chromatin and mRNA degradation , as is demonstrated by Lirnet's experimentally validated hypothesis of a relationship between two post-transcriptional regulatory pathways . An alternative approach is proposed by Jiang et al . [12] , whose method prioritizes non-synonymous SNPs as being disease-related , based on various features such as weight and biochemical properties . Their predictor was trained on a database of over 20 , 000 non-synonymous SNPs , annotated with a disease level for each SNP , and achieved a high prediction performance . Although this method takes the SNP-specific features as input and prioritizes individual SNPs , it does not incorporate the gene-based and network-based features that are used in our analysis , as well as the ones of Zhu and Suthram . It is also restricted to the analysis of non-synonymous coding SNPs , and focuses on the relevance to a single phenotype ( i . e . a disease ) . We test our approach using two eQTL data sets , selected to assess method's versatility . The first is the HapMap data set of Stranger et al . [4] , which contains expression profiles for lymphoblastoid cell lines generated from participants in the human HapMap study . The second is the yeast data set of Brem and Kruglyak [3] , which measures the mRNA expression and genotype of 112 recombinant progeny generated by mating of two genetically diverse strains of S . cerevisiae , a laboratory strain ( BY ) and a wild vineyard strain ( RM ) . We show statistically that the learned regulatory potential significantly improves the quality of the learned regulatory programs , as evaluated by the percent of the variance explained . We also evaluate the biological validity of our learned regulatory programs by comparing them to other biological data , not used within the algorithm . Our results clearly demonstrate that Lirnet produces more accurate regulatory programs than previous approaches , including Geronemo [7] and the recent methods of Suthram et al [10] and Zhu et al . [9] . We also provide a detailed analysis of some of the inferred yeast regulatory programs , and demonstrate that Lirnet can correctly identify the causative polymorphism within a large , linked region , even in regions containing several biologically plausible candidates . We study in greater depth one of the pathways produced by Lirnet , involving two modules related to post-transcriptional gene regulation . In this case , Lirnet suggested a three-tiered regulatory cascade: at the lowest level , a module comprising a set of genes that are bound by the sequence-specific RNA binding protein , Puf3; the module's predicted regulatory program , which utilizes factors involved in several distinct post-transcriptional regulatory processes , including members of the P-body complex , an RNA storage and degradation complex that can also modulate mRNA translation; and at the highest level , a chromosomal region containing the causal variation , and , using its learned regulatory prior , even a particular gene in the region – Mkt1 , whose protein product binds ( indirectly ) to the PolyA-binding protein at the 3′ region of mRNA transcripts [13] . We provide multiple forms of experimental data supporting Lirnet's computational prediction , including the causal role of Mkt1 . The resulting regulatory network for the yeast data and the software are freely available on our website http://dags . stanford . edu/lirnet/; the learned network can be effectively explored using our visualization tool , downloadable from the same website .
We briefly review the Lirnet method , referring to the Methods for a full description . Lirnet uses genotype and expression data of genetically diverse individuals ( eQTL data ) and aims to learn a regulatory prior concurrently with reconstructing a regulatory network . Building on earlier work [7] , [14] , Lirnet clusters genes into modules with the assumption that expression of the target genes in each module is governed by the same regulatory program . As with several other methods for the reconstruction of regulatory networks , Lirnet can accommodate two types of regulators: values of genotype markers ( genotype regulators ) , representing genetic polymorphisms on chromosomal regions [1]–[3]; and expression levels of genes that are known to have regulatory roles ( expression regulators ) , representing activity levels of genes that might regulate that module [7] , [9] , [14] , [15] . Lirnet's regulatory programs are based on linear regression , a choice designed to allow for the incorporation and learning of regulatory potentials . For each module m , the expression levels of genes in the module ( denoted by ym , j ) are modeled as a linear regression of candidate regulators ( denoted by x1 , … , xn ) : ym , j∼wm , 1x1+wm , 2x2+…+wm , nxn , where all genes in the module share the same weights wm , k A regulator r that has a zero weight wm , r has no effect on the expression of the targets in module m . A biologically plausible regulatory program should have a small number of regulators with a non-zero weight . To achieve this goal , we use the LASSO method [16] to select only the most significant regulators . In its simplest form , LASSO adds a fixed penalty term to the objective function that introduces a uniform bias towards sparsity in the weights w . Lirnet , inspired by our recent work on feature selection [17] , incorporates regulatory potential by allowing different regulators to have different sparsity biases; a regulator r whose regulatory potential is low for a module m will have a stronger bias towards wm , r = 0 , and thus a lower probability of being selected as an active regulator for that module . The regulatory potential Cr of regulator r is defined to be a function of its regulatory features fr . The method flexibly accommodates any property or combinations of properties of a regulator ( Figure 2 , Tables S1 , S2 , S3 ) that might be indicative of its likelihood of having a causal effect on its targets . These features can include features of the regulator alone , such as the location and significance of sequence polymorphisms , the function of the gene ( transcription factor , signaling protein , etc . ) , and conservation of the polymorphic site . They can also include features that involve both the regulator and the targets , such as the enrichment of the module with genes having known relationships to regulator ( e . g . transcription factor targets ) . The regulatory prior β encodes the importance given to each regulatory feature ( see Methods ) , and is automatically learned from the data , allowing less relevant regulatory features to be ignored and others to manifest their significance . The learning algorithm of Lirnet jointly estimates wm , r's and β by maximizing a joint objective that involves both . The algorithm iterates three steps until convergence ( Figure 1 ) : ( 1 ) learning the regulatory program for each module by estimating the weights wm , r's; ( 2 ) learning the regulatory priors β that reflect the importance of each regulatory feature; ( 3 ) computing the regulatory potential for each candidate regulator , module pair , based on the current β , thereby biasing their selection in the next iteration's regulatory programs . The output of Lirnet is thus threefold . First , it constructs a set of learned regulatory programs for the modules used in the analysis; for module m , these are all the regulators r with a non-zero weight wm , r . Second , it constructs a quantitative regulatory potential both for genes and for specific sequence polymorphisms within them , allowing us to rank candidates for the causal sequence variation and to prioritize hypotheses for further testing . Third , it produces a set of regulatory priors , which may provide insight on the properties of a polymorphism that tends to induce an effect on its downstream targets . To test the versatility and generality of Lirnet , we applied it to two very different eQTL data sets . The first is the eQTL data set of Brem and Kruglyak [3] , which measured the mRNA expression and genotype of 112 S . cerevisiae strains derived as the F2 progeny of a BY/RM cross . The second is the expression data measured in the lymphoblastoid cell lines of the 60 unrelated HapMap individuals in the CEU data set , using only the genotypes of a subset of markers ( the 500 K tag SNPs on the Affymetrix chips ) as regulators ( see Methods ) . Figure 2 shows the regulatory priors of the most significant regulatory features in each of these data sets , as identified by the Lirnet algorithm . Several aspects of the features automatically chosen as important by the algorithm are revealing . First the top learned regulatory potentials between two organisms as different as human and yeast are remarkably consistent , supporting the robustness of our approach . For example , aside from the feature indicating the stop codon , which affects only 43 genes , the strongest positive weight on the regulatory potential in both datasets is given to a feature denoting whether the sequence variation correlates with changes in the expression level of the closest gene ( cis-eQTL ) . This gene-level feature indicates that the polymorphism is already causal towards a change in the cell ( the expression level of the gene ) and hence may have additional downstream effects . The second most significant regulatory feature , also in both data sets , is the conservation score , consistent with the hypothesis that changes in residues conserved across millions of evolutionary years are more likely to have a causative influence . Also with a significant weight are features that evaluate the functional relevance of the position of the polymorphism , and the significance of the actual change . These features include , for example , the presence and type of amino acid changes , and polymorphisms in the 5′ or 3′ UTR . Surprisingly , in both data sets , synonymous SNPs weigh more heavily than non-synonymous ones . This phenomenon might arise from the fact that synonymous SNPs can have an effect on translational efficiency or mRNA destabilization [18] , consistent with recent findings that such SNPs are under significant purifying selection [19] , [20] . The selection of this regulatory feature by our automated method lends support to this hypothesis , and is worthy of further investigation . Aside from sequence-based features , we also provided the method a rough categorization of gene function , allowing it to learn which types of genes are likely to play a regulatory role . Despite receiving no prior knowledge about the relative importance of the various functional categories , the method automatically assigns high weight to functional categories with a regulatory role: In the human data , the highest weights among those features are given to genes involved in cell death , transport , cell growth , signal transduction , transcription , and cell communication . In the yeast data , ‘transcription regulator activity’ , ‘telomere organization and biogenetics’ , ‘protein folding’ , ‘glucose metabolic processes’ , and ‘RNA modification’ are chosen to be important . Notably , other work supports the differences between BY and RM in many of these processes , including glucose processing ( BY and RM demonstrate dramatically different growth rates on a number of carbon sources including glucose; D . Pe'er , unpublished data ) and telomere organization [7] , showing the value of allowing the regulatory priors to be tailored to particular data sets and organisms . Lirnet can also take advantage of other functional data , when available . For example , in the yeast data , the pairwise feature derived from ChIP-chip binding between the regulatory gene and targets in the module received a relatively high weight . We note , however , that the method is also effective when such data are not available for a given transcription factor and set of target genes ( see Oaf1 example below ) or when the features themselves are not available , as in the case of the human data . We next tested whether the learned regulatory potentials improved the quality of the learned regulatory program , by computing the proportion of genetic variance ( PGV ) explained by the learned program . We compared Lirnet with a uniform regulatory potential ( hereafter “flat” Lirnet ) , Lirnet with a learned regulatory potential , a standard single-marker linkage ( as in [1]–[3] ) and Geronemo [7] . The results ( Figure 3 ) demonstrate that Lirnet explains a dramatically larger fraction of the variance for a much larger set of genes than all the other methods . For example , in the yeast data , Lirnet with learned regulatory potentials explains over 50% of the PGV for 1 , 644 genes , compared to 1 , 457 genes for flat Lirnet , 828 genes for Geronemo and 230 genes for the method of Brem & Kruglyak . The same advantage translates across the spectrum of PGV values , and is arguably even greater at the tail , where many genes that are very poorly explained by other methods have a considerable fraction of PGV explained by Lirnet . A more refined PGV analysis , with an independent test set , shows that this dramatic improvement does not arise from overfitting to the test data ( Figure S1 ) . In fact , the Lirnet model has a comparable number of parameters to Geronemo and fewer parameters than the method of Brem and Kruglyak , due to the use of modules . Notably , even flat Lirnet considerably outperformed both the single-marker linkage and Geronemo methods , suggesting that Lirnet would still perform well in cases where the learned regulatory potential was less informed . The gap between Lirnet and flat Lirnet appears to increase when using more markers ( see Figure S1D for human results on 100 K markers ) , and is larger in the YRI than in the CEU data . These results are consistent with a hypothesis that the benefits of a non-uniform regulatory prior are more pronounced when regulatory potentials are aggregated over smaller regions of linkage disequilibrium . Thus , one can expect the benefits of Lirnet's learned regulatory priors to grow as we move to denser genotyping . We evaluated how well the learned regulatory program recovers known regulatory interactions . We begin with a comprehensive analysis of the quality of the learned regulatory programs , demonstrating that they are consistent with other sources of data indicating regulatory interactions that were not provided to the Lirnet algorithm . We then provide a comparison to the state-of-the-art method of [9] , which was applied to the same data . As a gold-standard set of regulator-target relationships is not available , we constructed a comparison test set from various datasets: deletion and over-expression microarrays [21] , [22]; chromatin immune-precipitation ( ChIP-chip ) binding experiments [23]; mRNA binding pull-down experiments [31]; transcription factor binding sites [65]; and a literature-curated set of signaling interactions from the Proteome database ( http://www . proteome . com/ ) . Although each of these data sets has its own limitations in terms both of false negatives and of false positives , agreement with these orthogonal data sources is a reasonable metric for evaluating the quality of a method's predictions . For a prediction that a regulator r regulates a module m , we defined it to be validated if there was significant overlap ( hypergeometric p<0 . 01 ) between the members of m and the putative targets of r , suggested by one of the above datasets . We note that none of these datasets was used for constructing the regulatory features for Lirnet: Lirnet used only the ChIP-chip data set of [24] , and all regulator-target pairs that appeared in these data were removed from the evaluation data [23] . Most of these data sets focus on regulatory relationships where r is a transcription factor , whereas Lirnet and the other methods we evaluate are also capable of identifying cases where r plays a different regulatory role , such as signaling , chromatin modification , or RNA degradation . Therefore , to increase the coverage of our validation effort , we also considered indirect regulatory relationships ( Two-Step Cascade in Table 1 ) , where a method predicted a regulator r that has some close relationship with a transcription factor t , and t is confirmed in the above data sets to regulate m . We considered cases where t and r have a reliable protein-protein interaction ( PPI ) ( Xenarios et al . 2000 ) ; and cases where r phosphorylates t in the Proteome data set . Table 1 summarizes the number of validated regulators for various models , applied to the same set of modules: Lirnet with a uniform regulator potential , Lirnet with the learned regulatory potential , Geronemo , and a random model . ( See also Table S5 for a full list of Lirnet predictions and their support . ) Overall , Lirnet recovers a larger fraction of the known regulatory interactions than the other methods . We note that the reference set supports only a subset of the predicted regulatory interactions . This fact is not surprising , as the data sources used for constructing the reference set focus on transcriptional regulation , whereas Lirnet and Geronemo cover a much larger set of regulatory relationships . Although we have made some attempt to expand our reference set to cover signaling interactions , the data set of literature-curated signaling interactions is only a small fraction of the total set of signaling interactions that presumably hold in yeast . Moreover , as shown in our previous study [7] , a large part of the regulatory interactions in these data sets represent chromatin modification and post-transcriptional regulation , which are not represented in our reference set . The BY/RM cross also exhibits a large amount of cis-regulation , which we did not explicitly model in the Lirnet analysis . Nevertheless , of 492 cis-regulated genes – those whose nearby marker is significantly predictive of its expression level ( t-test p-value<1e-5; 12 . 8% of the 3152 genes used in our analysis ) , 307 genes ( 76 . 4% ) are assigned to modules with cis-regulatory programs . More specifically , 149 genes are assigned to modules that have the genes' nearby markers as genetic marker regulators; 158 other genes are assigned to modules that have their nearby expression regulators having their markers in the regulators' regulatory programs , suggesting an indirect cis-effect from a locus to an expression level of a regulator in that region to a target . We also compared our results with those of the recent Bayesian network method of [9] . As this method infers regulatory interactions for individual genes rather than modules , we used the number of regulator-target pairs validated in the reference set as our evaluation metric . For a fair comparison , we removed from the reference set any data sets that are used for learning either model , leaving only the deletion and over-expression microarray data [21] , [22] . The results ( Figure 4A ) show that , for various levels of expression change in the deletion or over-expression data sets , the regulatory interactions inferred by Lirnet are more consistent with previously known regulatory relationships . In Figure 4A , we see that Lirnet recovers many more supported regulators than the Zhu et al . method . This large discrepancy is partially due to the fact that their method largely focuses on transcription factors , and hence is incapable of picking up many of the regulatory relationships that are uncovered by Lirnet . However , even if we focus attention only on the relationships that could be compared to the deletion/over-expression data ( Figure 4B ) , and evaluate the fraction that were validated in these data , we see that Lirnet significantly outperforms the Zhu et al . method . One of Lirnet's key features is its ability to identify a specific causative regulator in a linked chromosomal region , an ability also presented in several other recent methods [9] , [10] . Since other methods for comparison focused on identifying a causal gene not a specific SNP , we also prioritized genes for a direct comparison . We first compute the regulatory potential of all SNPs in the region of interest , relative to each module ( to account for pairwise regulatory features ) . We then rank each gene based on the highest-scoring SNP associated with it . Figure S2 shows the overall distribution of the regulatory potentials for both SNPs and genes . We can see that the vast majority of SNPs and of genes have a fairly low regulatory potential . The distribution begins to tail off at a regulatory potential of about 0 . 694; only 0 . 22% of SNPs and 2% of genes have a regulatory potential that exceeds this value . We first compare to the recent work of Zhu et al . , who focus on 13 “hot spots” – chromosomal regions identified to regulate expression levels of a number of genes in a previous study [1] . Previous work has identified regulators for several of these hot spots , and Zhu et al . provide new experimental validation for a number of others . To compare to these results , we considered the genes linked to each hot spot as a “module” , and applied the learned regulatory prior ( Figure 2 ) to individual SNPs in each region to compute each gene's regulatory potential ( see Methods ) . We sorted the genes in each hot spot based on their regulatory potential , and listed the top 3 genes for each hot spot . Figure 4B compares the result of the suggested causative genes in each region between Lirnet and the method of Zhu et al . ( 2008 ) . We see that , of the top Lirnet regulators , 14 regulators , spanning 11 hot spots , have experimental support ( see Methods ) , in comparison to 8 regulators ( 7 hot spots ) in the analysis of Zhu et al . Even if we consider only Lirnet's top regulator for each region , there is experimental support for 10 regulators . We also compare to the recent method of Suthram et al . [10] , which improves on earlier work of Tu et al . [11] . These methods consider a gene and a chromosomal region to which it is linked , and analyze the flow in a protein-protein interaction network to select a particular causal regulator within the region . Suthram et al . validate their results relative to a pre-defined set of 548 regulatory relationships , extracted from gene knockout or overexpression microarray studies [21] , [25] , similarly to our analysis above . The predicted network of Suthram et al . was not available , so we evaluated Lirnet using their protocol and the reference set , to allow for a direct comparison . For each target gene and linked region , we selected , as the Lirnet predicted regulator , the gene whose regulatory potential in the region was highest . We then evaluated these predictions using the reference set of Suthram et al . The results , shown in Table S6 , show that Lirnet significantly outperforms both the method of Suthram et al . and the previous method of Tu et al . [11] , according to this evaluation metric . We also performed an in-depth analysis of some of the specific regulatory modules produced by Lirnet for the yeast data set , and evaluated its ability to identify both the correct regulators and the specific polymorphisms that gave rise to the expression change in the targets . One example of the predictive power of assigning regulatory potential to individual SNPs within a large chromosomal region is the Zap1 module ( Figure 5A ) . The module contains ten target genes and two major regulators , the gene expression pattern of ZAP1 , which encodes a transcription factor ( TF ) that activates genes in response to Zinc [26] , [27] , and a genetic region on chromosome 10 that contains ZAP1 . Of the ten target genes in the module , six were among 40 probable Zap1 targets based on the presence of a consensus ZRE element and RNA expression patterns in zinc and in the absence of Zap1 ( p = 5 . 7×10−10 ) [27] . While the causative role of Zap1 in this data has previously been affirmed a number of times [3] , [7]–[9] , Lirnet automatically identified polymorphisms within Zap1 as the ones most likely , within the linked region , to play a causal role ( Table S7 ) . The regulatory potential of the identified SNP is the highest over all yeast SNPs ( Figure S2 ) . The most significant regulatory feature by far in this identification was the known binding relationship between Zap1 and two of its target genes , but other features also played a role ( Figure 5B ) . Thus , the method has identified a TF-target relationship for which there is significant biological support . Importantly , however , Lirnet is also able to predict such relationships when relevant functional data such as binding assays are not available . One such example is the peroxisome module ( Figure 6A ) , containing ten genes that are enriched for processes related to fatty acid metabolism ( hypergeometric p<4 . 8×10−6 ) and peroxisome organization and biogenesis ( hypergeometric p<5 . 5×10−6 ) , nine of which we considered for further analysis ( see Methods ) . Lirnet suggests two regulators: expression level of PIP2 ( alias: OAF2 ) , a gene that encodes a Zn ( 2 ) -Cys ( 6 ) TF that heterodimerizes with Oaf1 to regulate genes involved in peroxisomal functions via an ORE element [28] , and a genetic region between nucleotides 51 , 324 and 52 , 943 on chromosome 1 . Of the 11 genes in this region of chromosome 1 ( Figure 6B; Table S8 ) , Lirnet selected polymorphisms within OAF1 as having the highest regulatory potential; the regulatory potential value of OAF1 is within the top 1% over all genes ( Figure S2 ) . Several forms of data support the role of the Oaf1/Pip2 heterodimer in regulating this module . Of the nine target genes analyzed in the module , six contained the canonical ORE motif ( p = 1 . 8×10−6 ) . Moreover , five were in the top 1% of most significantly down-regulated genes in a microarray experiment that compared RNA expression levels of an oaf1Δ versus a wild-type ( BY ) strain under inducing conditions [29] ( p = 8 . 0×10−9 ) . We note that the PIP2 promoter itself is Oaf1-dependent and contains an ORE element [29] . Thus , differences in PIP2 expression patterns across the 112 segregants are also likely to be partly due to polymorphisms in OAF1 . However , the fact that PIP2 expression was selected in addition to the OAF1 genotype demonstrates Lirnet's ability to identify multiple relevant regulators , even when they are correlated . We note that OAF1 was not identified by previous methods analyzing this data set . The BY allele of OAF1 contains two non-conservative coding polymorphisms at conserved positions that are likely to alter its function: an R70W polymorphism in the DNA binding domain ( the highest-scoring SNP ) and a Q447P polymorphism in the ligand binding domain [30] . OAF1's high regulatory score ( Figure 6B ) is a combination of the correlation between its expression and genotype ( cis-regulation ) , shared GO process annotations with the target genes , the presence of non-synonymous coding mutations and their effects on protein properties ( e . g . , pKa and pI ) , and ( to a lesser extent ) its function as a transcriptional regulator . Importantly , the ChIP-chip data set [24] used in our analysis did not contain Oaf1 binding information and therefore did not influence the choice of the regulator , demonstrating Lirnet's effectiveness even when binding data are not available . Moreover , another gene in the same region of chromosome 1 ( PEX22 ) shares common functional annotations with many of the target genes , yet received a significantly lower score . This highlights the fact that functional annotations , although a useful source of prior knowledge , are not the primary cue used by Lirnet . Both of these characteristics are likely to play an important role in the analysis of data from other organisms , where binding data and functional annotations are both limited . Lirnet also suggested an intriguing hypothesis regarding a cascaded pathway involving two modules . The first ( hereafter the “Puf3 module” ) contains 153 co-expressed genes , highly enriched ( p<10−91 ) for nuclear genes with mitochondrial functions , and very highly enriched ( p<10−130 , Figure S3A ) for genes whose mRNA transcripts are bound by the sequence-specific RNA-binding protein , Puf3 [31] . This enrichment is specific to Puf3 binding and not just coincident with the preponderance of mitochondrial genes ( Figure S3B ) . Also , the Puf3 targets that are in the module show higher Puf3 motif score than those not ( Figure S3C; Text S1 ) . Indeed , RNA expression levels in a puf3Δ mutant in rich medium ( YPD ) showed a significant up-regulation ( p<10−37 ) mRNA levels of the module genes ( Figure S3D ) . While these results suggested that the highly coherent expression profile of this module was due , at least in part , to regulation of RNA stability via Puf3 , we found that neither PUF3 mRNA expression nor its genotype is correlated with expression of the module genes ( Figures 7A , S4 ) , suggesting that Puf3 itself was not the regulator driving the observed variability across the strains . The Lirnet analysis identified several genes as being involved in the regulatory program of the Puf3 module ( Figure 7A ) . Towards the top of the list , we find DHH1 ( ranked 1st ) and KEM1 ( ranked 4th ) , two components of the dynamic cellular structures called cytoplasmic processing bodies ( P-bodies ) [32]–[35] . P-bodies are sites of RNA storage [32] that can modulate mRNA translation or degradation: RNA transcripts are translationally silenced while stored in the P-body [36] and can be subsequently degraded or released back into the translating pool [37] . P-bodies contain the catalytic subunits of the mRNA de-capping enzyme Dcp1/Dcp2 [38] , [39] , whose activity is regulated by Dhh1 . However , the signals for determining mRNA localization to P-bodies and subsequent degradation or release have not been identified [32] . Thus , Lirnet suggested an intriguing regulatory connection between the Puf3-bound transcripts and a known posttranscriptional regulatory complex ( P-bodies ) . If Puf3 serves as a regulatory signal in one or more of the processes associated with P-bodies ( RNA targeting , degradation , or release back into the translating pool ) , we would expect Puf3 protein to be localized to P-bodies . We therefore used fluorescence microscopy to test the subcellular localization of Puf3 in wild-type BY cells . Indeed , under certain conditions ( see Methods ) , a Puf3-GFP fusion protein formed bright punctuate spots in the cytoplasm which co-localized with those of known P-body components , Dhh1 and Edc3 ( Figure 7B ) . These results are consistent with those of a previous study [31] that reported punctuate cytoplasmic Puf3-GFP fluorescence in the BY strain background , but did not test for co-localization with P-bodies . This finding demonstrates the role of p-bodies in the regulation of the Puf3 module genes , but does not elucidate the causal SNP responsible for the difference between strains . To identify this SNP , we explore the Lirnet predictions for the regulatory program determining P-body expression . Dhh1 and Kem1 are themselves members of another module that we call the post-transcriptional regulatory ( PTR ) module ( Figure 8A ) . This module also contains other regulators of the Puf3 module , including GCN1 and GCN20 , two members of a complex that regulates translational repression in response to nutrient starvation . Many other module members are associated , directly or indirectly , with post-transcriptional regulation ( Figure S6 ) . The sole regulator of the PTR module is a genotype marker located on Chromosome XIV ( Figure 8A ) , the same region that had previously shown as linked to several of these targets and members of the Puf3 module [1] , [40] . The smallest region linked to the PTR module spans more than 30 genes ( Figure 8B; Table S9 ) , making a systematic evaluation of the candidates a significant effort , even in a genetically tractable organism like yeast . We therefore used Lirnet to evaluate the genes in the region based on their regulatory potential for the PTR module . As shown in Figure 8B , MKT1 is ranked as the highest scoring gene in terms of the learned regulatory potential . The regulatory potential of MKT1 is within the top 1% over all genes ( Figure S2 ) . Mkt1 interacts with the Poly ( A ) -binding protein associated factor , Pbp1 , and is present at the 3′ end of RNA transcripts during translation [13] . Mkt1 contains a highly-conserved nuclease domain , homologous to the human XPG endonuclease [13] and is required for the translational regulation of an Mpt5/Puf5-dependent transcript ( HO ) [13] . In the BY strain background , MKT1 harbors two non-synonymous coding mutations ( Figure S7 ) . The first variation lies in its putative nuclease domain and makes a non-conservative amino acid change ( G30D ) in a residue highly conserved in yeast . Both this and the more conservative mutation ( R453K ) are outside a Pbp1 interacting domain , which maps to the C-terminus of the protein ( residues 601–760 ) [13] . MKT1 has thus far been implicated in four diverse processes , including propagation of a double stranded RNA virus [41] , growth at high temperature [42] , efficiency of sporulation [43] , and gene-specific translational regulation [13] . In many of these cases the BY variant , and , where examined , the G30D mutation specifically , appears to introduce a loss of function mutation [41] , [43] , [44] . Both these SNPs contributed significantly to MKT1's high regulatory potential , with the G30D SNP scoring somewhat higher . The properties that contributed the most to this selection are conservation , cis-regulation , the amino-acid properties of the coding SNP , and the common GO process with the targets . These properties are precisely the ones that an expert biologist would look for in a manual scan of the region; but Lirnet automatically learned the significance of these features and their relative importance , allowing it to correctly identify the correct polymorphism using a fully automated approach . Moreover , we note that the region contains a number of other plausible candidate genes , including transcription factors and a number of mitochondrial genes; nevertheless , Lirnet identified Mtk1 as top ranked . To test the effect of the loss of Mkt1 function on the PTR and Puf3 module genes , we deleted the MKT1 open reading frame in the RM background and measured genome-wide RNA expression by DNA microarray analysis . Consistent with Lirnet's predictions , we observed a modest but highly significant down-regulation ( KS p-value<10−23 ) of the Puf3 module in the RM mkt1Δ strain ( Figure 8C ) . While previous work demonstrated Mkt1's role in repressing the translation of an Mpt5/Puf5-dependent transcript [13] , our results suggest that Mkt1 plays a role in the RNA stability of Puf3-dependent transcripts . In the PTR module , 16 of 40 genes were among the top 5% most up-regulated genes ( hypergeometric p<10−10; Figure 8A ) , a set which includes the Puf3 module regulators DHH1 , KEM1 , GCN1 , and GCN20 , with expression changes ( 1 . 36 fold increase ) similar to the difference between RM and BY . A similar response of the Puf3 and PTR module genes was observed in an RM strain harboring the BY allele of MKT1 ( Figure S8 ) , further supporting the role of MKT1 in the RNA expression differences seen in the original population . Taken together , these results provide strong evidence that MKT1 contains a causative variation for these modules , and further demonstrate Lirnet's ability to identify the correct causal regulator even in a large linked region Thus , the Lirnet procedure automatically uncovered a comprehensive 3 tiered regulatory cascade in which MKT1 regulates P-body abundance , that consequently regulate Puf3 target abundance , providing significant detail and insight into the mechanism through which the Puf3 module is regulated ( Figure S9 ) . Other methods [8] , [11] recently applied to these data produced no hypothesis regarding this pathway . The analysis of Brem and Kruglyak [3] linked these genes and many others to a region on Chromosome XIV ( Figure S10 ) , but no causal mutation was identified . Geronemo picked up Dhh1 as a key regulator but failed to identify the causal SNP or gene involved ( Figure S11 ) . The recent analysis of Zhu et al . [9] , performed in parallel to our work , also identified the same genetic region , and provided experimental support for its causal role with respect to a large group of genes . However , their identification of MKT1 as the quantitative trait gene was manual , and based on biological intuition , rather than an automated method . Moreover , their method failed to elucidate the mechanism through which MKT1 regulates the gene expression , missing the role of both Puf3 and the P-bodies . Thus Lirnet is unique in its ability to automatically generate a comprehensive regulatory pathway from causal SNP via the intermediate regulator through which this SNP acts upon the linked genes ( Figure S9 ) .
Advances in technology , most notably the emerging availability of inexpensive sequencing , are likely to give rise to the production of large amounts of data measuring both genotype and expression across large cohorts of individuals , both in human and in model organisms . These data provide a unique potential to elucidate the biological mechanisms underlying complex traits , including both basic biological functions and traits related to human health , such as predisposition to disease or response to treatment . However , our ability to unravel complex traits depends not only on the availability of data , but also on our ability to construct more sophisticated models of the complex pathways underlying these traits , and to identify the polymorphisms that perturb them . The precise identification of specific causal polymorphisms is critical for understanding the mechanisms underlying disease , and for constructing targeted diagnostic tests and treatments . Lirnet provides a unified method that tackles these two inter-related problems: constructing a regulatory network from eQTL data , and learning the extent to which different regulators and sequence variations are likely to play a causal role in modifying expression data . Like other methods that allow for combinatorial regulation , Lirnet provides the potential for uncovering multiple factors underlying complex traits . The use of carefully regularized linear regression allows Lirnet to construct high-quality , biologically plausible regulation programs . Our results demonstrate that many of the regulatory programs inferred by Lirnet have significant support in data sets not used for constructing the network . The key novel component in the Lirnet method is its ability to learn a model of the regulatory potential of individual SNPs and genes , which estimates how likely they are to play a causal role in gene expression . This capability serves two important roles: it allows us to exploit prior knowledge in constructing better regulatory networks , by selecting regulators that are more likely to play a causal role; and it allows us to select a particular polymorphism within a large linked region as the most likely causal regulator . Other methods have been proposed that address one or both of these goals . A number of methods make use of prior knowledge in constructing regulatory networks . The pre-determined selection of candidate regulators [7] , [14] is a form of prior knowledge on the set of regulators . Other methods prioritize the choice of regulatory program using pairwise relationships between TFs and their targets , based on ChIP-chip data or on binding site data [9] , [45] . Various types of prior knowledge has also been used for selecting a causal gene within a linked region , including: correlation of expression between regulator and targets [8] , [9] , [46]–[48] , TF binding data [9] , or paths in a protein-protein interaction networks [10] , [11] . Several important differences distinguish Lirnet from these previous approaches . First , Lirnet avoids the use of special-purpose methods and hand-selected parameters for utilizing different types of prior knowledge . Rather , it automatically learns the regulatory potential from data , allowing it to utilize any set of regulatory features that appear relevant in a given organism and data set , without additional engineering effort . Our results comparing to two state-of-the-art methods [9] , [10] demonstrate that the Lirnet method , with its automatically learned priors , provides significantly better reconstructions of regulatory interactions and better ability to identify the causal polymorphism . At a more qualitative level , Lirnet's ability to flexibly accommodate new types of features will allow it to utilize different types of high-throughput functional data . Second , Lirnet is able to make use of sequence features , such as conservation or significance of the sequence change , in a deeper way than simply eliminating all candidate genes without polymorphisms in the coding sequence [46]; as we saw , this property allows the method to be used in less well-characterized organisms , such as human , where functional data , such as transcription-factor binding or functional characterization , are scarce . The use of sequence-based features allows Lirnet to identify not only the gene that induces the expression change , but also particular sequence polymorphisms within the gene that underlie the functional change . This property is critical in obtaining a mechanistic understanding of the perturbation underlying the phenotype . Lirnet's ability to identify the causal regulator , and even the specific SNP , is likely to be even more valuable in higher-level organisms , where linked regions are long and contain many polymorphisms , and where experiments to test different candidate hypotheses are far more difficult . There are several limitations to our work that provide directions for further developments . First , we have exploited only a basic set of regulatory features; it is likely that improved results can be obtained with a richer set of regulatory features [49] . In particular , a deeper study of the effect of different sequence features , including , for example , synonymous SNPs , may give rise to insights about the effect of different sequence perturbations . Moreover , additional data sets that indicate regulatory interactions continue to be produced , and can be usefully adopted as regulatory features . In particular , all of the data sets that we used to produce our benchmark set of regulator-target interactions ( such as differential expression subject to regulator deletion or over-expression ) can also be used as meta-features , as can other high-throughput data such as signaling interactions [50] , [51] or genetic interactions [52] , [53] . The flexibility of Lirnet allows these features to be easily integrated into the model . More broadly , Lirnet currently utilizes prior knowledge only regarding regulators and regulator-target interactions . We often have data relating to relationships between targets ( such as protein-protein interactions ) , and even between regulators ( such as cooperative or competitive binding ) . It would be interesting to extend the method to exploit such relationships . One exciting opportunity is the application of the concept of a regulatory potential to the task of identifying the causal polymorphisms underlying phenotypes of interest , such as disease or drug response . In particular , it seems plausible that a sequence variation that is more likely to be causal relative to gene expression traits may also have a higher chance of playing a causal role for other phenotypes . If so , then a regulatory potential learned from eQTL data can help narrow down hypotheses in association or linkage studies . This capability could be of value in several settings: in reducing the burden of multiple hypothesis testing by favoring hypotheses that are more likely to be causal [54] , [55]; in identifying plausible regions for resequencing or for follow-up in a larger population; and in prioritizing particular polymorphisms that may be worthy of follow-on experiments . This idea may allow eQTL data from model organisms to be used to increase the power in human disease studies , where expression data from relevant tissues is not readily available .
We applied our analysis to the expression and the genotype data generated from 112 meiotic recombinant progeny of two yeast strains: BY4716 ( BY; a laboratory strain ) and RM11-1a ( RM; a natural isolate ) [3] . Our expression data and genotype data were selected as previously described [7] . Our 305 candidate expression regulators were selected using the process previously described [7] , and are listed in our accompanying website ( http://dags . stanford . edu/lirnet/ ) . We applied Lirnet to the eQTL dataset [4] , [56] of human HapMap individuals −60 European ( CEU ) and 60 African ( YRI ) individuals . Among 47 , 297 probes in the expression data [4] , we picked 7 , 324 whose standard deviation is greater than 0 . 03 and used them for our analysis . The phase II HapMap data [56] contain genotypes for nearly 4 million SNPs . To perform our experiment in a setting that is closer to that of a real association study , we selected only the SNPs that are on a commercial genotyping chip , namely Affymetrix GeneChip 100 k & 500 k , and used only their genotypes in our analysis . We first identified orthologous genes between BY and RM . We downloaded the genome sequences of S288C ( isogenic to BY ) and RM from the Saccharomyces Genome Database ( http://www . yeastgenome . org/ ) and Broad Institute of Fungal Genome Initiative ( http://www . broad . mit . edu ) , respectively ( sequences were retrieved on 12 January 2005 ) . In order to define orthologous genes between BY and RM , we used reciprocal best BLAST hit [57] ( protein sequences of S288C were downloaded from SGD on 12 January 2005 ) . Out of 6 , 683 genes in 16 nuclear chromosomes , 6 , 292 ( 94 . 1493% ) have reciprocal best matches between the two strains . We also retrieved the genomic sequences , 500 bp upstream/downstream of each orthologous pair . We aligned the DNA sequences of the ortholog pair by using LAGAN [58] , and retrieved SNPs between the orthologs . We constructed a set of regulatory features that describes each single nucleotide variation ( SNP ) in terms of various intrinsic characteristics . We identified orthologs between BY and RM , and constructed a list of SNPs , as described above . For human regulatory features , we downloaded data from dbSNP containing a list of human SNPs and their various properties . For each SNP , we defined six kinds of features that can determine its regulatory potential . First , we characterized each SNP in terms of its location relative to genes , resulting in seven regulatory features ( 1 & 12–17 in Table S1 ) . Each non-synonymous coding SNP can change various properties of the corresponding amino acid ( AA ) , which can affect the regulatory role of its gene . Therefore , we described each non-synonymous coding SNP in 10 ways in terms of changes in various properties caused by the corresponding AA change based on various data sources [59] , [60] ( 2–11 in Table S1 ) . A sequence polymorphism on the genomic site that is strongly conserved is more likely to affect the regulatory network . Thus , we characterized each SNP in terms of the conservation score on its genomic site ( 18 in Table S1 ) . The conservation score was computed based on comparison of protein sequences across a large group of species . For yeast data , we downloaded the aligned sequences from Wapinski et al . [61] . For human data , we downloaded the conservation scores from the UCSC human genome browser ( “Most Conserved” track ) . We also incorporated the feature indicating whether the SNP is likely to regulate the expression of the gene in which it resides ( 19 in Table S1 ) . Because regulatory potential of a SNP is likely to be affected by the function of the gene where it resides , we defined a set of regulatory features that indicate whether the gene belongs to each of 87 Gene Ontology ( GO ) categories related to regulatory roles ( 20 in Table S1 ) . For human data , we used 48 GO Slim biological process categories . Finally , a SNP might have different regulatory potential over different modules . We defined three ‘pair-wise features’ that describe how likely a SNP is to regulate a particular module ( 21–23 in Table S1 ) . For each module , we picked GO categories – biological process and molecular function – that are significantly enriched in the module genes; and transcription factors ( TFs ) whose putative targets appear significantly in the module , based on the ChIP-chip binding data [24] . More precisely , we picked the GO categories and TFs whose hypergeometric p-value is below 10−3 after a false discovery rate ( FDR ) correction . Then , for a combination of a SNP and a module , we constructed three features based on whether: ( 1 ) the gene containing the SNP belongs to the module's GO process; ( 2 ) the gene containing the SNP belongs to the module's GO function; and ( 3 ) the SNP resides in the module's TF . In all cases , we took -log ( p-value ) to be the value of the regulatory feature , so that a regulator-module pair where the enrichment is highly significantly will have a higher-valued regulatory feature . Overall , for each SNP n , this process results in a set of 22 regulatory features and 87 ( for yeast ) / 48 ( for human ) features based on the gene function , listed in Table S1 . Based on the regulatory features of each individual SNP , we modeled the regulatory potential of each genetic marker , representing how likely sequence variations on the marker's chromosomal region regulate expression levels of genes . We also defined a regulatory potential for each e-regulator , representing how likely the regulator's expression is to regulate other genes' expression . These potentials are based on the regulatory potential of individual SNPs . We model the probability that each SNP n causes expression variation ( regulatory potential of n ) as: ( 1 ) where βk is the parameter called regulatory prior that determines the impact of each regulatory feature on the regulatory potential: higher values of βk encode the fact that the presence of the feature fnk increases the probability of having a regulatory effect . The learning algorithm of Lirnet automatically estimates the value of the β parameters from data . In our analysis , we focus only on regulatory features that are likely to increase the regulatory potential , and hence restrict βk to be non-negative; this assumption can easily be relaxed in the context of other feature sets . A SNP with many important regulatory features ( with high β's ) will have a higher regulatory prior , but the sigmoid function introduces a saturation effect , preventing the regulatory potential from increasing unboundedly and swamping the data . Due to linkage disequilibrium , each marker i can represent genotypes of the chromosomal region where it resides . We therefore define the regulatory potential of each genetic marker as an aggregate of the regulatory potentials of the individual SNPs in the corresponding chromosomal region . We assigned each SNP to the region associated with its nearest genotyped marker ( or tag SNP ) . Then , for each region r , we aggregated the contributions of all SNPs ( in the region ) , each modeled based on ( 1 ) , by summing them up and taking a sigmoid function: ( 2 ) Therefore , a region that contains a number of SNPs with high regulatory potentials is likely to have a high regulatory potential , but the outer-most sigmoid function again prevents it from increasing unboundedly . We note that a region that contains a large number of moderately relevant SNPs can also achieve a high regulatory potential . This method of aggregation tends to prefer regions with more SNPs , which is arguably justified , as they are also more likely to contain a causal polymorphism . However , other methods of aggregation are also plausible . We experimented with several other approaches; the one selected achieved the highest performance in prediction of expression profiles in test data not used for training the model . We also model the regulatory potential of candidate expression regulators based on their regulatory features . We used the regulatory features of SNPs ( Table S1 ) for constructing those of an expression regulator . The regulatory features consist of five categories: ( 1 ) 7 features each representing the number of SNPs in the gene region having one of the features 1 & 12–17 in Table S1; ( 2 ) 1 feature representing the conservation score of the gene region ( analogous to 18 in Table S1 ) ; ( 3 ) 1 binary feature indicating whether the gene is cis-regulated ( analogous to 19 in Table S1 ) ; ( 4 ) 87 ( for yeast ) / 48 ( for human ) binary features indicating whether the gene belongs to each of the GO categories listed in Table S12 ( analogous to 20 in Table S1 ) ; and ( 5 ) three pairwise binary features indicating whether the gene belongs to a GO process category enriched in the module , whether the gene belongs to a GO function category enriched in the module and whether the gene is the TF whose putative binding targets are enriched in the module ( analogous to 21–23 in Table S1 ) . We define the regulatory potential of r to be the probability that each candidate e-regulator r causes expression variation , which we model as follows: ( 3 ) where grk represents the k'th regulatory feature of e-regulator r ( explained above ) and αk is the weight assigned to the k'th regulatory feature . Lirnet attempts to reconstruct regulatory programs that define the regulatory interactions between each group of co-regulated genes ( called a module ) and its regulatory factors ( regulators ) . Candidate regulators of a module consist of binary genotype values of genetic markers and expression levels of genes that are not in the module . We model the expression level of each gene g in a module m ( denoted by ym , g ) as a linear combination of the potential regulators ( denoted by x1 , … , xn ) : ( 4 ) where ε represents a zero mean Gaussian noise , and x and y are standardized . Our objective is to estimate the weights ( wm , 1 , … , wm , n ) for each module m , from the data that best reflect the regulatory relationship between x's and y . More precisely , given x and y , we aim to construct the network by maximizing the joint log-likelihood Log P ( w , y|x ) = Log P ( y|x , w ) +Log P ( w ) , where for each module and its regulators P ( y|x , w ) ∼Ν ( Σiwixi , σ2 ) and P ( w ) represents the prior probability distribution of w . We model the prior probabilities on the weights based on the regulator's regulatory potential: For a regulator r , which can be either a region or an e-regulator , the prior probability is modeled as: ( 5 ) The regulatory potentials , Pr ( Regulator r is causal ) , are defined in ( 2 ) and ( 3 ) as functions of β and α , respectively . C0 and C1 represent the maximum and minimum regularization parameters Ci , respectively . The prior on the weight wr is an L1 prior , which tends to move the weights of less relevant coefficients to 0 [16]; the larger Cr , the stronger the bias towards 0 . As the regulatory potential increases , we have that Cr decreases , reducing the tendency of the learning algorithm to set the regulator's coefficient to 0 . To avoid the singularity problem ( when the number of parameters is greater than the number of data instances ) and a degeneracy problem that occurs when some of the x's are correlated , we introduced an additional L2 regularization term ( also called a ridge term ) with a regularization parameter D . The Lirnet algorithm estimates α , β , and w by solving the following optimization problem: ( 6 ) where Cr's are defined in ( 5 ) , θ = {α}∪{β} . Unlike previous approaches that use tree regression [7] , [62] or stepwise linear regression [48] , this approach deals well with correlated regulators . In particular , when several regulators are highly correlated , both tree regression and stepwise linear regression will select one representative within the set , often arbitrarily; with that regulator selected , the correlated regulators have little explanatory power , and will not be added to the regulator set . This approach is susceptible to making arbitrary decisions based on random fluctuations in the data . Lirnet uses an iterative coordinate descent algorithm to minimize the above objective function ( Figure 1 ) , iterating over two steps , where in one step we optimize over w's given the current α and β , and in the other step we optimize over α and β given the current w's . To learn θ , we solved the optimization problem of Eq . ( 6 ) with the current weights w's . To estimate w's , we modified least angle regression ( LARS ) [63] to allow it to handle the L2 regularization term ( the third term in ( 6 ) ) ; LARS is known to be one of the most efficient algorithms for solving this type of regularized regression problem . The regularization parameters C0 , C1 , D and E were determined through a 10 fold cross validation procedure: The arrays were divided into 10 groups; in each run , we train on 9/10 of the arrays , and compute predictive accuracy on the held out 1/10; the parameters were selected to maximize the average performance over the 10 runs . The final set of regulators was determined by using the chosen D and E over the entire set of arrays . Lirnet can be used both to construct a regulatory program for a pre-determined set of modules , or , as a step in an iterative procedure whereby modules are adapted dynamically to match the learned regulatory programs . In this iterative process , first developed in the module networks algorithm [14] , one starts with an initial assignment of genes to modules , for which regulatory programs are subsequently learned . Each gene is then reassigned to the module whose current regulatory program best explains its expression profile . We ran Lirnet both on a set of Geronemo modules , and using this iterative process initialized from these modules . The PGV analysis ( see below ) showed essentially no difference between these two runs ( data not shown ) . To facilitate comparison to the previous results , we therefore used the Lirnet program for the Geronemo modules . We applied Lirnet to the human eQTL dataset: genotype data from Phase II HapMap Project −60 European ( CEU ) and 60 African ( YRI ) individuals [56] – and expression profiles from the same individuals [4] . We treated SNPs on the Affymetrix GeneChip Human Mapping 100 k/ 500 k Array sets as genetic markers ( i . e . tagging SNPs ) , and assumed we observed the genotype of only those SNPs . The regulatory features were constructed for 6 , 515 , 224 SNPs downloaded from NCBI dbSNP database ( http://www . ncbi . nlm . nih . gov/SNP ) , based on various data sources such as dbSNP , UCSC genome browser ( http://www . genome . ucsc . edu/ ) , and gene ontology [64] . The list of regulatory features for the human data can be found in Table S3 . We divided the SNPs into regions corresponding to each of the tag SNP , assigning each gene to its closest tag SNP , and defined the regulatory potential of each individual SNP and each region , according to Eq ( 1 ) & ( 2 ) in Methods , similarly to the experiments on the yeast data . The learned regulatory prior is listed in Figure 2 and Table S3 . As a baseline method for comparison , we used the standard single-marker linkage model ( as in [1]–[3] ) . For each ( gene , marker ) pair , we performed a linear regression using the gene's expression level as a response variable and the marker as a predictor , and chose the marker that achieves the best fit . For the comparison on the yeast data , we used the published results [1] . We compared those results to those of our implementation of single-marker linkage model ( explained above ) , and the results were almost identical ( data not shown ) . For the human data , we used our implementation . We estimated the percentage of genetic variance ( PGV ) explained by the identified genetic regulators , following the procedure of Brem & Kruglyak , as also used for Geronemo [7] . In brief , we randomly divided the data of 112 segregants into a detection set ( training data ) and an estimation set ( test data ) . We used the method on the detection set to learn the regulation programs and ( where relevant ) the modules , and used the estimation set to calculate the PGV for these regulation programs . The PGV formula uses a corrected ANOVA , which automatically accounts for model complexity determined by the number of predictors . We repeated this process 10 times with different random splits of data , and estimated PGV of each gene by taking the average of its PGV over 10 runs . We note that , in the protocol of Brem & Kruglyak , the set of regulators is chosen on the detection set , but the actual parameters are estimated using ANOVA on the estimation set . Thus , there is a risk that more complex regulatory program will be able to overfit the training data , producing misleadingly good results . Although the number of parameters in our model is not larger than the number in the Geronemo model , we wanted to demonstrate directly that overfitting is not a factor in these results . We therefore also used an alternative PGV protocol , where the entire regulatory program – both the choice of regulators and the parameters – are derived from the detection set alone , and then the resulting model is estimated on the test set . In the results from this protocol ( Figure S1 ) , the proposed models also considerably outperformed Geronemo . We constructed a set of putative regulator-target pairs for the biological evaluation of the methods . We used five kinds of datasets: ( 1 ) deletion and over-expression microarrays [21] , [22]; ( 2 ) chromatin immunce-precipitation ( ChIP-chip ) binding experiments [23]; ( 3 ) mRNA binding pull-down experiments [31]; ( 4 ) transcription factor binding sites [65]; and ( 5 ) a literature-curated set of signaling interactions from the Proteome database ( http://www . proteome . com ) . For ( 1 ) , we considered as targets the genes whose expression changes are within the top 10% of all genes in terms of the magnitude of the expression change . For ( 2 ) , for each transcription factor , we picked the genes with a p-value of p<0 . 01 . For ( 3 ) , ( 4 ) and ( 5 ) , we downloaded the lists of putative targets suggested by the corresponding papers . For each of 13 regions that are identified to contain many candidate regulators of expression variation [1] , we constructed a list of genes that have experimental supports of their regulatory role on the targets linked to that region , based on microarray data from deletion experiments [1] , [9] , [21]–[23] and ChIP-chip binding data [23] ( see Figure 4B legend ) . For each region , we checked for each candidate regulator whether the candidate is “supported” by any of these data: ( 1 ) For the deletion microarray dataset , we considered the candidate regulator to be supported if there is a significant overlap ( p<0 . 01; hypergeometric distribution ) between its putative target based on the deletion data ( within top 20% of differentially expressed genes ) and the targets defined by the eQTL data [1]; ( 2 ) For the ChIP-chip binding data , we considered a regulator to be supported if there is a significant overlap between the putative targets with a binding significance of p<0 . 01 and the linked targets . To identify the causal SNP in a chromosomal region chosen by Lirnet as a g-regulator , we ranked each SNP using its regulatory potential , computed from its regulatory features and the learned coefficients ( as in Figure 2 ) . We then ranked the genes according to the SNP of highest regulatory potential in the gene region ( coding region , 500 bp upstream , 100 bp downstream ) . Unless stated , all S . cerevisiae strains used in this study are isogenic with a GAL2+ derivative of S288c [66] and are also isogenic to the BY parental strain [3] . Strains used to test the effect of deleting MKT1 in the RM background we constructed by transforming a URA3-marked deletion allele into RM11-1a . The “allele swap” strains ( RM mkt1-by ) , replacing the RM allele of MKT1 ( G30 , R453 ) with the BY allele ( D30 , K453 ) , were constructed by standard methods of PCR and transformation into the RM mkt1Δ::URA3 deletion strain . The presence of the appropriate variations and absence of any secondary mutations in the substituted region was confirmed by DNA sequence analysis . All strains were constructed by standard methods of PCR amplification and yeast transformation; details are available upon request . Microarray expression analysis of puf3Δ and mkt1Δ in the BY background used strains from the homozygous yeast deletion collection [67] ( Open Biosystems ) with the BY4743 isogenic parental strain as a control . Microarray expression analysis of mkt1Δ and mkt1-by in the RM strain background used the isogenic parental strain RM11-1a as a control . Strains containing Puf3 , Dhh1 , and Edc3 GFP protein fusions were taken from the collection described by Huh et al . [68] . Strains containing inframe protein fusions to the Red fluorescent protein , tdimer2 , were constructed using pKT176 [69] by standard methods of yeast methods of PCR amplification and yeast transformation into either the strains containing the GFP tagged protein or the isogenic wild-type strain BY4741 . Unless stated , all strains were grown in YPD medium and harvested in mid-log phase . Total yeast RNA was isolated by hot phenol method [70] . For both standard and tiling array analysis ( see Text S1 ) , total RNA was converted to cDNA and labeled with Alexa 647 and Alexa 555 ( Molecular Probes ) using the Atlas PowerScript Fluorescent Labeling Kit ( Clontech ) and an oligo ( dT ) primer as described by the manufacturer . Labeled cDNA samples were hybridized to either a stock yeast expression array ( Agilent-011445 Yeast Oligo Microarray G4140A ) or a custom yeast tiling array ( described below ) and processed according to manufacturer's instructions ( Agilent Technologies ) . Arrays were scanned using a ScanArray Express HT ( Perkin Elmer ) at a constant laser power of 90% and various photomultiplier tube gains as described in Dudley et al . ( 2002 ) . Signal and background intensities were measured using GENEPIX image analysis software ( Axon Instruments ) and data from multiple intensity scans were combined onto a common scale using the MASLINER linear regression method [71] . The log2 ratio of intensities of the signal and the background was calculated for each array element , and the standard normalization techniques described in Yang et al . [72] were applied to the log2 ratio values . We used global normalization and intensity-dependent normalization by using LOWESS ( locally weighted scatter-plot smoothing ) [73] with parameters relevant to our experimental setting , single slides and single print tips . Live yeast cells containing GFP and tdimer2 fluorescently tagged proteins were visualized with a Nikon Eclipse TE2000-E inverted microscope under 100× objective with oil . GFP was detected using a FITC filter , and tdimer2 was detected using HCRed1 . Images were captured using a Hamamatsu Orca-ER CCD digital camera . Image capture and analysis used Metamorph 6 . 3R5 and Adobe Photoshop software . P-bodies , observed as bright punctate spots in the cytoplasm of cells containing a fluorescently labeled P-body protein , form in live cells after approximately 10 minutes in water or minimal medium lacking glucose under a microscope coverslip . Under the same conditions after approximately 12 minutes , a Puf3-GFP fusion protein formed similar fluorescent spots ( Figure S5 ) , most of which overlapped the P-bodies ( Table S10 ) . When present in the same cells , punctate spots of Puf3-GFP fluorescence overlap with the punctate pattern formed by known P-body components ( 69/75 = 92% of P-body spots are also Puf3 spots ) , showing localization of Puf3 to P-bodies ( Table S10 ) . Of the ten target genes in the peroxisome module , the proximity and orientation of one ( YAL049C ) suggested that its co-expression could be the result of cross hybridization to the OAF1 probe in the original microarray data; thus , it was removed from further consideration . To evaluate the dependence of the remaining target genes on Oaf1 , we examined a published microarray dataset [29] comparing RNA expression oaf1Δ to a wild-type ( BY ) strain in the presence of oleate ( an inducing condition ) . This dataset also included an estimate of the likelihood of differential expression [74] . We sorted RNA expression levels by the log10 ratios and filtered for λ values greater than 36 . 23 to arrive at the top 1% ( 63 ) most significantly down regulated genes ( Table S11 ) . The revised RM11-1a PUF3 DNA sequence determined by this study will be deposited in GenBank ( NCBI ) prior to publication . All Microarray datasets will be deposited in the GEO database prior to publication . | Gene expression data of genetically diverse individuals ( eQTL data ) provide a unique perspective on the effect of genetic variation on cellular pathways . However , the burden of multiple hypotheses , combined with the challenges of linkage disequilibrium , makes it difficult to correctly identify causal polymorphisms . Researchers traditionally apply heuristics for selecting among plausible hypotheses , favoring polymorphisms that are more conserved , that lead to significant amino acid change , or that reside in genes whose function is related to that of the targets . But how do we know how much weight to attribute to different regulatory features ? We describe Lirnet , which learns from eQTL data how to weight regulatory features and induce a regulatory potential for sequence variations . Lirnet assesses these weights simultaneously to learning a regulatory network , finding weights that lead to a more predictive network . We show that Lirnet constructs high-accuracy regulatory programs and demonstrate its ability to correctly identify causative polymorphisms . Lirnet can flexibly use any regulatory features , including sequence features that are available for any sequenced organism , and automatically learn their weights in a dataset-specific way . This feature makes it especially advantageous for mammalian systems , where many forms of prior knowledge used in simple model organisms are incomplete or unavailable . | [
"Abstract",
"Introduction",
"Results",
"Discussion",
"Methods"
] | [
"computational",
"biology/genomics",
"computational",
"biology"
] | 2009 | Learning a Prior on Regulatory Potential from eQTL Data |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.